repo
stringclasses 1
value | number
int64 1
25.3k
| state
stringclasses 2
values | title
stringlengths 1
487
| body
stringlengths 0
234k
⌀ | created_at
stringlengths 19
19
| closed_at
stringlengths 19
19
| comments
stringlengths 0
293k
|
|---|---|---|---|---|---|---|---|
transformers | 5,900 | closed | Is there any api for intermediate layer outputs? | I would like to get all 12 Bertlayer outputs. Can you please help? | 07-20-2020 02:42:37 | 07-20-2020 02:42:37 | You need to pass `output_hidden_states=True` when calling your model, or set this option when instantiating it. For instance, if you're using a pretrained model:
```
from transformers import DistilBertModel
model = DistilBertModel.from_pretrained('distilbert-base-uncased', output_hidden_states=True)
``` |
transformers | 5,899 | closed | Smaller output vocabulary for GPT-2 | # ❓ Questions & Help
I noticed that by default, GPT2LMHeadModel returns prediction scores of shape (batch_size, sequence_length, config.vocab_size) ([docs link](https://huggingface.co/transformers/model_doc/gpt2.html#gpt2lmheadmodel)). Is there any way for me to limit the output vocabulary to only a subset of words?
I want to take the existing weights from GPT-2, but re-train a new top linear layer with a smaller vocabulary. I suppose I could mask the logits at the end, but then it feels like a waste of computational power to even predict them.
**A link to [original question on the forum](https://discuss.huggingface.co/t/smaller-output-vocabulary-for-gpt-2/366)** | 07-20-2020 02:03:02 | 07-20-2020 02:03:02 | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,898 | closed | Fix pack_dataset.py | right now it just sorts the examples and prints an incorrect logger message! | 07-20-2020 00:29:06 | 07-20-2020 00:29:06 | |
transformers | 5,897 | closed | examples/seq2seq: add label_smoothing cross entropy option | Started! | 07-20-2020 00:28:20 | 07-20-2020 00:28:20 | |
transformers | 5,896 | closed | MT: automate/experiment with pruning embeddings | This isn't really a github issue, but follow pytorch/fairseq#2120 and see if it works. If it does, make it easy to use! | 07-20-2020 00:27:42 | 07-20-2020 00:27:42 | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,895 | closed | examples/seq2seq/finetune.py and BART supports TPU | - [ ] test the code on tpu
- [ ] if it doesn't work well: change code as little as possible to get it working.
- [ ] add a script/command that works | 07-20-2020 00:26:12 | 07-20-2020 00:26:12 | Do you want to wait for a stable release for the torch-xla? Also shouldn't the end user pass in "(num_tpu_cores=8)" when they are creating the lightning's trainer. It should automatically handle the rest for us i think. Also we have "bloat16" for TPUs as well.<|||||>It totally could work out of the box. In which case this issue could be as simple as running a shell command on a tpu machine, seeing that it works well, and then checking in the working command or commenting here :).
<|||||>Alright cool! Let's see if I can make an attempt to do it! Very new to this examples section of hugging face. <|||||>Hey! So i tried running it on the [colab](https://colab.research.google.com/drive/16q2GWrnZ0Tjg1OxJQUcaWKCWwn3Jh5z0?usp=sharing) first, but it seems there's some error wrt bart model. The traceback it threw, (another issue open as well https://github.com/huggingface/transformers/issues/5915).
To reproduce the same, I executed the finetune.sh with
```bash
!sh finetune.sh \
--data_dir /content/xsum \
--model_name_or_path facebook/bart-base \
--output_dir=xsum_results \
--train_batch_size=2 \
--eval_batch_size=2 \
--num_train_epochs 1 \
--n_tpu_cores 8 \
--tpu_cores 8
```
And modified the args in the shell snip when we invoke the python finetune.py (removed fp16 and gpus to 0)
```
Attempted to call `variable.set_data(tensor)`, but `variable` and `tensor` have incompatible tensor type.
Traceback (most recent call last):
File "/usr/local/lib/python3.6/dist-packages/torch_xla/distributed/xla_multiprocessing.py", line 119, in _start_fn
fn(gindex, *args)
File "/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/distrib_parts.py", line 222, in tpu_train
self.run_pretrain_routine(model)
File "/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/trainer.py", line 1196, in run_pretrain_routine
False)
File "/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/evaluation_loop.py", line 293, in _evaluate
output = self.evaluation_forward(model, batch, batch_idx, dataloader_idx, test_mode)
File "/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/evaluation_loop.py", line 470, in evaluation_forward
output = model.validation_step(*args)
File "finetune.py", line 145, in validation_step
return self._generative_step(batch)
File "finetune.py", line 176, in _generative_step
decoder_start_token_id=self.decoder_start_token_id,
File "/usr/local/lib/python3.6/dist-packages/torch/autograd/grad_mode.py", line 15, in decorate_context
return func(*args, **kwargs)
File "/usr/local/lib/python3.6/dist-packages/transformers/generation_utils.py", line 248, in generate
if self.get_output_embeddings() is None:
File "/usr/local/lib/python3.6/dist-packages/transformers/modeling_bart.py", line 1113, in get_output_embeddings
return _make_linear_from_emb(self.model.shared) # make it on the fly
File "/usr/local/lib/python3.6/dist-packages/transformers/modeling_bart.py", line 190, in _make_linear_from_emb
lin_layer.weight.data = emb.weight.data
```<|||||>Makes sense. You could try to instantiate
```python
self.lm_head = _make_linear_from_emb(self.model.shared)
```
in BartForConditionalGeneration.__init__
and then have `get_output_embeddings` return `self.lm_head`.
<|||||>Alright! Thanks; Will keep you posted 🎉;
Edit -1:
So things seems to work after i make the changes you said. The trouble is that training seems to be frozen on TPUs.<|||||>`emb.weight.data` is...
```
tensor([[-0.0370, 0.1117, 0.1829, ..., 0.2054, 0.0578, -0.0750],
[ 0.0055, -0.0049, -0.0069, ..., -0.0030, 0.0038, 0.0087],
[-0.0448, 0.4604, -0.0604, ..., 0.1073, 0.0310, 0.0477],
...,
[-0.0138, 0.0278, -0.0467, ..., 0.0455, -0.0265, 0.0125],
[-0.0043, 0.0153, -0.0567, ..., 0.0496, 0.0108, -0.0099],
[ 0.0053, 0.0324, -0.0179, ..., -0.0085, 0.0223, -0.0020]],
device='xla:1')
```
...and `lin_layer.weight.data` is...
```
tensor([[-1.0449e-03, 4.0973e-03, -9.7727e-04, ..., 8.2363e-04,
-3.2153e-03, 3.5317e-03],
[ 2.3644e-03, 3.5527e-03, -1.2428e-03, ..., -1.0983e-04,
-2.1916e-03, 5.3099e-05],
[-4.2492e-03, 3.8183e-04, 3.2527e-03, ..., -4.4359e-03,
7.6555e-04, -4.1728e-03],
...,
[-4.3412e-03, 2.8537e-03, 7.9720e-04, ..., 2.9499e-03,
2.6357e-03, -3.5283e-03],
[ 3.7042e-03, -3.0546e-03, 3.9206e-03, ..., -2.3771e-03,
4.3551e-03, 1.1703e-04],
[ 3.5616e-03, -3.1224e-03, 1.3898e-03, ..., -2.1096e-05,
5.4077e-04, 1.6183e-03]])
```
... and you try to do...
```
lin_layer.weight.data = emb.weight.data
```
Isn't the problem that they are on different devices?<|||||>We can remove the thing which does it on fly(refer to sshleifer comment above). That won't work in case of TPUs.<|||||>Good to know. Once you get everything working it would be great to have all the required changes consolidated into one PR.<|||||>I am not quite sure that lightning does this optimization when we use multiple TPU cores (only available at the nightly-xla's). Refer [here](https://github.com/pytorch/xla/issues/1870#issuecomment-623603323).<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,894 | closed | fix typo in training_args_tf.py | 07-19-2020 21:08:11 | 07-19-2020 21:08:11 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=h1) Report
> Merging [#5894](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/09a2f40684f77e62d0fd8485fe9d2d610390453f&el=desc) will **decrease** coverage by `1.19%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5894 +/- ##
==========================================
- Coverage 78.49% 77.30% -1.20%
==========================================
Files 146 146
Lines 26210 26210
==========================================
- Hits 20573 20261 -312
- Misses 5637 5949 +312
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/training\_args\_tf.py](https://codecov.io/gh/huggingface/transformers/pull/5894/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmluZ19hcmdzX3RmLnB5) | `47.45% <ø> (ø)` | |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5894/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.39%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5894/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.71% <0.00%> (+0.50%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5894/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=footer). Last update [09a2f40...b9754e4](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,893 | closed | fix typo in training_args.py | 07-19-2020 21:08:05 | 07-19-2020 21:08:05 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=h1) Report
> Merging [#5893](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/09a2f40684f77e62d0fd8485fe9d2d610390453f&el=desc) will **increase** coverage by `0.17%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5893 +/- ##
==========================================
+ Coverage 78.49% 78.66% +0.17%
==========================================
Files 146 146
Lines 26210 26210
==========================================
+ Hits 20573 20619 +46
+ Misses 5637 5591 -46
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/training\_args.py](https://codecov.io/gh/huggingface/transformers/pull/5893/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmluZ19hcmdzLnB5) | `77.55% <ø> (ø)` | |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5893/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5893/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `95.53% <0.00%> (+69.51%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=footer). Last update [09a2f40...762ff74](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,892 | closed | Benchmark: traceback does not describe real problem | To reproduce:
- execute a command that results in 'N/A'
- long traceback describes tuple unpacking. The real problem is what caused the N/A and is far above in a small logger.error statement.
Example Command:
```bash
python examples/benchmarking/run_benchmark.py --models facebook/mbart-large-en-ro
```
Traceback:
```python
1 / 1
ERROR:transformers.benchmark.benchmark_utils:<class 'transformers.configuration_bart.MBartConfig'>
<class 'transformers.configuration_bart.MBartConfig'>
Traceback (most recent call last):
File "examples/benchmarking/run_benchmark.py", line 29, in <module>
main()
File "examples/benchmarking/run_benchmark.py", line 25, in main
benchmark.run()
File "/home/shleifer/transformers_fork/src/transformers/benchmark/benchmark_utils.py", line 665, in run
memory, inference_summary = self.inference_memory(model_name, batch_size, sequence_length)
ValueError: too many values to unpack (expected 2)
```
This may not be worth a fix, and probably won't be fixed soon, but posting for others who might run into this or have ideas for a fix. | 07-19-2020 19:07:04 | 07-19-2020 19:07:04 | Working command after MbartConfig change:
```bash
export d=mbart_benchmark_data
python examples/benchmarking/run_benchmark.py \
--models facebook/mbart-large-en-ro \
--log_filename $d/log.txt \
--inference_memory_csv \
$d/inference_memory.csv \
--train_memory_csv $d/train_memory.csv \
--train_time_csv $d/train_time.csv \
--inference_time_csv $d/inference_time.csv \
--fp16 --log_print --training --save_to_csv \
--batch_sizes 4 8 12 16
```
<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,891 | closed | [WIP] Fix mbart benchmark | - [ ] undo benchmarking changes
- [ ] publish working command + results | 07-19-2020 18:50:38 | 07-19-2020 18:50:38 | fixed indirectly by #6441 |
transformers | 5,890 | closed | How to finetune distillbart from distilbart-cnn-12-6 checkpoint using cnn_daily mail or gigawords dataset? | I would like to ask about how to finetune distillbart on gigaword and cnn dailymail with the starting checkpoint distilbart-cnn-12-6.
I did use the gigaword dataset provided by tensorflow but it replaces numbers by this character: "#", as a result, my summaries have # instead of numbers, is it normal that it has those # ?
Also is it really possible to finetune distillbart from the checkpoint distilbart-cnn-12-6 with cnn daily mail?
| 07-19-2020 14:53:43 | 07-19-2020 14:53:43 | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,889 | closed | AttributeError: 'GPT2LMHeadModel' object has no attribute 'h' | # ❓ Questions & Help
when I run run_generation.py（I use TF1.14 fine tuned gpt2）,the program is broken with the error message AttributeError: 'GPT2LMHeadModel' object has no attribute 'h'
Environment:
WIN10
pytorch 1.3.1
python 3.6
Part of the code has been modified
| 07-19-2020 07:17:44 | 07-19-2020 07:17:44 | You can't load directly a TF checkpoint if it's not been generated by the TensorFlow side of the library (which can't be the case here since you're using a TF1 checkpoint and the lib requires TF2.0). You probably have to use the [conversion script](https://huggingface.co/transformers/converting_tensorflow_models.html).<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,888 | closed | Reading transformer package from local codes and NOT the pip installed version | Hi @sshleifer ,
There seem to be a problem of python package structure. I am getting an error similar to a previous issue (https://github.com/huggingface/transformers/issues/5303) but with the token_classification/run_ner.py file.
File "examples/token-classification/run_ner.py", line 42, in
from modeling_auto import AutoModelForTokenClassification
File "transformers/src/transformers/modeling_auto.py", line 22, in
from .configuration_auto import (
ImportError: attempted relative import with no known parent package
I have not installed transformers library using pip because I want to use the local codes (cloned from transformers library). After reading various stackoverflow suggestions (https://stackoverflow.com/questions/16981921/relative-imports-in-python-3 and https://napuzba.com/a/import-error-relative-no-parent), I believe that when I am importing the transformer package locally from my own directory, then it is not able read+load transformer as a package.
I am using python3.7
Can you please suggest how to read transformer as a package from local codes.
Thanks... | 07-19-2020 03:28:24 | 07-19-2020 03:28:24 | How did you install the package ?
According to the [README](https://github.com/huggingface/transformers#from-source) you should do :
```console
git clone https://github.com/huggingface/transformers
cd transformers
pip install -e .
```
<|||||>Hi ,
Apologies for raising this issue, the error was something specific to singularity images that I have been using to run transformers on HPC.
Please pardon this issue.
Anyway thanks @Colanim. I knew the steps, but the error was from somewhere else.
|
transformers | 5,887 | closed | Seq2Seq: same MultiGPU test failing twice! | This is not run by CI, but if you go on a machine with multiple GPU, you will see an interesting unittest failure:
```bash
pytest examples -k multigpu
```
```
===================================== FAILURES ======================================
_____________________ TestSummarizationDistiller.test_multigpu ______________________
self = <seq2seq.test_seq2seq_examples.TestSummarizationDistiller testMethod=test_multigpu>
@require_multigpu
def test_multigpu(self):
updates = dict(no_teacher=True, freeze_encoder=True, gpus=2, sortish_sampler=False,)
> self._test_distiller_cli(updates)
examples/seq2seq/test_seq2seq_examples.py:117:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
examples/seq2seq/test_seq2seq_examples.py:182: in _test_distiller_cli
self.assertIn(ckpt_name, contents)
E AssertionError: 'val_avg_rouge2=0.0000-step_count=2.ckpt' not found in {'metrics.json', 'hparams.pk
l', 'git_log.json'}
============================== short test summary info ==============================
FAILED examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller::test_multigpu
========================= 1 failed, 25 deselected in 9.75s ==========================
...
'step_count': 4,
'test_avg_gen_time': 0.9470310211181641,
'test_avg_loss': 10.737613677978516,
'test_avg_rouge1': 0.0,
'test_avg_rouge2': 0.0,
'test_avg_rougeL': 0.0,
'test_avg_summ_len': 141.0,
'test_loss': tensor(10.7376, device='cuda:0'),
'test_rouge2': tensor(0., device='cuda:0')}
--------------------------------------------------------------------------------
Testing: 100%|█████████████████████████████████████████| 1/1 [00:01<00:00, 1.49s/it]
F
===================================== FAILURES ======================================
_____________________ TestSummarizationDistiller.test_multigpu ______________________
self = <seq2seq.test_seq2seq_examples.TestSummarizationDistiller testMethod=test_multigpu>
@require_multigpu
def test_multigpu(self):
updates = dict(no_teacher=True, freeze_encoder=True, gpus=2, sortish_sampler=False,)
> self._test_distiller_cli(updates)
examples/seq2seq/test_seq2seq_examples.py:117:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
examples/seq2seq/test_seq2seq_examples.py:193: in _test_distiller_cli
self.assertEqual(len(metrics["val"]), desired_n_evals)
E AssertionError: 3 != 5
============================== short test summary info ==============================
FAILED examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller::test_multigpu
========================= 1 failed, 25 deselected in 47.48s =========================
```
The same `SummarizationDistiller.test_multigpu` fails twice. I think we need to use `pl.Trainer.argparse.from_args` in the unittests.
How do you guys do it @nateraw ? | 07-19-2020 01:01:06 | 07-19-2020 01:01:06 | Spent an hour testing the new version on real data:
(1) `./train_mbart_enro_multigpu` works!
(2) `./train_mbart_enro_multigpu --logger_name wandb` ~doesn't work~ works.
(3) trainer.test seems to try to call `main` again, and then runs into an "output directory" already exists error.
It hangs for me with versions:
```bash
pytorch-lightning==0.8.5
torch==1.5.1+cu101
```
cc @nateraw @williamFalcon <|||||>@stas00 do you think multi-gpu test coverage (at least on machines with multiple GPU, not CI yet) is possible?<|||||>> @stas00 do you think multi-gpu test coverage (at least on machines with multiple GPU, not CI yet) is possible?
Are you referring to `pytest --cov`? Yes, of course, it should work.
The main issue with multiprocessing and coverage is a potential unregistered coverage if the sub-process hasn't finished its work right away and is hanging around. But we already have this issue in the benchmarking tests - I need to go back and pound at it some more.<|||||>Sorry, I'm referring to getting that test to pass.
At the moment it fails (twice!).<|||||>Oh, I see what you meant. Let me investigate.<|||||>Yeah, multigpu is a problem. Here is a totally different manifestation of it - hanging and other subtests failing:
```
pytest --disable-warnings examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller
――――――――――――――――――――――――――――――――――――――――― TestSummarizationDistiller.test_distill_checkpointing_with_teacher ―――――――――――――――――――――――――――――――――――――――――
self = <seq2seq.test_seq2seq_examples.TestSummarizationDistiller testMethod=test_distill_checkpointing_with_teacher>
def test_distill_checkpointing_with_teacher(self):
updates = dict(
student_encoder_layers=2,
student_decoder_layers=1,
max_epochs=4,
val_check_interval=0.25,
alpha_hid=2.0,
model_name_or_path="IGNORE_THIS_IT_DOESNT_GET_USED",
)
model = self._test_distiller_cli(updates, check_contents=False)
ckpts = list(Path(model.output_dir).glob("*.ckpt"))
> self.assertEqual(1, len(ckpts))
E AssertionError: 1 != 0
examples/seq2seq/test_seq2seq_examples.py:173: AssertionError
examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller.test_distill_checkpointing_with_teacher ⨯ 14% █▌ using module BartTranslationDistiller
Epoch 2: 100%|████████████████████████████████████████████████████████████████████████████████████| 4/4 [00:00<00:00, 11.53it/s, loss=2.970, v_num=19]
――――――――――――――――――――――――――――――――――――――――――――――――――― TestSummarizationDistiller.test_distill_mbart ――――――――――――――――――――――――――――――――――――――――――――――――――――
self = <seq2seq.test_seq2seq_examples.TestSummarizationDistiller testMethod=test_distill_mbart>
def test_distill_mbart(self):
updates = dict(
student_encoder_layers=2,
student_decoder_layers=1,
num_train_epochs=4,
val_check_interval=0.25,
alpha_hid=2.0,
task="translation",
model_name_or_path="IGNORE_THIS_IT_DOESNT_GET_USED",
tokenizer_name=MBART_TINY,
teacher=MBART_TINY,
src_lang="en_XX",
tgt_lang="ro_RO",
)
model = self._test_distiller_cli(updates, check_contents=False)
assert model.model.config.model_type == "mbart"
ckpts = list(Path(model.output_dir).glob("*.ckpt"))
> self.assertEqual(1, len(ckpts))
E AssertionError: 1 != 0
examples/seq2seq/test_seq2seq_examples.py:224: AssertionError
examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller.test_distill_mbart ⨯ 29% ██▉ using module SummarizationModule
Epoch 2: 100%|███████████████████████████████████████████████████████████████████████████████████| 4/4 [00:00<00:00, 4.23it/s, loss=10.785, v_num=19]
――――――――――――――――――――――――――――――――――――――――――――――――― TestSummarizationDistiller.test_distill_no_teacher ―――――――――――――――――――――――――――――――――――――――――――――――――
self = <seq2seq.test_seq2seq_examples.TestSummarizationDistiller testMethod=test_distill_no_teacher>
def test_distill_no_teacher(self):
updates = dict(student_encoder_layers=2, student_decoder_layers=1, no_teacher=True)
> self._test_distiller_cli(updates)
examples/seq2seq/test_seq2seq_examples.py:159:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
self = <seq2seq.test_seq2seq_examples.TestSummarizationDistiller testMethod=test_distill_no_teacher>
updates = {'no_teacher': True, 'student_decoder_layers': 1, 'student_encoder_layers': 2}, check_contents = True
def _test_distiller_cli(self, updates, check_contents=True):
default_updates = dict(
label_smoothing=0.0,
early_stopping_patience=-1,
train_batch_size=1,
eval_batch_size=2,
max_epochs=2,
alpha_mlm=0.2,
alpha_ce=0.8,
do_predict=True,
model_name_or_path="sshleifer/tinier_bart",
teacher=CHEAP_ARGS["model_name_or_path"],
val_check_interval=0.5,
alpha_encoder_loss=0.4,
)
default_updates.update(updates)
args_d: dict = CHEAP_ARGS.copy()
tmp_dir = make_test_data_dir()
output_dir = tempfile.mkdtemp(prefix="output_")
args_d.update(data_dir=tmp_dir, output_dir=output_dir, **default_updates)
model = distill_main(argparse.Namespace(**args_d))
if not check_contents:
return model
contents = os.listdir(output_dir)
contents = {os.path.basename(p) for p in contents}
ckpt_files = [p for p in contents if p.endswith("ckpt")]
> assert len(ckpt_files) > 0
E AssertionError: assert 0 > 0
E + where 0 = len([])
examples/seq2seq/test_seq2seq_examples.py:271: AssertionError
examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller.test_distill_no_teacher ⨯ 43% ████▍
examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller.test_distill_t5 s 57% █████▊
../../../../../home/stas/anaconda3/envs/main-38/lib/python3.8/unittest/case.py::TestSummarizationDistiller.test_hub_configs s 71% ███████▎
examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller.test_loss_fn ✓ 86% ████████▋
examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller.test_multigpu s 100% ██████████
============================================================== short test summary info ===============================================================
FAILED examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller::test_distill_checkpointing_with_teacher - AssertionError: 1 != 0
FAILED examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller::test_distill_mbart - AssertionError: 1 != 0
FAILED examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller::test_distill_no_teacher - AssertionError: assert 0 > 0
Results (18.02s):
1 passed
3 failed
- examples/seq2seq/test_seq2seq_examples.py:161 TestSummarizationDistiller.test_distill_checkpointing_with_teacher
- examples/seq2seq/test_seq2seq_examples.py:206 TestSummarizationDistiller.test_distill_mbart
- examples/seq2seq/test_seq2seq_examples.py:157 TestSummarizationDistiller.test_distill_no_teacher
3 skipped
```
and it hangs. getting a trace:
```
Thread 0x00007fc81efd6700 (most recent call first):
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 306 in wait
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/queue.py", line 179 in get
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/tensorboard/summary/writer/event_file_writer.py", line 232 in run
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 932 in _bootstrap_inner
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 890 in _bootstrap
Thread 0x00007fc80e850700 (most recent call first):
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 306 in wait
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/queue.py", line 179 in get
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/tensorboard/summary/writer/event_file_writer.py", line 232 in run
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 932 in _bootstrap_inner
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 890 in _bootstrap
Thread 0x00007fc8248b0700 (most recent call first):
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 306 in wait
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/queue.py", line 179 in get
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/tensorboard/summary/writer/event_file_writer.py", line 232 in run
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 932 in _bootstrap_inner
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 890 in _bootstrap
Thread 0x00007fc8250b1700 (most recent call first):
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 306 in wait
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/queue.py", line 179 in get
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/tensorboard/summary/writer/event_file_writer.py", line 232 in run
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 932 in _bootstrap_inner
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 890 in _bootstrap
Thread 0x00007fc81cbec700 (most recent call first):
<no Python frame>
Thread 0x00007fc81e7d5700 (most recent call first):
<no Python frame>
Thread 0x00007fc826ffd700 (most recent call first):
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 306 in wait
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/queue.py", line 179 in get
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/tensorboard/summary/writer/event_file_writer.py", line 232 in run
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 932 in _bootstrap_inner
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py", line 890 in _bootstrap
Current thread 0x00007fc91d2e2740 (most recent call first):
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/torch/distributed/rendezvous.py", line 172 in _env_rendezvous_handler
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 423 in init_process_group
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pytorch_lightning/core/lightning.py", line 973 in init_ddp_connection
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pytorch_lightning/trainer/distrib_data_parallel.py", line 506 in ddp_train
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pytorch_lightning/trainer/distrib_data_parallel.py", line 462 in spawn_ddp_children
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py", line 992 in fit
File "/mnt/nvme1/code/huggingface/transformers-master/examples/lightning_base.py", line 380 in generic_train
File "/mnt/nvme1/code/huggingface/transformers-master/examples/seq2seq/finetune.py", line 420 in main
File "/mnt/nvme1/code/huggingface/transformers-master/examples/seq2seq/distillation.py", line 511 in distill_main
File "/mnt/nvme1/code/huggingface/transformers-master/examples/seq2seq/test_seq2seq_examples.py", line 267 in _test_distiller_cli
File "/mnt/nvme1/code/huggingface/transformers-master/examples/seq2seq/test_seq2seq_examples.py", line 157 in test_multigpu
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/unittest/case.py", line 633 in _callTestMethod
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/unittest/case.py", line 676 in run
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/unittest/case.py", line 736 in __call__
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/unittest.py", line 278 in runtest
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py", line 153 in pytest_runtest_call
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/callers.py", line 187 in _multicall
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py", line 84 in <lambda>
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py", line 93 in _hookexec
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/hooks.py", line 286 in __call__
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py", line 247 in <lambda>
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py", line 294 in from_call
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py", line 246 in call_runtest_hook
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py", line 207 in call_and_report
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py", line 117 in runtestprotocol
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py", line 100 in pytest_runtest_protocol
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/callers.py", line 187 in _multicall
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py", line 84 in <lambda>
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py", line 93 in _hookexec
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/hooks.py", line 286 in __call__
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/main.py", line 321 in pytest_runtestloop
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/callers.py", line 187 in _multicall
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py", line 84 in <lambda>
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py", line 93 in _hookexec
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/hooks.py", line 286 in __call__
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/main.py", line 296 in _main
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/main.py", line 240 in wrap_session
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/main.py", line 289 in pytest_cmdline_main
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/callers.py", line 187 in _multicall
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py", line 84 in <lambda>
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py", line 93 in _hookexec
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/hooks.py", line 286 in __call__
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/config/__init__.py", line 157 in main
File "/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/config/__init__.py", line 180 in console_main
File "/home/stas/anaconda3/envs/main-38/bin/pytest", line 8 in <module>
```
If I run just these 3 tests - all is green. So its the multigpu test.
Ah, it somehow starts running pytest second time half-way through.
Rechecked with py37 and py38 - same, same.
p.s. checkout `pip install pytest-sugar` - makes running tests a brighter experience.<|||||>I will try to make a small reproducible test where the problem still happens, and then we can try to sort it out.<|||||>OK, so first there is some strange issue with the properties - seems to be related to https://github.com/pytorch/pytorch/issues/43542, the multigpu test fails with:
```
―――――――――――――――――――――――――――――――――――――――――――――――――――――― TestSummarizationDistiller.test_multigpu ――――――――――――――――――――――――――――――――――――――――――――――――――――――
self = <seq2seq.test_seq2seq_examples.TestSummarizationDistiller testMethod=test_multigpu>
@require_multigpu
def test_multigpu(self):
updates = dict(
no_teacher=True,
freeze_encoder=True,
gpus=2,
sortish_sampler=True,
)
> self._test_distiller_cli(updates, check_contents=False)
examples/seq2seq/test_seq2seq_examples.py:154:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
examples/seq2seq/test_seq2seq_examples.py:264: in _test_distiller_cli
model = distill_main(argparse.Namespace(**args_d))
examples/seq2seq/distillation.py:495: in distill_main
return ft_main(args, model=model)
examples/seq2seq/finetune.py:399: in main
trainer: pl.Trainer = generic_train(
examples/lightning_base.py:380: in generic_train
trainer.fit(model)
../../github/00pytorch/pytorch-lightning/pytorch_lightning/trainer/trainer.py:301: in fit
results = self.accelerator_backend.train()
../../github/00pytorch/pytorch-lightning/pytorch_lightning/accelerators/ddp_backend.py:133: in train
self.ddp_train_tmp(process_idx=self.task_idx, mp_queue=None, model=model)
../../github/00pytorch/pytorch-lightning/pytorch_lightning/accelerators/ddp_base_backend.py:148: in ddp_train_tmp
optimizers, lr_schedulers, optimizer_frequencies = self.trainer.init_optimizers(model)
../../github/00pytorch/pytorch-lightning/pytorch_lightning/trainer/optimizers.py:32: in init_optimizers
optim_conf = model.configure_optimizers()
examples/lightning_base.py:152: in configure_optimizers
scheduler = self.get_lr_scheduler()
examples/lightning_base.py:122: in get_lr_scheduler
self.opt, num_warmup_steps=self.hparams.warmup_steps, num_training_steps=self.total_steps
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
self = SummarizationModule(
(model): BartForConditionalGeneration(
(model): BartModel(
(shared): Embedding(50265,... )
)
(layernorm_embedding): LayerNorm((24,), eps=1e-05, elementwise_affine=True)
)
)
)
)
name = 'total_steps'
def __getattr__(self, name: str) -> Union[Tensor, 'Module']:
if '_parameters' in self.__dict__:
_parameters = self.__dict__['_parameters']
if name in _parameters:
return _parameters[name]
if '_buffers' in self.__dict__:
_buffers = self.__dict__['_buffers']
if name in _buffers:
return _buffers[name]
if '_modules' in self.__dict__:
modules = self.__dict__['_modules']
if name in modules:
return modules[name]
> raise ModuleAttributeError("'{}' object has no attribute '{}'".format(
type(self).__name__, name))
E torch.nn.modules.module.ModuleAttributeError: 'SummarizationModule' object has no attribute 'total_steps'
```
I fixed it with:
```
diff --git a/examples/lightning_base.py b/examples/lightning_base.py
index e7c41a3e..f5e858c7 100644
--- a/examples/lightning_base.py
+++ b/examples/lightning_base.py
@@ -119,8 +119,7 @@ class BaseTransformer(pl.LightningModule):
def get_lr_scheduler(self):
get_schedule_func = arg_to_scheduler[self.hparams.lr_scheduler]
scheduler = get_schedule_func(
- self.opt, num_warmup_steps=self.hparams.warmup_steps, num_training_steps=self.total_steps
- )
+ self.opt, num_warmup_steps=self.hparams.warmup_steps, num_training_steps=self.total_steps())
scheduler = {"scheduler": scheduler, "interval": "step", "frequency": 1}
return scheduler
@@ -159,12 +158,11 @@ class BaseTransformer(pl.LightningModule):
def test_epoch_end(self, outputs):
return self.validation_end(outputs)
- @property
def total_steps(self) -> int:
"""The number of total training steps that will be run. Used for lr scheduler purposes."""
num_devices = max(1, self.hparams.gpus) # TODO: consider num_tpu_cores
effective_batch_size = self.hparams.train_batch_size * self.hparams.accumulate_grad_batches * num_devices
- dataset_size = len(self.train_loader.dataset)
+ dataset_size = len(self.train_dataloader().dataset)
return (dataset_size / effective_batch_size) * self.hparams.max_epochs
def setup(self, mode):
```
it works fine the first time, but when it re-reruns itself, it behaves differently - so the above fix was needed.<|||||>WIP: https://github.com/huggingface/transformers/pull/7281
|
transformers | 5,886 | closed | DataCollatorForLanguageModeling - Shift labels for left-to-right LM? | Hi,
Shouldn't the labels be shifted 1 step right in the data collation step at -https://github.com/huggingface/transformers/blob/09a2f40684f77e62d0fd8485fe9d2d610390453f/src/transformers/data/data_collator.py#L88
| 07-19-2020 00:30:53 | 07-19-2020 00:30:53 | Actually never mind. At least GPT2 does the shifting internally. |
transformers | 5,885 | closed | feat: allow prefix for any generative model | Allow `padding_text` for any generative model.
### Motivations
* some models require some text at the beginning such as GPT-2 which should always have a white space. GPT-2 model could now have `config.padding_text` set to `" "` to ensure this is done
* this argument adds more customization to huggingface inference widget (related issue #5553). For huggingtweets models I can set `model.config.task_specific_params['text-generation']['padding_text'] = '<|endoftext|> '`
See [example](https://gist.github.com/borisdayma/89aaf1587390340add44c3c7081bffcd) of use of `padding_text` to create negative/positive sentences.
*Side note*: this is not completely obvious that a leading space should always be added for GPT-2 as `" \n This"` is not tokenized the same as `" \nThis"` which GPT-2 may have seen more in its training data (unless white spaces are added after new lines as well)
### Future improvements
In the future, we should also add `prefix` and `suffix`. This could let us do things similar to GPT-3 with the inference widget and specific tasks:
* `padding_text` would be used to initialize tasks (ex: 10 examples conversation with a bot). This is always cropped out and never displayed in the output
* `prefix` would add required text (ex: `" Human: "`). This is displayed in output.
* `suffix` would add required text after prompt (ex: `" \n AI: "`). This is displayed in output. | 07-18-2020 19:46:28 | 07-18-2020 19:46:28 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=h1) Report
> Merging [#5885](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/42fddacd1cac3cc57c3326aa51a409f5090b1261?el=desc) will **decrease** coverage by `0.88%`.
> The diff coverage is `75.00%`.
[](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5885 +/- ##
==========================================
- Coverage 78.47% 77.58% -0.89%
==========================================
Files 157 157
Lines 28569 28571 +2
==========================================
- Hits 22420 22168 -252
- Misses 6149 6403 +254
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/pipelines.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9waXBlbGluZXMucHk=) | `80.50% <75.00%> (+0.55%)` | :arrow_up: |
| [src/transformers/modeling\_reformer.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19yZWZvcm1lci5weQ==) | `16.87% <0.00%> (-79.30%)` | :arrow_down: |
| [src/transformers/configuration\_reformer.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3JlZm9ybWVyLnB5) | `21.62% <0.00%> (-78.38%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `22.58% <0.00%> (-72.26%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_flaubert.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9mbGF1YmVydC5weQ==) | `24.53% <0.00%> (-63.81%)` | :arrow_down: |
| [src/transformers/modeling\_marian.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19tYXJpYW4ucHk=) | `60.00% <0.00%> (-30.00%)` | :arrow_down: |
| [src/transformers/activations.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9hY3RpdmF0aW9ucy5weQ==) | `85.00% <0.00%> (-5.00%)` | :arrow_down: |
| [src/transformers/tokenization\_dpr.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fZHByLnB5) | `53.15% <0.00%> (-4.51%)` | :arrow_down: |
| [src/transformers/modeling\_bart.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19iYXJ0LnB5) | `95.05% <0.00%> (-0.35%)` | :arrow_down: |
| [src/transformers/tokenization\_utils\_base.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHNfYmFzZS5weQ==) | `93.49% <0.00%> (-0.28%)` | :arrow_down: |
| ... and [13 more](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree-more) | |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=footer). Last update [42fddac...00efd7d](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>Another solution would be to just allow `prefix` element like in `TranslationPipeline`. The main difference will be that it becomes part of the display and it should not affect too much future plans such as suggested above with `padding_text`, `prefix` and `suffix`.
I still think that `padding_text` as proposed here will be useful in the future to easily provide "GPT-3 type" widgets.<|||||>I updated my [example](https://gist.github.com/borisdayma/89aaf1587390340add44c3c7081bffcd) of the feature with the case when it loads model config.<|||||>Let me know if I can provide any additional information or make any changes<|||||>Just checking if you need anything else from me @julien-c<|||||>I’m off for a couple of days so I’ll let others chime in<|||||>Enjoy your time off!<|||||>@julien-c @mfuntowicz sorry I'm just pinging.
I have a [demo](https://huggingface.co/huggingtweets/julien_c) I'm excited about and this feature just makes predictions so much better.
*Note: not on master branch - dev version [here](https://colab.research.google.com/github/borisdayma/huggingtweets/blob/feat-artifacts/huggingtweets-demo.ipynb)*<|||||>Thoughts @patrickvonplaten @sshleifer? (text generation pipeline option)<|||||>This PR adds the padding functionality from `run_generation` to pipelines (and allow for custom text).
However the main idea would be to eventually allow "GPT-3 style" API's.
It could be renamed "prefix" but I suggest to make a distinction between padding and prefix where you would potentially want to display prefix & suffix in an interface (for example huggingface widget) but not necessarily the "padding".
This kind of API could be used for T5 and can let people create more modular inference examples.<|||||>@sshleifer I gave more thoughts to it and I agree we should just name it `prefix` and maybe add a "deprecated warning" on `run_generation`.
It should be sufficient for now and we will always be able to add a more complex behavior in the future if necessary.
Note: I don't understand the CI errors, quality is good on my side and the other tests errors seem unrelated to this PR<|||||>style: you need to rebase and then `pip install black --upgrade`
Other CI failures are spurious, you can ignore them.
My last request would be to add test coverage for the new argument in test_pipelines.py.
see:
```
test_torch_text_generation
test_tf_text_generation
```
<|||||>Changes:
* renamed `padding_text` to `prefix` in pipelines and `run_generation.py` (with warning of deprecated argument)
* since PR was made, we cannot tokenize empty text anymore so we now check for empty "prefix"
* since PR was made, we cannot pass extra args to `generate` so we explicitly add `prefix` to pipeline `__call__` method
* added tests for pipeline using prefix
Notes:
* we set default `prefix` to `None` (vs `""`) as we could pass `""` to overwrite model default value.
* sorry I should have done a merge instead of rebase as it changes the history. New commits start at "feat: rename padding_text to prefix"<|||||>Thanks @LysandreJik. Yes I noticed later that my results were better without the initial space as well.
Intuitively it seemed to me like a waste of encoding differentiating between having a space or not (for example even after new lines) as the start of a document can simply be identified by the special token.
I thought there were more examples without the space within the original training set which would help my fine-tuning, as long as I do the same both for training and inference.
In the end, I get my best results with just `<|endoftext|>` between each sample and stripping every extra space.<|||||>@LysandreJik let me know if anything else is required on my side.
Since it's a relatively large PR it can be hard to remain in sync with master.<|||||>I think we can merge this: @LysandreJik I leave it to you to merge :-) <|||||>+ @mfuntowicz can you update the transformers dependency in the inference API when this lands on master, so that @borisdayma can showcase his magic ;)<|||||>@mfuntowicz can you let me know when you update the inference API so I do some tests on the widget?<|||||>@borisdayma Everything has been deployed on the inference API, servers are up.
Thanks for taking care of this 💪 🙏 <|||||>That was fast! Thanks it looks to be working great! |
transformers | 5,884 | closed | Add ComVE model cards | Add `aliosm/ComVE*` model cards. These models were part of this work: https://sentic.net/transformer-models-for-commonsense-reasoning.pdf. | 07-18-2020 17:53:14 | 07-18-2020 17:53:14 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=h1) Report
> Merging [#5884](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/4b506a37e3e0ff679235961ba14dd9397843ef3a&el=desc) will **increase** coverage by `1.39%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5884 +/- ##
==========================================
+ Coverage 77.27% 78.67% +1.39%
==========================================
Files 146 146
Lines 26210 26210
==========================================
+ Hits 20254 20620 +366
+ Misses 5956 5590 -366
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5884/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `43.98% <0.00%> (-49.38%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5884/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |
| [src/transformers/file\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5884/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.49% <0.00%> (+0.29%)` | :arrow_up: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5884/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (+1.25%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5884/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `95.53% <0.00%> (+69.51%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5884/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=footer). Last update [4b506a3...d67656b](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>This is really cool, thanks for sharing |
transformers | 5,883 | closed | Xlnet outputs | Reopening #5770 again | 07-18-2020 15:16:09 | 07-18-2020 15:16:09 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=h1) Report
> Merging [#5883](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/3653d01f2af0389207f2239875a8ceae41bf0598&el=desc) will **increase** coverage by `1.36%`.
> The diff coverage is `100.00%`.
[](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5883 +/- ##
==========================================
+ Coverage 77.26% 78.62% +1.36%
==========================================
Files 146 146
Lines 25948 25958 +10
==========================================
+ Hits 20048 20409 +361
+ Misses 5900 5549 -351
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/configuration\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3V0aWxzLnB5) | `96.52% <ø> (ø)` | |
| [src/transformers/configuration\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3hsbmV0LnB5) | `94.33% <ø> (ø)` | |
| [src/transformers/modeling\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ194bG5ldC5weQ==) | `81.75% <100.00%> (+0.26%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=footer). Last update [3653d01...0c73c6d](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>Did @LysandreJik @patrickvonplaten or @sgugger get a chance to review this before merging, @TevenLeScao?<|||||>@patrickvonplaten and @joeddav reviewed it (@LysandreJik was off), there's quite a bit of discussion in #5770 if you wanna take a look! |
transformers | 5,882 | closed | Revert "Xlnet outputs" | This seems to break a test that isn't a CI. Think I've fixed it, will revert and re-open. | 07-18-2020 15:11:08 | 07-18-2020 15:11:08 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=h1) Report
> Merging [#5882](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/13be4872123094c37eb5fab939b38967b0ad2cd0&el=desc) will **increase** coverage by `0.21%`.
> The diff coverage is `100.00%`.
[](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5882 +/- ##
==========================================
+ Coverage 78.27% 78.48% +0.21%
==========================================
Files 146 146
Lines 26210 26200 -10
==========================================
+ Hits 20515 20563 +48
+ Misses 5695 5637 -58
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/configuration\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3V0aWxzLnB5) | `96.52% <ø> (ø)` | |
| [src/transformers/configuration\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3hsbmV0LnB5) | `94.33% <ø> (ø)` | |
| [src/transformers/modeling\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ194bG5ldC5weQ==) | `81.49% <100.00%> (-0.27%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `26.02% <0.00%> (-69.52%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `43.98% <0.00%> (-49.38%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (+0.25%)` | :arrow_up: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `82.31% <0.00%> (+1.28%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `95.18% <0.00%> (+74.91%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=footer). Last update [13be487...1be083d](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
transformers | 5,881 | closed | Xlnet outputs | Reopening #5770 since push failed. | 07-18-2020 14:41:12 | 07-18-2020 14:41:12 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=h1) Report
> Merging [#5881](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/3653d01f2af0389207f2239875a8ceae41bf0598&el=desc) will **increase** coverage by `0.54%`.
> The diff coverage is `100.00%`.
[](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5881 +/- ##
==========================================
+ Coverage 77.26% 77.81% +0.54%
==========================================
Files 146 146
Lines 25948 25958 +10
==========================================
+ Hits 20048 20198 +150
+ Misses 5900 5760 -140
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/configuration\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5881/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3V0aWxzLnB5) | `96.52% <ø> (ø)` | |
| [src/transformers/configuration\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5881/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3hsbmV0LnB5) | `94.33% <ø> (ø)` | |
| [src/transformers/modeling\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5881/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ194bG5ldC5weQ==) | `81.75% <100.00%> (+0.26%)` | :arrow_up: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5881/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5881/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=footer). Last update [3653d01...cfd7c66](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
transformers | 5,880 | closed | How to get probabilities from MarianMT models? | Hey,
I try to use MarianMT models for distillation, but I didn't found way to get raw probabilities from model.
I use model as in examples:
```
from transformers import MarianMTModel, MarianTokenizer
class Translator:
def __init__(self, src, dst):
model_name = f'Helsinki-NLP/opus-mt-{src}-{dst}'
self.tokenizer = MarianTokenizer.from_pretrained(model_name)
self.model = MarianMTModel.from_pretrained(model_name)
def __call__(self, text):
if not isinstance(text, list):
text = [text]
preds = self.model.generate(**self.tokenizer.prepare_translation_batch(text))
print(preds)
translation = [self.tokenizer.decode(t, skip_special_tokens=True) for t in preds]
return translation
```
How can I get translation and probs via Huggingface API?
Thanks in advance! | 07-18-2020 13:23:30 | 07-18-2020 13:23:30 | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,879 | closed | Create README.md | 07-18-2020 13:03:38 | 07-18-2020 13:03:38 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=h1) Report
> Merging [#5879](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `1.11%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5879 +/- ##
==========================================
+ Coverage 77.54% 78.66% +1.11%
==========================================
Files 146 146
Lines 26200 26200
==========================================
+ Hits 20318 20609 +291
+ Misses 5882 5591 -291
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5879/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5879/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5879/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=footer). Last update [eae6d8d...0c9ff72](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,878 | closed | Create README.md | 07-18-2020 13:02:29 | 07-18-2020 13:02:29 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=h1) Report
> Merging [#5878](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `1.11%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5878 +/- ##
==========================================
+ Coverage 77.54% 78.66% +1.11%
==========================================
Files 146 146
Lines 26200 26200
==========================================
+ Hits 20318 20610 +292
+ Misses 5882 5590 -292
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5878/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5878/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.46% <0.00%> (ø)` | |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5878/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=footer). Last update [eae6d8d...e9b1654](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,877 | closed | Create README.md | 07-18-2020 13:01:07 | 07-18-2020 13:01:07 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=h1) Report
> Merging [#5877](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `0.78%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5877 +/- ##
==========================================
+ Coverage 77.54% 78.33% +0.78%
==========================================
Files 146 146
Lines 26200 26200
==========================================
+ Hits 20318 20524 +206
+ Misses 5882 5676 -206
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5877/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5877/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5877/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.21% <0.00%> (-1.26%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5877/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=footer). Last update [eae6d8d...c7a4977](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,876 | closed | Create README.md | 07-18-2020 12:59:19 | 07-18-2020 12:59:19 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=h1) Report
> Merging [#5876](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `0.93%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5876 +/- ##
==========================================
+ Coverage 77.54% 78.48% +0.93%
==========================================
Files 146 146
Lines 26200 26200
==========================================
+ Hits 20318 20564 +246
+ Misses 5882 5636 -246
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5876/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `26.02% <0.00%> (-69.52%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5876/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |
| [src/transformers/file\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5876/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.49% <0.00%> (+0.29%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5876/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=footer). Last update [eae6d8d...befb998](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,875 | closed | Create README.md | 07-18-2020 12:57:57 | 07-18-2020 12:57:57 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=h1) Report
> Merging [#5875](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `0.81%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5875 +/- ##
==========================================
+ Coverage 77.54% 78.36% +0.81%
==========================================
Files 146 146
Lines 26200 26200
==========================================
+ Hits 20318 20531 +213
+ Misses 5882 5669 -213
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_gpt2.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9ncHQyLnB5) | `65.42% <0.00%> (-29.91%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `83.98% <0.00%> (-4.91%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_pytorch\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9weXRvcmNoX3V0aWxzLnB5) | `88.05% <0.00%> (-1.26%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.46% <0.00%> (-1.01%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=footer). Last update [eae6d8d...103b7af](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,874 | closed | Create README.md | 07-18-2020 12:55:40 | 07-18-2020 12:55:40 | ||
transformers | 5,873 | closed | Create README.md | 07-18-2020 12:53:49 | 07-18-2020 12:53:49 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=h1) Report
> Merging [#5873](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `0.79%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5873 +/- ##
==========================================
+ Coverage 77.54% 78.34% +0.79%
==========================================
Files 146 146
Lines 26200 26200
==========================================
+ Hits 20318 20527 +209
+ Misses 5882 5673 -209
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5873/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5873/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5873/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.96% <0.00%> (-0.51%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5873/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=footer). Last update [eae6d8d...f782f75](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,872 | closed | Create README.md | 07-18-2020 12:51:52 | 07-18-2020 12:51:52 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=h1) Report
> Merging [#5872](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `1.11%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5872 +/- ##
==========================================
+ Coverage 77.54% 78.66% +1.11%
==========================================
Files 146 146
Lines 26200 26200
==========================================
+ Hits 20318 20609 +291
+ Misses 5882 5591 -291
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5872/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5872/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5872/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=footer). Last update [eae6d8d...857c1c2](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,871 | closed | Create README.md | 07-18-2020 12:49:47 | 07-18-2020 12:49:47 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=h1) Report
> Merging [#5871](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `1.11%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5871 +/- ##
==========================================
+ Coverage 77.54% 78.66% +1.11%
==========================================
Files 146 146
Lines 26200 26200
==========================================
+ Hits 20318 20609 +291
+ Misses 5882 5591 -291
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5871/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5871/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5871/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=footer). Last update [eae6d8d...823159f](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,870 | closed | [cleanup] Less aggressive warnings about checkpoint mismatches | These are often spurious, so, in general, I made them shorter, and more about what happened than how to fix it.
Also, removed warnings for missing keys in seq2seq checkpoints. LM head is often made on the fly (to make the download cheaper).
Some examples:
- key 'encoder.version' in `bart.large` should not be trained more
- no keys in `T5ForConditionalGeneration` should be trained more.
| 07-18-2020 12:16:08 | 07-18-2020 12:16:08 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=h1) Report
> Merging [#5870](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/dad5e12e54bc2cf80a24b3430b5c847fc213a73e&el=desc) will **decrease** coverage by `0.49%`.
> The diff coverage is `85.71%`.
[](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5870 +/- ##
==========================================
- Coverage 78.48% 77.98% -0.50%
==========================================
Files 146 146
Lines 26200 26203 +3
==========================================
- Hits 20563 20435 -128
- Misses 5637 5768 +131
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5870/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `87.02% <85.71%> (-0.13%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5870/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |
| [src/transformers/tokenization\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5870/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25feGxuZXQucHk=) | `66.66% <0.00%> (-23.43%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5870/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5870/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.96% <0.00%> (-0.26%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5870/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=footer). Last update [dad5e12...02f32f1](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
transformers | 5,869 | closed | Silenced error while downloading pretrained model | # 🐛 Bug
Currently, if user failed to download a model, they will receive uninformative and confusing message (see #5787). This is most hurting for users behind a proxy. This happens because of this [line](https://github.com/huggingface/transformers/blob/0533cf470659b97c6279bd04f65536a1ec88404a/src/transformers/file_utils.py#L681). My opinion is that there should be at least a warning, and if `force_download` is `True`, then error should be raised. I can make a PR if you agree with that.
| 07-18-2020 11:57:30 | 07-18-2020 11:57:30 | This sounds reasonable<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,868 | closed | [cleanup] squad processor | 07-18-2020 11:48:00 | 07-18-2020 11:48:00 | ||
transformers | 5,867 | closed | Update tokenizers to 0.8.1.rc to fix Mac OS X issues | As mentioned in huggingface/tokenizers#321 tokenizers and thus transformers fails on macOS 10.11+. The issue was fixed in `0.8.1.rc2` so this PR updates the dependency to the new version.
I ran all tests on my mac with Mac OS High Sierra 10.13.6 as desribed in the [readme](https://github.com/huggingface/transformers/blob/master/README.md#tests), but not the examples:
```
1607 passed, 304 skipped, 15294 warnings in 465.74s (0:07:45)
```
| 07-18-2020 11:43:24 | 07-18-2020 11:43:24 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=h1) Report
> Merging [#5867](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/ba2400189b2242620868096ae49babf93bd9ce00&el=desc) will **decrease** coverage by `0.39%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5867 +/- ##
==========================================
- Coverage 78.48% 78.09% -0.40%
==========================================
Files 146 146
Lines 26200 26200
==========================================
- Hits 20564 20461 -103
- Misses 5636 5739 +103
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5867/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5867/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5867/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.96% <0.00%> (-0.51%)` | :arrow_down: |
| [src/transformers/generation\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5867/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3V0aWxzLnB5) | `97.11% <0.00%> (+0.28%)` | :arrow_up: |
| [src/transformers/file\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5867/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.49% <0.00%> (+0.29%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5867/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=footer). Last update [ba24001...4d35935](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>Awesome! Thank you @sepal <|||||>Could someone provide more detailed instructions to update the tokenizers in the version that solves the issue? I have a mac (10.13) and encountered the same problem. I tried through terminal to pip install tokenizers 0.8.1.rs2 but ended up with this error.
ERROR: Failed building wheel for tokenizers
Failed to build tokenizers
ERROR: Could not build wheels for tokenizers which use PEP 517 and cannot be installed directly |
transformers | 5,866 | closed | T5Tokenizer adds EOS token if not already added | T5 Tokenizer should add `</s>` to the end of sequences. Since some users are doing this on their own, this PR only adds `</s>` if it has already been added.
On my machine, this makes zero shot validation BLEU go from 27.87 -> 27.65. Since this change is needed for finetuning, and the empirical difference is small and doesn't happen on Stas' machine, I would recommend merging this.
If others want to test, the command takes about 3 mins to run on brutasse.
### Zero Shot BLEU Scores
For english -> romanian
I grabbed the WMT english-romanian dataset:
```bash
wget https://s3.amazonaws.com/datasets.huggingface.co/translation/wmt_en_ro.tar.gz
tar -xzvf wmt_en_ro.tar.gz
```
Then ran evaluation (without finetuning) on the validation split:
```bash
export DATA_DIR=wmt_en_ro
python run_eval.py t5-base \
$DATA_DIR/val.source t5_val_generations.txt \
--reference_path $DATA_DIR/val.target \
--score_path t5_enro_bleu_eos.json \
--task translation_en_to_ro \
--device cuda \
--fp16 \
--bs 32
```
(this branch) (with EOS):27.65
master (no EOS): 27.87
```
sacrebleu==1.4.3
torch==1.5.1
```
Will merge and fix tests if others have positive results. | 07-18-2020 11:29:53 | 07-18-2020 11:29:53 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=h1) Report
> Merging [#5866](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/7e6397a7d8e7433aa4c4cafba98e08e5c73f087c?el=desc) will **decrease** coverage by `0.67%`.
> The diff coverage is `89.47%`.
[](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5866 +/- ##
==========================================
- Coverage 80.10% 79.42% -0.68%
==========================================
Files 156 156
Lines 28411 28426 +15
==========================================
- Hits 22758 22578 -180
- Misses 5653 5848 +195
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/tokenization\_t5.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdDUucHk=) | `95.32% <89.47%> (-1.42%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `22.58% <0.00%> (-72.26%)` | :arrow_down: |
| [src/transformers/tokenization\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fcm9iZXJ0YS5weQ==) | `87.67% <0.00%> (-10.96%)` | :arrow_down: |
| [src/transformers/tokenization\_utils\_base.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHNfYmFzZS5weQ==) | `86.58% <0.00%> (-7.19%)` | :arrow_down: |
| [src/transformers/tokenization\_transfo\_xl.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdHJhbnNmb194bC5weQ==) | `38.73% <0.00%> (-3.76%)` | :arrow_down: |
| [src/transformers/tokenization\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fb3BlbmFpLnB5) | `82.57% <0.00%> (-1.52%)` | :arrow_down: |
| [src/transformers/tokenization\_utils\_fast.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHNfZmFzdC5weQ==) | `92.85% <0.00%> (-1.43%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `80.96% <0.00%> (-1.30%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.51%)` | :arrow_down: |
| [src/transformers/tokenization\_bert.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fYmVydC5weQ==) | `91.07% <0.00%> (-0.45%)` | :arrow_down: |
| ... and [2 more](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree-more) | |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=footer). Last update [7e6397a...d977bff](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>FWIW, I get identical results of `27.84` with this branch and the master.<|||||>Happy to eventually remove the check to see if it's already there.<|||||>I think we can keep it like this right now with a warning for future versions. It would create a breaking change to users, and I feel it would be especially hard to debug an unknown drop in performance due to an additional token being added, right?<|||||>Will this behavior cause problems for the unsupervised setting? Per the [docs](https://huggingface.co/transformers/model_doc/t5.html#training), `</s>` is not added during denoising training:
```
input_ids = tokenizer.encode('The <extra_id_0> walks in <extra_id_1> park', return_tensors='pt')
labels = tokenizer.encode('<extra_id_0> cute dog <extra_id_1> the <extra_id_2> </s>', return_tensors='pt')
model(input_ids=input_ids, labels=labels)
```
Not sure if this will cause problems. (Also, Aa a somewhat related question, should the sentinel tokens in the `labels` be excluded from the loss in this setting, as I believe is the case with `[MASK]` in BERT?).<|||||>I'm not sure about either question:
made an issue verifying the docs: https://github.com/huggingface/transformers/issues/7904
Feel free to make an issue about the sentinel tokens question. I'd tag thomwolf/patrickvonplaten.
|
transformers | 5,865 | closed | [seq2seq] distillation.py accepts trainer arguments | cc @nateraw | 07-18-2020 11:21:35 | 07-18-2020 11:21:35 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5865?src=pr&el=h1) Report
> Merging [#5865](https://codecov.io/gh/huggingface/transformers/pull/5865?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/529850ae7bca0ff388778c3c0d66240834cf56c3&el=desc) will **not change** coverage.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5865?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5865 +/- ##
=======================================
Coverage 78.48% 78.48%
=======================================
Files 146 146
Lines 26200 26200
=======================================
Hits 20563 20563
Misses 5637 5637
```
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5865?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5865?src=pr&el=footer). Last update [529850a...259bc29](https://codecov.io/gh/huggingface/transformers/pull/5865?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
transformers | 5,864 | closed | Create README.md | 07-18-2020 11:02:11 | 07-18-2020 11:02:11 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=h1) Report
> Merging [#5864](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/ba2400189b2242620868096ae49babf93bd9ce00&el=desc) will **decrease** coverage by `0.39%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5864 +/- ##
==========================================
- Coverage 78.48% 78.09% -0.40%
==========================================
Files 146 146
Lines 26200 26200
==========================================
- Hits 20564 20460 -104
- Misses 5636 5740 +104
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5864/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5864/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5864/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.96% <0.00%> (-0.51%)` | :arrow_down: |
| [src/transformers/file\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5864/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.49% <0.00%> (+0.29%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5864/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=footer). Last update [ba24001...84114cc](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,863 | closed | not able to reproduce accuracy at the end of same epoch | # 🐛 Bug
update: Found what is causing this bug
Your optimization [code](https://github.com/huggingface/transformers/blob/master/src/transformers/optimization.py#L70) uses a linear scheduler for learning rate with warmup which is **dependent on the total number of training steps** (== no of batches*total epochs). This injects randomness (as the total epoch count changes) during every scheduler.step(), which affects the weights of the trained model, which in turn reflects in accuracy....
The effect of this is very minimal in the in-domain setting. However, since I was evaluating on a cross-domain-setting it ended up being heavily amplified. Can you please make the warm up dependent on the current epoch and not on total epochs/step size.
----original post
When am running say for 25 epochs, I wanted to run an eval on dev and test partition at the end of every epoch. So I moved the evaluate() code in trainery.py to the end of all batches. However the accuracy I get at the end of say epoch 1 (in the big 25 epoch run) is different than what I get when I run the entire code for just 1 epoch. This might be the same answer to [this](https://github.com/huggingface/transformers/issues/5264) issue also
## Information
Model I am using (Bert, XLNet ...):Bert (base-uncased)
Language I am using the model on (English, Chinese ...):English
The problem arises when using:
* [ ] the official example scripts: (give details below)
* [x ] my own modified scripts: (give details below): the only modification is I call the same evaluate function to after epochs. You can see that [here](https://github.com/mithunpaul08/transformers/blob/master/src/transformers/trainer.py#L1592).
The tasks I am working on is:
* [ ] an official GLUE/SQUaD task: (give the name)
* [ x] my own task or dataset: (give details below)
I am training and testing on [FEVER](https://s3-eu-west-1.amazonaws.com/fever.public/wiki-pages.zip) . My code should download it automatically though.
## To reproduce
Steps to reproduce the behavior:
- Kindly please clone my [repo](https://github.com/mithunpaul08/transformers)
- conda create --name huggingface python=3
- source activate huggingface
- pip install -r ./examples/requirements.txt
from home folder (i.e /huggingface/) do:
- cd mithun_scripts/
- ./run_all.sh
(should run i think)
<!-- If you have code snippets, error messages, stack traces please provide them here as well.
Important! Use code tags to correctly format your code. See https://help.github.com/en/github/writing-on-github/creating-and-highlighting-code-blocks#syntax-highlighting
Do not use screenshots, as they are hard to read and (more importantly) don't allow others to copy-and-paste your code.-->
## Expected behavior
<!-- A clear and concise description of what you would expect to happen. -->
## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version:
- Platform:
- Python version:
- PyTorch version (GPU?):
- Tensorflow version (GPU?):
- Using GPU in script?:
- Using distributed or parallel set-up in script?:
| 07-18-2020 08:33:02 | 07-18-2020 08:33:02 | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,862 | closed | Potential security vulnerability regarding Hosted Interface API? | # 🐛 Bug
## Information
Model I am using (Bert, XLNet ...): distilbert-base-cased-distilled-squad with huggingface.co's API
Language I am using the model on (English, Chinese ...): English
The problem arises when using:
* [ ] the official example scripts: (give details below)
* [X] my own modified scripts: (give details below)
The tasks I am working on is:
* [ ] an official GLUE/SQUaD task: (give the name)
* [X] my own task or dataset: Question and answering
## To reproduce
Steps to reproduce the behavior:
Run this code on a notebook
```
answer = requests.post('https://api-inference.huggingface.co/models/distilbert-base-cased-distilled-squad',
headers = {"Content-Type": "application/json"},
data = {"question":'What is my name?' ,
"context" :"My name is Batik"})
```
Result : 500 Error . With the following content when I executed `print(answer.content)`
```
b'{"error":"JSONDecodeError(\'Expecting value: line 1 column 1 (char 0)\')","traceback":" File \\"/home/hf/api-inference/server.py\\", line 251, in model_forward\\n inputs = await request.json()\\n File \\"/home/hf/api-inference/.env/lib/python3.8/site-packages/starlette/requests.py\\", line 227, in json\\n self._json = json.loads(body)\\n File \\"/usr/lib/python3.8/json/__init__.py\\", line 357, in loads\\n return _default_decoder.decode(s)\\n File \\"/usr/lib/python3.8/json/decoder.py\\", line 337, in decode\\n obj, end = self.raw_decode(s, idx=_w(s, 0).end())\\n File \\"/usr/lib/python3.8/json/decoder.py\\", line 355, in raw_decode\\n raise JSONDecodeError(\\"Expecting value\\", s, err.value) from None\\n"}'
```
## Expected behavior
I imagine we don't want to give the Error traceback for security reasons
## Environment info
I tried using Python's request and Postman and have the same outcome. Furthermore, I'm wondering why I have a Error 500. Was my code wrong?
| 07-18-2020 05:51:36 | 07-18-2020 05:51:36 | Your `requests` syntax is incorrect, you're not actually sending the request body as json.
This works:
```python
answer = requests.post(
"https://api-inference.huggingface.co/models/distilbert-base-cased-distilled-squad",
json={"question": "What is my name?", "context": "My name is Batik"},
)
```
Regarding the stack trace, I do not see an issue with returning it (it's json-encoded), happy to discuss more. |
transformers | 5,861 | closed | [seq2seq] add back clargs | 07-18-2020 04:05:36 | 07-18-2020 04:05:36 | cc @nateraw <|||||># [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=h1) Report
> Merging [#5861](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/529850ae7bca0ff388778c3c0d66240834cf56c3&el=desc) will **decrease** coverage by `0.25%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5861 +/- ##
==========================================
- Coverage 78.48% 78.23% -0.26%
==========================================
Files 146 146
Lines 26200 26200
==========================================
- Hits 20563 20497 -66
- Misses 5637 5703 +66
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `83.70% <0.00%> (-2.51%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |
| [src/transformers/generation\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3V0aWxzLnB5) | `97.11% <0.00%> (+0.28%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `95.53% <0.00%> (+69.51%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=footer). Last update [529850a...c7b897e](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>better way in #5865 |
|
transformers | 5,860 | closed | issue with loading pretrained model - xlnet | # ❓ Questions & Help
<!-- The GitHub issue tracker is primarly intended for bugs, feature requests,
new models and benchmarks, and migration questions. For all other questions,
we direct you to the Hugging Face forum: https://discuss.huggingface.co/ .
You can also try Stack Overflow (SO) where a whole community of PyTorch and
Tensorflow enthusiast can help you out. In this case, make sure to tag your
question with the right deep learning framework as well as the
huggingface-transformers tag:
https://stackoverflow.com/questions/tagged/huggingface-transformers
-->
I ran the following code:
from transformers import XLNetTokenizer, XLNetForSequenceClassification
import torch
tokenizer = XLNetTokenizer.from_pretrained('xlnet-base-cased')
model = XLNetForSequenceClassification.from_pretrained('xlnet-base-cased')
## Details
<!-- Description of your issue -->
and I have the following error:
> OSError: Can't load config for 'xlnet-base-cased'. Make sure that:
> 'xlnet-base-cased' is a correct model identifier listed on 'https://huggingface.co/models'
> or 'xlnet-base-cased' is the correct path to a directory containing a config.json file
<!-- You should first ask your question on the forum or SO, and only if
you didn't get an answer ask it here on GitHub. -->
The error didn't appear when I ran the same piece of code two days ago so I think it might be caused by recent update. | 07-18-2020 03:59:12 | 07-18-2020 03:59:12 | |
transformers | 5,859 | closed | [seq2seq] organize commands into scripts/ subdir | whoever takes this, please test them after! | 07-18-2020 02:48:39 | 07-18-2020 02:48:39 | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,858 | closed | wrong args name: n_gpu -> gpus | The correct argument name should be `gpus`.
The change fixed the following error:
```bash
warnings.warn(*args, **kwargs)
Traceback (most recent call last):
File "/home/.conda/envs/hf/lib/python3.6/site-packages/pytorch_lightning/utilities/parsing.py", line 114, in __getattr__
return self[key]
KeyError: 'n_gpu'
``` | 07-18-2020 02:32:38 | 07-18-2020 02:32:38 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=h1) Report
> Merging [#5858](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/615be03f9d961c0c9722fe10e7830e011066772e&el=desc) will **decrease** coverage by `0.20%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5858 +/- ##
==========================================
- Coverage 78.66% 78.46% -0.21%
==========================================
Files 146 146
Lines 26200 26200
==========================================
- Hits 20611 20558 -53
- Misses 5589 5642 +53
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5858/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `26.02% <0.00%> (-69.52%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5858/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `84.96% <0.00%> (-1.76%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5858/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=footer). Last update [615be03...713f748](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>great catch! was fixed on master by #5798 so I don't think we need this. Let us know if further issues! |
transformers | 5,857 | closed | Update README.md | Add nlp dataset used | 07-17-2020 22:15:12 | 07-17-2020 22:15:12 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=h1) Report
> Merging [#5857](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/615be03f9d961c0c9722fe10e7830e011066772e&el=desc) will **decrease** coverage by `0.58%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5857 +/- ##
==========================================
- Coverage 78.66% 78.08% -0.59%
==========================================
Files 146 146
Lines 26200 26200
==========================================
- Hits 20611 20459 -152
- Misses 5589 5741 +152
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5857/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5857/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `26.02% <0.00%> (-69.52%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5857/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5857/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.96% <0.00%> (-0.76%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5857/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5857/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=footer). Last update [615be03...e7f330a](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
transformers | 5,856 | closed | Update README.md | Add dataset used as it is now part of nlp package | 07-17-2020 22:13:46 | 07-17-2020 22:13:46 | 👍 awesome
<|||||>Even if dataset is not in `nlp` yet, you can still link to it, as it will display a missing dataset page with a call to action to add it to nlp: e.g. https://huggingface.co/datasets/dlkfjldsfjdslf |
transformers | 5,855 | closed | docs: fix model sharing file names | Documentation references `vocab.txt` while I think it should be `vocab.json`.
Additional notes:
* I also have `training_args.bin` and `merges.txt` which I understand are not needed to be uploaded
* I didn't get `added_tokens.json` but I understand it's not always present | 07-17-2020 21:34:18 | 07-17-2020 21:34:18 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=h1) Report
> Merging [#5855](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/8279471506fab5733dab3e2d3a1542010c976d8a?el=desc) will **decrease** coverage by `2.76%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5855 +/- ##
==========================================
- Coverage 79.87% 77.11% -2.77%
==========================================
Files 181 181
Lines 35788 35788
==========================================
- Hits 28587 27597 -990
- Misses 7201 8191 +990
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/configuration\_reformer.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3JlZm9ybWVyLnB5) | `21.62% <0.00%> (-78.38%)` | :arrow_down: |
| [src/transformers/modeling\_reformer.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19yZWZvcm1lci5weQ==) | `16.71% <0.00%> (-77.89%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_funnel.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9mdW5uZWwucHk=) | `18.53% <0.00%> (-75.51%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `24.25% <0.00%> (-73.56%)` | :arrow_down: |
| [src/transformers/activations\_tf.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9hY3RpdmF0aW9uc190Zi5weQ==) | `54.16% <0.00%> (-20.84%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |
| [src/transformers/file\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `83.13% <0.00%> (-0.25%)` | :arrow_down: |
| [src/transformers/modeling\_bert.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19iZXJ0LnB5) | `88.48% <0.00%> (+0.16%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `86.90% <0.00%> (+0.63%)` | :arrow_up: |
| [src/transformers/modeling\_auto.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19hdXRvLnB5) | `87.04% <0.00%> (+1.03%)` | :arrow_up: |
| ... and [5 more](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree-more) | |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=footer). Last update [8279471...d60525f](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>It's more general than vocab.txt or vocab.json so I'll let @sgugger chime in as I think he wrote those lines.<|||||>I didn't know that when I wrote it (and the file was vocab.json for me ;-) ). This looks good to me with the suggestion, thanks for clarifying!<|||||>Looks good on my side, the only file removed from my example would be `training_args.bin` which makes sense.<|||||>Let me know if I need to do any other changes<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
<|||||>Oops, sorry this got lost in my notifications. Don't hesitate to ping me if it happens again @borisdayma, will merge as soon as the CI is green. |
transformers | 5,854 | closed | Revert "XLNet `use_cache` refactor" | Reverts huggingface/transformers#5770 | 07-17-2020 18:33:37 | 07-17-2020 18:33:37 | |
transformers | 5,853 | closed | [BartModel] Question for BartModel Output Shape when I pass the 'decoder_input_ids' | # ❓ Questions & Help
<!-- The GitHub issue tracker is primarly intended for bugs, feature requests,
new models and benchmarks, and migration questions. For all other questions,
we direct you to the Hugging Face forum: https://discuss.huggingface.co/ .
You can also try Stack Overflow (SO) where a whole community of PyTorch and
Tensorflow enthusiast can help you out. In this case, make sure to tag your
question with the right deep learning framework as well as the
huggingface-transformers tag:
https://stackoverflow.com/questions/tagged/huggingface-transformers
-->
## Details
<!-- Description of your issue -->
<!-- You should first ask your question on the forum or SO, and only if
you didn't get an answer ask it here on GitHub. -->
Hi I'm trying to using model named 'BartForConditionalGeneration' for text generation,
The main question and problem is, when I pass the 'decoder_input_ids' to the model as the input of forward in model,
the first index of model's output shows the tensor shape as [batch_size, 1 , vocab_size] all the time.
I might misunderstood the model and Docs of Bart overall,
when I only pass the Input_ids without decoder_input_ids, it worked right.
but It never make the shape of tensor like [batch_size, target_seq_len, vocab_size] when I put 'decoder_input_ids'
I think when I put the decoder_input_ids as [batch_size, target_seq_len], the Model's output should be like
[batch_size, target_seq_len, vocab_size] (for prediction scores)
the reason why I tried to pass the decoder_input_ids was , I think Bart is seq2seq model, and their output logits will be computed just like the basic transformer, so I wanted to pretrain this model.
Please help me to solve this problem. If I misunderstood the concept of Bart at all, just comment me.
I need a clean view for my situation
I used mBartTokenizer and also 'facebook/mbart-large-en-ro' pretrained weights,
I used Model BartForConditionalGeneration.from_pretrained('facebook/mbart-large-en-ro')
It also didn't worked for normal BartModel (with same pretrained config)
I'm really confused about it so the whole question looks like a mass, Sorry for that
Thank you for reading my questions. Please Help Me.
| 07-17-2020 16:58:09 | 07-17-2020 16:58:09 | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,852 | closed | Exception in device=TPU:1: 'ascii' codec can't decode byte 0xc2 in position 37: ordinal not in range(128) | I get the error message when trying to run the seq2seq example on 8 TPU cores on XSUM and bart-large.
Run this from `examples/seq2seq`:
```
export PYTHONPATH="../":"${PYTHONPATH}"
python finetune.py \
--learning_rate=3e-5 \
--gpus 0 \
--n_tpu_cores 8 \
--do_train \
--do_predict \
--n_val 1000 \
--val_check_interval 1.0 \
--sortish_sampler \
--data_dir ${PWD}/xsum \
--train_batch_size=1 \
--eval_batch_size=1 \
--output_dir=xsum_results \
--num_train_epochs 1 \
--model_name_or_path facebook/bart-large
```
BTW there is also a bug in `lightning_base.py`. The `gpus` argument is given twice if `n_tpu_cores` is specified (second time through `**train_params`) so I replaced Trainer creation lines with these:
```
n_gpus = train_params.pop("gpus", args.gpus)
trainer = pl.Trainer(
logger=logger,
accumulate_grad_batches=args.gradient_accumulation_steps,
gpus=n_gpus,
max_epochs=args.num_train_epochs,
early_stop_callback=early_stopping_callback,
gradient_clip_val=args.max_grad_norm,
checkpoint_callback=checkpoint_callback,
callbacks=[logging_callback] + extra_callbacks,
fast_dev_run=args.fast_dev_run,
val_check_interval=args.val_check_interval,
weights_summary=None,
resume_from_checkpoint=args.resume_from_checkpoint,
**train_params,
)
``` | 07-17-2020 16:51:49 | 07-17-2020 16:51:49 | ...it also happens on CPU<|||||>I solved it. If you happen to use a Google Cloud VM with the `Debian GNU/Linux 9 Stretch + PyTorch/XLA` image then locale is not set correctly. Add the following to `/etc/default/locale`:
```
LC_ALL=en_US.UTF-8
```<|||||>...thank you, it works! |
transformers | 5,851 | closed | Covid-19 - TPU V3-1024 - T5 11B: Tensorflow to Pytorch conversion failed | We are training a large scale T5-11B model using TPU V3-1024 for a Covid-19 project
We tried to convert the TensorFlow checkpoint to the Pytorch version, but it did fail.
Could you please help us to figure out the problem since this model is very important for Covid-19 research.
# 🐛 Bug
## Information
Model I am using (Bert, XLNet ...):
T5
Language I am using the model on (English, Chinese ...):
Protein Sequences
The problem arises when using:
* [x] the official example scripts: (give details below)
* [ ] my own modified scripts: (give details below)
The tasks I am working on is:
* [ ] an official GLUE/SQUaD task: (give the name)
* [x] my own task or dataset: (give details below)
## To reproduce
Steps to reproduce the behavior:
1. The config file:
```
{
"architectures": [
"T5WithLMHeadModel"
],
"d_ff": 65536,
"d_kv": 128,
"d_model": 1024,
"decoder_start_token_id": 0,
"dropout_rate": 0.1,
"eos_token_id": 1,
"initializer_factor": 1.0,
"is_encoder_decoder": true,
"layer_norm_epsilon": 1e-06,
"model_type": "t5",
"n_positions": 512,
"num_heads": 128,
"num_layers": 24,
"output_past": true,
"pad_token_id": 0,
"relative_attention_num_buckets": 32,
"vocab_size": 128
}
```
2. conversion command:
```
python convert_t5_original_tf_checkpoint_to_pytorch.py \
--tf_checkpoint_path xxx/tensorflow \
--config_file xxx/t5-11b-config.json \
--pytorch_dump_path xxx/pytorch
```
3. Error:
```
Building PyTorch model from configuration: T5Config {
"architectures": [
"T5WithLMHeadModel"
],
"d_ff": 65536,
"d_kv": 128,
"d_model": 1024,
"decoder_start_token_id": 0,
"dropout_rate": 0.1,
"eos_token_id": 1,
"initializer_factor": 1.0,
"is_encoder_decoder": true,
"layer_norm_epsilon": 1e-06,
"model_type": "t5",
"n_positions": 512,
"num_heads": 128,
"num_layers": 24,
"output_past": true,
"pad_token_id": 0,
"relative_attention_num_buckets": 32,
"vocab_size": 128
}
INFO:transformers.modeling_t5:Converting TensorFlow checkpoint from /mnt/lsf-nas-1/lsf/job/repo/elnaggar/prot-transformers/models/t5/tensorflow/bfd100
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/relative_attention_bias with shape [128, 32]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/relative_attention_bias_slot_v with shape [128, 32]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/EncDecAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/EncDecAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/EncDecAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/EncDecAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/EncDecAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/EncDecAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/EncDecAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/EncDecAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/EncDecAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/EncDecAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/EncDecAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/EncDecAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_002/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_002/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_002/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_002/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_002/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_002/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_002/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/layer_002/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight decoder/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/relative_attention_bias with shape [128, 32]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/relative_attention_bias_slot_v with shape [128, 32]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_000/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_001/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_002/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_003/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_004/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_005/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_006/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_007/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_008/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_009/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_010/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_011/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_012/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_013/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_014/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_015/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_016/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_017/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_018/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_019/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_020/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_021/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_022/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/SelfAttention/k with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/SelfAttention/k_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/SelfAttention/k_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/SelfAttention/o with shape [16384, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/SelfAttention/o_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/SelfAttention/o_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/SelfAttention/q with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/SelfAttention/q_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/SelfAttention/q_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/SelfAttention/v with shape [1024, 16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/SelfAttention/v_slot_vc with shape [16384]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/SelfAttention/v_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_000/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_001/DenseReluDense/wi/kernel with shape [1024, 65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_001/DenseReluDense/wi/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_001/DenseReluDense/wi/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_001/DenseReluDense/wo/kernel with shape [65536, 1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_001/DenseReluDense/wo/kernel_slot_vc with shape [65536]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_001/DenseReluDense/wo/kernel_slot_vr with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_001/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/block_023/layer_001/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/rms_norm/scale with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight encoder/rms_norm/scale_slot_v with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight global_step with shape []
INFO:transformers.modeling_t5:Loading TF weight shared/embedding with shape [128, 1024]
INFO:transformers.modeling_t5:Loading TF weight shared/embedding_slot_vc with shape [1024]
INFO:transformers.modeling_t5:Loading TF weight shared/embedding_slot_vr with shape [128]
INFO:transformers.modeling_t5:Transposing numpy weight of shape (1024, 16384) for ['decoder', 'block_000', 'layer_000', 'SelfAttention', 'k']
INFO:transformers.modeling_t5:Initialize PyTorch weight ['decoder', 'block_000', 'layer_000', 'SelfAttention', 'k']
INFO:transformers.modeling_t5:Skipping decoder/block_000/layer_000/SelfAttention/k_slot_vc
INFO:transformers.modeling_t5:Skipping decoder/block_000/layer_000/SelfAttention/k_slot_vr
INFO:transformers.modeling_t5:Transposing numpy weight of shape (16384, 1024) for ['decoder', 'block_000', 'layer_000', 'SelfAttention', 'o']
INFO:transformers.modeling_t5:Initialize PyTorch weight ['decoder', 'block_000', 'layer_000', 'SelfAttention', 'o']
INFO:transformers.modeling_t5:Skipping decoder/block_000/layer_000/SelfAttention/o_slot_vc
INFO:transformers.modeling_t5:Skipping decoder/block_000/layer_000/SelfAttention/o_slot_vr
INFO:transformers.modeling_t5:Transposing numpy weight of shape (1024, 16384) for ['decoder', 'block_000', 'layer_000', 'SelfAttention', 'q']
INFO:transformers.modeling_t5:Initialize PyTorch weight ['decoder', 'block_000', 'layer_000', 'SelfAttention', 'q']
INFO:transformers.modeling_t5:Skipping decoder/block_000/layer_000/SelfAttention/q_slot_vc
INFO:transformers.modeling_t5:Skipping decoder/block_000/layer_000/SelfAttention/q_slot_vr
INFO:transformers.modeling_t5:Transposing numpy weight of shape (128, 32) for ['decoder', 'block_000', 'layer_000', 'SelfAttention', 'relative_attention_bias']
INFO:transformers.modeling_t5:Initialize PyTorch weight ['decoder', 'block_000', 'layer_000', 'SelfAttention', 'relative_attention_bias']
INFO:transformers.modeling_t5:Skipping decoder/block_000/layer_000/SelfAttention/relative_attention_bias_slot_v
INFO:transformers.modeling_t5:Transposing numpy weight of shape (1024, 16384) for ['decoder', 'block_000', 'layer_000', 'SelfAttention', 'v']
INFO:transformers.modeling_t5:Initialize PyTorch weight ['decoder', 'block_000', 'layer_000', 'SelfAttention', 'v']
INFO:transformers.modeling_t5:Skipping decoder/block_000/layer_000/SelfAttention/v_slot_vc
INFO:transformers.modeling_t5:Skipping decoder/block_000/layer_000/SelfAttention/v_slot_vr
INFO:transformers.modeling_t5:Skipping decoder/block_000/layer_000/rms_norm/scale
Traceback (most recent call last):
File "xxx/convert_t5_original_tf_checkpoint_to_pytorch.py", line 61, in <module>
convert_tf_checkpoint_to_pytorch(args.tf_checkpoint_path, args.config_file, args.pytorch_dump_path)
File "xxx/convert_t5_original_tf_checkpoint_to_pytorch.py", line 36, in convert_tf_checkpoint_to_$ytorch
load_tf_weights_in_t5(model, config, tf_checkpoint_path)
File "xxx/modeling_t5.py", line 102, in load_tf_weights_in_t5
pointer = getattr(pointer, "weight")
File "xxx/anaconda3/envs/transformers/lib/python3.7/site-packages/torch/nn/modules/module.py", line 594, in __getattr__
type(self).__name__, name))
AttributeError: 'T5LayerSelfAttention' object has no attribute 'weight'
```
## Expected behavior
T5 tensorflow model should be converted to pytorch model.
## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 2.11.0
- Platform: Linux-4.15.0-101-generic-x86_64-with-debian-buster-sid
- Python version: 3.7.7
- PyTorch version (GPU?): 1.5.0 (False)
- Tensorflow version (GPU?): 2.2.0 (False)
- Using GPU in script?: No
- Using distributed or parallel set-up in script?: No
| 07-17-2020 16:37:47 | 07-17-2020 16:37:47 | @thomwolf @LysandreJik @julien-c @patrickvonplaten @VictorSanh @sshleifer @mfuntowicz @sgugger your feedback will be highly appreciated.<|||||>I have checked the "operative_config.gin", and this is the encoder/decoder configuration 👍
```
# Parameters for decoder/LayerStack:
# ==============================================================================
decoder/LayerStack.dropout_rate = None
decoder/LayerStack.norm_epsilon = None
decoder/LayerStack.recompute_grads = False
decoder/LayerStack.sublayers_final = \
[@transformer.sublayer_rms_norm, @transformer.sublayer_dropout]
decoder/LayerStack.sublayers_initial = [@transformer.sublayer_dropout]
decoder/LayerStack.sublayers_per_layer = \
[@transformer.sublayer_rms_norm,
@transformer.sublayer_call_layer,
@transformer.sublayer_dropout,
@transformer.sublayer_residual]
# Parameters for encoder/LayerStack:
# ==============================================================================
encoder/LayerStack.dropout_rate = None
encoder/LayerStack.norm_epsilon = None
encoder/LayerStack.recompute_grads = False
encoder/LayerStack.sublayers_final = \
[@transformer.sublayer_rms_norm, @transformer.sublayer_dropout]
encoder/LayerStack.sublayers_initial = [@transformer.sublayer_dropout]
encoder/LayerStack.sublayers_per_layer = \
[@transformer.sublayer_rms_norm,
@transformer.sublayer_call_layer,
@transformer.sublayer_dropout,
@transformer.sublayer_residual]
# Parameters for make_bitransformer:
# ==============================================================================
make_bitransformer.decoder_name = 'decoder'
make_bitransformer.encoder_name = 'encoder'
# Parameters for decoder/make_layer_stack:
# ==============================================================================
decoder/make_layer_stack.block_scope = True
decoder/make_layer_stack.layers = \
[@mesh_tensorflow.transformer.transformer_layers.SelfAttention,
@mesh_tensorflow.transformer.transformer_layers.EncDecAttention,
@mesh_tensorflow.transformer.transformer_layers.DenseReluDense]
decoder/make_layer_stack.num_layers = %num_layers
# Parameters for encoder/make_layer_stack:
# ==============================================================================
encoder/make_layer_stack.block_scope = True
encoder/make_layer_stack.layers = \
[@mesh_tensorflow.transformer.transformer_layers.SelfAttention,
@mesh_tensorflow.transformer.transformer_layers.DenseReluDense]
encoder/make_layer_stack.num_layers = %num_layers
```<|||||>I also upgraded the transformers package to the latest version, but still the same problem. <|||||>I also tried to load the model using the TF checkpoint directly and it doesn't work:
```
import transformers
tokenizer = transformers.T5Tokenizer.from_pretrained("xxx")
config = transformers.T5Config.from_json_file("t5-11b-config.json")
model = transformers.TFT5Model(config,"xxx/")
encoded_input = tokenizer("A A", return_tensors='tf')
model(encoded_input)
```
Tensor:
`{'input_ids': <tf.Tensor: shape=(1, 2), dtype=int32, numpy=array([[71, 71]], dtype=int32)>, 'attention_mask': <tf.Tensor: shape=(1, 2), dtype=int32, numpy=array([[1, 1]], dtype=int32)>}`
Error:
```
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/mnt/lsf-nas-1/lsf/job/repo/elnaggar/anaconda3/envs/transformers/lib/python3.7/site-packages/tensorflow/python/keras/engine/base_layer.py", line 968, in __call__
outputs = self.call(cast_inputs, *args, **kwargs)
File "/mnt/lsf-nas-1/lsf/job/repo/elnaggar/anaconda3/envs/transformers/lib/python3.7/site-packages/transformers/modeling_tf_t5.py", line 1064, in call
training=training,
File "/mnt/lsf-nas-1/lsf/job/repo/elnaggar/anaconda3/envs/transformers/lib/python3.7/site-packages/tensorflow/python/keras/engine/base_layer.py", line 968, in __call__
outputs = self.call(cast_inputs, *args, **kwargs)
File "/mnt/lsf-nas-1/lsf/job/repo/elnaggar/anaconda3/envs/transformers/lib/python3.7/site-packages/transformers/modeling_tf_t5.py", line 610, in call
raise ValueError("You have to specify either inputs or inputs_embeds")
ValueError: You have to specify either inputs or inputs_embeds
```<|||||>Not solving your issue but curious, are you training your model on TPU with pytorch? I'm interested to know if pytorch or transformers supports training on v3-1024 TPU. <|||||>@misrasaurabh1 No we didn't use the pytorch version.<|||||>@sshleifer @patrickvonplaten Any update regarding the conversion from TF to Pytorch ?<|||||>For your first issue, I wonder if the t5 repo has changed, because our conversion script still works on a t5-base checkpoint I received last week. The checkpoint was probably trained a while ago though.
we also can't load t5-11b in the inference API, although the traceback is different.
@mfuntowicz have you looked at that at all/found anything?
I can help with your second issue. Try sending unpacking the elements of your input dict, so that they comply with the new T5 signature on master:
https://github.com/huggingface/transformers/blob/4dc65591b5c61d75c3ef3a2a883bf1433e08fc45/src/transformers/modeling_tf_t5.py#L1115<|||||>For the second issue, I was able to solve it by using the following:
```
ids = tokenizer.encode(seq, return_tensors='tf')
embedding = model(ids,decoder_input_ids=ids)
```
For some reason, I have to send also the ids as decoder_input_ids.
Is this correct @sshleifer or did I miss something ?
For the first issue, I could not solve it.
I can send you the T5 11B checkpoint privately if you need it.<|||||>Regarding "we also can't load t5-11b in the inference API, although the traceback is different.":
This applies to both Tensorflow and Pytorch versions ?<|||||>Please see https://github.com/huggingface/transformers/issues/5986#issuecomment-663090043<|||||>@julien-c :
My 11B checkpoint is only 20GB. It was trained on very small vocab.
I am not trying to use the pertained t5-11b from T5.
It is a custom new pertained model.
I don't believe #5986 (comment) is related to my problem.<|||||>Ok, my apologies for not reading your issue carefully enough. For the inference API, that's probably the reason though.<|||||>No problem, my main issue is the conversion from tensorflow to Pytorch for my new pertained 11B model.<|||||>@misrasaurabh1 @patrickvonplaten Any progress ?<|||||>I will take a deeper look next week on Monday when I'm back regarding the conversion from tensorflow to PyTorch.
Regarding, the `decoder_input_ids` ("For some reason, I have to send also the ids as decoder_input_ids.") -> yes for T5 you have to input both `input_ids` and `decoder_input_ids`, for the first forward call this is usually:
```python
from transformers import T5Model, T5Tokenizer
import torch
model = T5Model.from_pretrained("t5-small") # your model of choice here
tokenizer = T5Tokenizer.from_pretrained("t5-small")
input_ids = tokenizer("This sentence is encoded and only has to be passed once.", return_tensors="pt").input_ids
decoder_input_ids = torch.tensor([[model.config.decoder_start_token_id]])
# first call uses the start token id for T5 decoder
outputs = model(input_ids, decoder_input_ids=decoder_input_ids)
# now next decoder token can be sampled from decoder outputs
next_decoder_input_ids = torch.argmax(outputs.last_hidden_state, dim=-1)
decoder_input_ids = torch.cat([decoder_input_ids, next_decoder_input_ids], dim=-1)
# second call can reuse encoder outputs and continues auto-regressive language generation
# model.generate(...) does that automatically
outputs = model(encoder_outputs=outputs.decoder_past_key_values[0], decoder_input_ids=decoder_input_ids)
# The same holds true for TF
from transformers import TFT5Model, T5Tokenizer
import tensorflow as tf
model = TFT5Model.from_pretrained("t5-small") # your model of choice here
tokenizer = T5Tokenizer.from_pretrained("t5-small")
input_ids = tokenizer("This sentence is encoded and only has to be passed once.", return_tensors="tf").input_ids
decoder_input_ids = tf.convert_to_tensor([[model.config.decoder_start_token_id]], dtype=tf.dtypes.int32)
# first call uses the start token id for T5 decoder
outputs = model(input_ids, decoder_input_ids=decoder_input_ids)
# now next decoder token can be sampled from decoder outputs
next_decoder_input_ids = tf.cast(tf.argmax(outputs[0], axis=-1), tf.dtypes.int32)
decoder_input_ids = tf.concat([decoder_input_ids, next_decoder_input_ids], axis=-1)
# second call can reuse encoder outputs and continues auto-regressive language generation
# model.generate(...) does that automatically
# inputs (encoder inputs) has to be defined because of keras convention => it's not needed though after first
# pass, so set to None
model(None, encoder_outputs=outputs[1][0], decoder_input_ids=decoder_input_ids)
```
So this means that the `decoder_input_ids` should not be the same as the `input_ids` (encoder inputs) but should correspond to the auto-regressively generated text by the model, that starts with the `decoder_start_token_id`. Hope this makes it a bit clearer.<|||||>Thanks @patrickvonplaten for the clarification.
In this case, I would recommend to update the documentation on:
https://github.com/huggingface/transformers/blob/master/src/transformers/modeling_tf_t5.py#L972
Just to clarify one thing, If I am only interested on feature extraction and not generation, in this case my code should looks like:
```
from transformers import TFT5Model, T5Tokenizer
import tensorflow as tf
model = TFT5Model.from_pretrained("t5-small") # your model of choice here
tokenizer = T5Tokenizer.from_pretrained("t5-small")
input_ids = tokenizer("This sentence is encoded and only has to be passed once.", return_tensors="tf").input_ids
decoder_input_ids = tf.convert_to_tensor([[model.config.decoder_start_token_id]], dtype=tf.dtypes.int32)
# first call uses the start token id for T5 decoder
outputs = model(input_ids, decoder_input_ids=decoder_input_ids)
features = outputs[2]
```
Correct ?
Hopefully next week you will figure out the problem of Pytorch conversion, and let me know if you need our checkpoint, we can send it to you privately. <|||||>Yes, we should update the example to clarify a bit...`T5` is an encoder-decoder model, so it's usually used for conditional generation IMO, so to model the following distribution:
with y being the `decoder_input_ids` and x being the `input_ids`.
So for feature extraction, is the idea to extract the features of the first `decoder_input_ids` token, so:
=> so the hidden states of the first output? In this case, your code above is correct.
Can you give me a bit more details of the task and how / with what objective (masked language modeling, ...) the model was trained?
<|||||>Would be great if you can send me the weights of your trained model per mail (Weshare, google drive) then I check out the problem with the conversion next week <|||||>Thanks @patrickvonplaten for the clarification. I am glad that my understanding is correct, and I am extracting the features from T5 correctly.
Regarding your questions:
1. Task: We are training different language models for the language of life "protein sequences".
2. Objective: Span de-noising.
3. Models: 11B and 3B models from T5.
You can find more details on my GitHub repo and my paper first version:
https://github.com/agemagician/ProtTrans
https://www.biorxiv.org/content/10.1101/2020.07.12.199554v2
I have sent you an email with the 11b model weights, config file and spm model.
I wish you a nice weekend and I hope you could figure out the problem next week.
Thanks in advance.<|||||>We have also tested the 3b parameter model with shared enc/dec weights and it did failed.
```
2020-08-03 18:20:42.993212: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1
Building PyTorch model from configuration: T5Config {
"architectures": [
"T5WithLMHeadModel"
],
"d_ff": 5120,
"d_kv": 64,
"d_model": 2048,
"decoder_start_token_id": 0,
"dropout_rate": 0.0,
"eos_token_id": 1,
"initializer_factor": 1.0,
"is_encoder_decoder": true,
"layer_norm_epsilon": 1e-06,
"model_type": "t5",
"n_positions": 512,
"num_heads": 32,
"num_layers": 24,
"output_past": true,
"pad_token_id": 0,
"relative_attention_num_buckets": 32,
"vocab_size": 128
}
INFO:transformers.modeling_t5:Converting TensorFlow checkpoint from /content/models/ProtT5-BFD
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/relative_attention_bias with shape [32, 32]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/relative_attention_bias_slot_v with shape [32, 32]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_000/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_001/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_002/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_003/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_004/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_005/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_006/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_007/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_008/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_009/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_010/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_011/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_012/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_013/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_014/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_015/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_016/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_017/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_018/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_019/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_020/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_021/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_022/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/dense_relu_dense/DenseReluDense/wi/kernel with shape [2048, 5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/dense_relu_dense/DenseReluDense/wi/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/dense_relu_dense/DenseReluDense/wi/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/dense_relu_dense/DenseReluDense/wo/kernel with shape [5120, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/dense_relu_dense/DenseReluDense/wo/kernel_slot_vc with shape [5120]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/dense_relu_dense/DenseReluDense/wo/kernel_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/dense_relu_dense/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/dense_relu_dense/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/EncDecAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/EncDecAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/EncDecAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/EncDecAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/EncDecAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/EncDecAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/EncDecAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/EncDecAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/EncDecAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/EncDecAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/EncDecAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/EncDecAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/enc_dec_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/SelfAttention/k with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/SelfAttention/k_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/SelfAttention/k_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/SelfAttention/o with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/SelfAttention/o_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/SelfAttention/o_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/SelfAttention/q with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/SelfAttention/q_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/SelfAttention/q_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/SelfAttention/v with shape [2048, 2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/SelfAttention/v_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/SelfAttention/v_slot_vr with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/block_023/self_attention/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/logits/kernel with shape [2048, 128]
INFO:transformers.modeling_t5:Loading TF weight decoder/logits/kernel_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/logits/kernel_slot_vr with shape [128]
INFO:transformers.modeling_t5:Loading TF weight decoder/rms_norm/scale with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight decoder/rms_norm/scale_slot_v with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight global_step with shape []
INFO:transformers.modeling_t5:Loading TF weight shared/embedding with shape [128, 2048]
INFO:transformers.modeling_t5:Loading TF weight shared/embedding_slot_vc with shape [2048]
INFO:transformers.modeling_t5:Loading TF weight shared/embedding_slot_vr with shape [128]
INFO:transformers.modeling_t5:Skipping decoder/block_000/dense_relu_dense/DenseReluDense/wi/kernel
INFO:transformers.modeling_t5:Skipping decoder/block_000/dense_relu_dense/DenseReluDense/wi/kernel
INFO:transformers.modeling_t5:Skipping decoder/block_000/dense_relu_dense/DenseReluDense/wi/kernel
Traceback (most recent call last):
File "/usr/local/lib/python3.6/dist-packages/transformers/convert_t5_original_tf_checkpoint_to_pytorch.py", line 61, in <module>
convert_tf_checkpoint_to_pytorch(args.tf_checkpoint_path, args.config_file, args.pytorch_dump_path)
File "/usr/local/lib/python3.6/dist-packages/transformers/convert_t5_original_tf_checkpoint_to_pytorch.py", line 36, in convert_tf_checkpoint_to_pytorch
load_tf_weights_in_t5(model, config, tf_checkpoint_path)
File "/usr/local/lib/python3.6/dist-packages/transformers/modeling_t5.py", line 104, in load_tf_weights_in_t5
pointer = getattr(pointer, "weight")
File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 772, in __getattr__
type(self).__name__, name))
torch.nn.modules.module.ModuleAttributeError: 'T5Block' object has no attribute 'weight'
```<|||||>Ok, I will start with the 3B model then! Starting downloading the model now -> will take a look tomorrow (my internet connection is so slow at the moment that it'll take a couple of hours to download both models).
Just to be sure:
You have trained the model using this repo: https://github.com/google-research/text-to-text-transfer-transformer right?
And then you ran
```
python convert_t5_original_tf_checkpoint_to_pytorch.py
```
-> which did fail no? <|||||>Thanks @patrickvonplaten.
Yes, I have used the official repo:
https://github.com/google-research/text-to-text-transfer-transformer
Yes, I have used "convert_t5_original_tf_checkpoint_to_pytorch.py", which failed.
An important note:
1) 11B used separate weights for encoder/decoder.
2) 3B used a shared weights for encoder and decoder.<|||||>Hey @agemagician,
I changed the conversion function to handle the 3B weights you send me. I have locally saved the converted pytorch model and you can do the conversion yourself using [this](https://github.com/huggingface/transformers/tree/adapt_t5_for_covid_19_3b) branch also with the changes shown in this PR: https://github.com/huggingface/transformers/pull/6388.
A couple of things that might be important:
1) I noticed that the input embeddings, called `shared/embedding` and output logits, called `decoder/logits/kernel` were different meaning that for the decoder the input and output embeddings are **not** shared. We usually share input and output embeddings in T5, so in your case I added a couple of lines to make sure that the correct input and output embeddings are converted and **not** tied. See these changes: https://github.com/huggingface/transformers/blob/abec44887d48e3ef3ccde50f3331f64aa2b3cbbc/src/transformers/modeling_t5.py#L1094
2) For convenience, I did not tie the weights of the encoder / decoder now, but just copied them. This meants that the resulting 3B T5 model has a size of ~8GB (2 times bigger because float32 is used instead of TPU's bfloat16 and again 2 times bigger because encoder and decoder are not shared). Because of the small word embedding matrix the model is still smaller than the official `t5-3b`: https://huggingface.co/t5-3b. You can reduce the size of your model, by calling `model.half()` and deleting `model.encoder` before doing `model.save_pretrained("./")`. When loading you will have to write a short script though that instantiates the encoder and correctly sets the encoder weights. I would not recommend to do this, but instead just use the 8GB model in PyTorch.
3) I did not test the conversion of the 11B model because I currently do not have the RAM capacity to do so (32GB is just not enough for such a large model). The script should work though. You should make sure to set `is_tied` to `False` in this line though: https://github.com/huggingface/transformers/blob/abec44887d48e3ef3ccde50f3331f64aa2b3cbbc/src/transformers/modeling_t5.py#L64. Also note that the conversion will automatically initialize the model in fp32, so that the PyTorch model alone requires 40GB of RAM. You can later save it as `model.half()`. I hope you have enough RAM space to be able to convert the model. It will be hard to debug for me without the necessary resources, so in case you experience errors, I would suggest to start debugging in this function: https://github.com/huggingface/transformers/blob/abec44887d48e3ef3ccde50f3331f64aa2b3cbbc/src/transformers/modeling_t5.py#L64 and see what names seem to be wrong.<|||||>Thanks a lot @patrickvonplaten for fixing the conversion script.
We highly appreciate your effort.
This will definitely make it easier for researchers to use our T5 protein language models.
I have checked the following models:
1. 3B shared.
2. 3B non-shared.
3. 11B non-shared.
All of them were converted correctly without any issue.
However, following your example above for using the models, the decoder generates out of vocabulary ids.
Despite, that both the vocab_size on the config file and the spm is set correctly to 128.
also I got the following warning when loading the model:
```
Some weights of the model checkpoint at xxx/3b-shared/pytorch/ were not used when initializing T5Model: ['lm_head.weight']
- This IS expected if you are initializing T5Model from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPretraining model).
- This IS NOT expected if you are initializing T5Model from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).
```
Error:
```
---------------------------------------------------------------------------
IndexError Traceback (most recent call last)
<ipython-input-66-c65c815d3456> in <module>
2 # model.generate(...) does that automatically
3
----> 4 outputs = model(encoder_outputs=outputs.decoder_past_key_values[0], decoder_input_ids=decoder_input_ids)
/mnt/lsf-nas-1/lsf/job/repo/elnaggar/anaconda3/envs/transformers/lib/python3.8/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
720 result = self._slow_forward(*input, **kwargs)
721 else:
--> 722 result = self.forward(*input, **kwargs)
723 for hook in itertools.chain(
724 _global_forward_hooks.values(),
/mnt/lsf-nas-1/lsf/job/repo/elnaggar/anaconda3/envs/transformers/lib/python3.8/site-packages/transformers/modeling_t5.py in forward(self, input_ids, attention_mask, encoder_outputs, decoder_input_ids, decoder_attention_mask, decoder_past_key_values, use_cache, inputs_embeds, decoder_inputs_embeds, head_mask, output_attentions, output_hidden_states, return_tuple, **kwargs)
995
996 # Decode
--> 997 decoder_outputs = self.decoder(
998 input_ids=decoder_input_ids,
999 attention_mask=decoder_attention_mask,
/mnt/lsf-nas-1/lsf/job/repo/elnaggar/anaconda3/envs/transformers/lib/python3.8/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
720 result = self._slow_forward(*input, **kwargs)
721 else:
--> 722 result = self.forward(*input, **kwargs)
723 for hook in itertools.chain(
724 _global_forward_hooks.values(),
/mnt/lsf-nas-1/lsf/job/repo/elnaggar/anaconda3/envs/transformers/lib/python3.8/site-packages/transformers/modeling_t5.py in forward(self, input_ids, attention_mask, encoder_hidden_states, encoder_attention_mask, inputs_embeds, head_mask, past_key_value_states, use_cache, output_attentions, output_hidden_states, return_tuple)
701 if inputs_embeds is None:
702 assert self.embed_tokens is not None, "You have to intialize the model with valid token embeddings"
--> 703 inputs_embeds = self.embed_tokens(input_ids)
704
705 batch_size, seq_length = input_shape
/mnt/lsf-nas-1/lsf/job/repo/elnaggar/anaconda3/envs/transformers/lib/python3.8/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
720 result = self._slow_forward(*input, **kwargs)
721 else:
--> 722 result = self.forward(*input, **kwargs)
723 for hook in itertools.chain(
724 _global_forward_hooks.values(),
/mnt/lsf-nas-1/lsf/job/repo/elnaggar/anaconda3/envs/transformers/lib/python3.8/site-packages/torch/nn/modules/sparse.py in forward(self, input)
122
123 def forward(self, input: Tensor) -> Tensor:
--> 124 return F.embedding(
125 input, self.weight, self.padding_idx, self.max_norm,
126 self.norm_type, self.scale_grad_by_freq, self.sparse)
/mnt/lsf-nas-1/lsf/job/repo/elnaggar/anaconda3/envs/transformers/lib/python3.8/site-packages/torch/nn/functional.py in embedding(input, weight, padding_idx, max_norm, norm_type, scale_grad_by_freq, sparse)
1812 # remove once script supports set_grad_enabled
1813 _no_grad_embedding_renorm_(weight, input, max_norm, norm_type)
-> 1814 return torch.embedding(weight, input, padding_idx, scale_grad_by_freq, sparse)
1815
1816
IndexError: index out of range in self
```
When looking into the decoder_input_ids, I found the following out of id values:
`tensor([[ 0, 150]])`
It seems like the decoder ignores the actual vocab size.
Any idea what is the problem here ?<|||||>I cannot reproduce the error with my weights. Can you add a code sample that helps me to reproduce the error?<|||||>I imagine you don't use a tokenizer, right?<|||||>Hi @patrickvonplaten ,
Thanks for reply and I hope you had a good week-end.
I am using the tokenizer.
I have created a Colab notebook which replicates my issue:
https://colab.research.google.com/drive/1NccKDsKObYYTjf8EPcrUUDfs0PyaRsaN?usp=sharing
You can see at the last cell:
`tensor([[ 0, 121, 121, 905]])`
905 is out of vocab id.
Note:
The model used in the Colab notebook is a dummy checkpoint for the 3B model (.i.e.: checkpoint 0). <|||||>I see what the problem is - you should use `T5ForConditionalGeneration` instead of `T5Model`. The T5Model does not have the `lm_head` on top and therefore returns a matrix of length 1024 instead of 128.
Here a notebook that runs without error:
https://colab.research.google.com/drive/1u3VxPkFfrm0Z3f6SMIZTXbazFUhXNe0Z?usp=sharing.
Also, you have to make sure that the model you pass for the conversion script is a `T5ForConditionalGeneration` model and not only a `T5Model`.
It might be possible that the checkpoint of your google colab is not correct because of this.
Can you also check this 3b checkpoint that I converted as explained above:
```python
config = T5Config.from_pretrained("patrickvonplaten/t5_3b_covid")
model = T5ForConditionalGeneration.from_pretrained("patrickvonplaten/t5_3b_covid")
```
As explained above, the output word embedding (`lm_head`) of the 'patrickvonplaten/t5_3b_covid' model are different to the input embedding, so it would be good if you check this model to your checkpoint in the notebook to see which performs as expected (hopefully at least one of them).
Let me know if this works. PS. I'm currently working on adding better functionality to share encoder and decoder weights for Seq2Seq models. Once this is done we can reduce this checkpoint to half its size.
<|||||>Thanks @patrickvonplaten for the clarification.
Using T5ForConditionalGeneration did solve my issue, and now my models work as excepted.
Thanks a lot for optimizing the shared encoder/decoder weights.
This will be really helpful to allow large models 3B/11B to fit on a single GPU for inference.
This will be the last point to cover before closing this issue.
<|||||>One more thing, that will be very useful for T5 large models 3B/11B is to enable gradient_checkpointing as Bert Model:
https://github.com/huggingface/transformers/blob/0735def8e1200ed45a2c33a075bc1595b12ef56a/src/transformers/modeling_bert.py#L461
This will be very useful to fine-tune such models on a single GPU.
I think it should be adjust here, if I am not mistaken:
https://github.com/huggingface/transformers/blob/master/src/transformers/modeling_t5.py#L752
<|||||>Will open a PR to tie encoder / decoder weights tomorrow!
Regarding the gradient checkpointing - yes it would be nice to add this. Feel free to open a new feature request "T5 Checkpointing" where you can tag me or @LysandreJik. <|||||>Thanks a lot @patrickvonplaten , you are the best.
Perfect, then I will add a new request for Checkpointing.
Feel free to close this issue after the PR.<|||||>PR is ready for review - should be merged this in the next few days. <|||||>Perfect, thanks a lot @patrickvonplaten.
Your help and support is highly appreciated by us. |
transformers | 5,850 | closed | pip install error | Hi, I tried to install transformers library via `pip install transformers` and I got tokenizer install error.
The error logs are as follows.
```
ERROR: Could not find a version that satisfies the requirement tokenizers==0.8.1.rc1 (from transformers) (from versions: 0.0.2, 0.0.3, 0.0.4, 0.0.5, 0.0.6, 0.0.7, 0.0.8, 0.0.9, 0.0.10, 0.0.11, 0.0.12, 0.0.13, 0.1.0, 0.1.1, 0.2.0, 0.2.1, 0.3.0, 0.4.0, 0.4.1, 0.4.2, 0.5.0, 0.5.1, 0.5.2, 0.6.0, 0.7.0, 0.8.0)
ERROR: No matching distribution found for tokenizers==0.8.1.rc1 (from transformers)
```
I googled about it but I couldn't find the way to solve it.
Is there I can do to handle this issue?
Thank you. | 07-17-2020 16:22:49 | 07-17-2020 16:22:49 | @devJWSong, how did you solve this?<|||||>Oh, actually I didn't solve it.
I just installed downgraded version which is 2.11.0. and it worked.
I still don't know the reason but I think it is the problem from my virtual environment setting since when I tried to install the recent version in the different environment, it worked...<|||||>its error occurs to me too.... could you give me another solution about that problems? ...<|||||>Any updates on this problem?<|||||>Try changing index-url and trusted-host in pip config.
I had same issue with the environment with index-url='http://ftp.daumkakao.com/pypi/simple' and trusted-host='ftp.daumkakao.com', but everything worked well with the environment without such config.
+)
try `pip install transformers -i https://pypi.python.org/simple`<|||||>Thank you! It worked. :) |
transformers | 5,849 | closed | Update shortcut name for reformer in pretrained_models.srt | Adding `google/` prefix on shortcut name for reformer in `pretrained_models.srt` according to mapping in `configuration_reformer.py` | 07-17-2020 16:15:37 | 07-17-2020 16:15:37 | I agree with the change but I think you probably broke the rst<|||||>> I agree with the change but I think you probably broke the rst
Ah, my bad, I have updated the rst file<|||||># [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5849?src=pr&el=h1) Report
> Merging [#5849](https://codecov.io/gh/huggingface/transformers/pull/5849?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/9d37c56bab8f7f1f1aa0b65be039516072254e77&el=desc) will **decrease** coverage by `1.19%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5849?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5849 +/- ##
==========================================
- Coverage 78.48% 77.28% -1.20%
==========================================
Files 146 146
Lines 26200 26200
==========================================
- Hits 20563 20249 -314
- Misses 5637 5951 +314
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5849?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5849/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.39%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5849/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5849?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5849?src=pr&el=footer). Last update [9d37c56...1103b9e](https://codecov.io/gh/huggingface/transformers/pull/5849?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>I check that this update has been included in the main repo directly, closing this PR |
transformers | 5,848 | closed | AttributeError: type object 'BertConfig' has no attribute 'pretrained_config_archive_map' | Hi,
After changing from `BERT_PRETRAINED_MODEL_ARCHIVE_MAP` to `BERT_PRETRAINED_MODEL_ARCHIVE_LIST` in #5842 , I encountered another issue:
> File "main.py", line 24, in <genexpr>
> ALL_MODELS = sum((tuple(conf.pretrained_config_archive_map.keys()) for conf in (BertConfig, XLNetConfig)), ())
> AttributeError: type object 'BertConfig' has no attribute 'pretrained_config_archive_map'
Is it also a breaking change ?
What is the replacing name for `pretrained_config_archive_map` now ?
Thank you, | 07-17-2020 16:05:04 | 07-17-2020 16:05:04 | it'd better to downgrade the version: pip install transformer==2.0.0 (which solves the compatible issue).<|||||>`$ pip install transformers==2.0.0`
works<|||||>Hi, is there any other approach other than downgrading?<|||||>try
> `$from transformers import ALBERT_PRETRAINED_CONFIG_ARCHIVE_MAP`
or
`$from transformers import BERT_PRETRAINED_CONFIG_ARCHIVE_MAP`<|||||>Hi! Could you let us know the use cases of using the archive maps?<|||||>
Hi,
<img width="712" alt="截屏2021-02-26 下午9 38 01" src="https://user-images.githubusercontent.com/7147150/109307249-49d5f000-787b-11eb-91eb-f076046f180a.png">
I used it to provide the hint of cmd args configuration.<|||||>I see, thank you! The issue is that these archive maps are outdated and don't mean much since the model hub.
In that case, it would be better to have a way to see all checkpoints on the hub corresponding to that model architecture, right?<|||||>If anyone has already solved the issue please let me know. I am having same issue.
https://hjlabs.in<|||||>`$from transformers import ALBERT_PRETRAINED_CONFIG_ARCHIVE_MAP`
this solved my issue. Thank you everyone for support. you are great.
https://hjlabs.in |
transformers | 5,847 | closed | Create README.md | 07-17-2020 15:26:58 | 07-17-2020 15:26:58 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5847?src=pr&el=h1) Report
> Merging [#5847](https://codecov.io/gh/huggingface/transformers/pull/5847?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/9d37c56bab8f7f1f1aa0b65be039516072254e77&el=desc) will **increase** coverage by `0.00%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5847?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5847 +/- ##
=======================================
Coverage 78.48% 78.48%
=======================================
Files 146 146
Lines 26200 26200
=======================================
+ Hits 20563 20564 +1
+ Misses 5637 5636 -1
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5847?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5847/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.46% <0.00%> (+0.25%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5847?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5847?src=pr&el=footer). Last update [9d37c56...07227e0](https://codecov.io/gh/huggingface/transformers/pull/5847?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,846 | closed | Decode [UNK] from tokenizer | I am using Bert for NER and I have a problem when the tokenizer uses the token [UNK] to mask a word. How can I easily know which was the original work given its index? | 07-17-2020 15:06:41 | 07-17-2020 15:06:41 | @RodSernaPerez I am not sure what you are doing, but the model will do that if it doesn't find your word in the vocab.txt file. You have to use `BertTokenizer` before training, so that it splits all of the words that are not in the vocab file into word tokens. I think it will mark words that it wasn't able to tokenize with `[UNK]`. As to your question, what do you mean by index? I am assuming you know where the `[UNK]` was in white space tokenized word. You can concatenate BERT word tokens first and then you can try to find word marked with `[UNK]` in that list of concatenated tokens.<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,845 | closed | Added model card for neuraly/bert-base-italian-cased-sentiment | Hello, we are making this pull request to add our Italian sentiment model to your repository of transformers.
Thank you again for hosting the models :) | 07-17-2020 12:57:32 | 07-17-2020 12:57:32 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5845?src=pr&el=h1) Report
> Merging [#5845](https://codecov.io/gh/huggingface/transformers/pull/5845?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/0b6c255a95368163d2b1d37635e5ce5bdd1b9423&el=desc) will **decrease** coverage by `1.15%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5845?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5845 +/- ##
==========================================
- Coverage 78.50% 77.35% -1.16%
==========================================
Files 146 146
Lines 26049 26049
==========================================
- Hits 20450 20150 -300
- Misses 5599 5899 +300
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5845?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5845/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.39%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5845/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `83.95% <0.00%> (-2.26%)` | :arrow_down: |
| [src/transformers/generation\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5845/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3V0aWxzLnB5) | `97.11% <0.00%> (+0.28%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5845/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5845?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5845?src=pr&el=footer). Last update [0b6c255...cf779f6](https://codecov.io/gh/huggingface/transformers/pull/5845?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>Hello guys, nice job. The tutorial seems broken to me.
https://github.com/huggingface/transformers/issues/14194 |
transformers | 5,844 | closed | problem about Custom class inheriting from <TFBertPreTrainedModel> | # ❓ Questions & Help
## Details
<!-- Description of your issue -->
I constructed class inheriting from TFBertPreTrainedModel. as follows:
```
class Model(transformers.TFBertPreTrainedModel):
def __init__(self,config, *inputs, **kwargs):
super(Model,self).__init__(config, *inputs, **kwargs)
self.bert = transformers.TFBertMainLayer(config,name="bert")
@tf.function
def call(self, inputs, training=None, mask=None,**kwargs):
out = self.bert(inputs)
print(out)
return out
```
main progress as follow:
```
if __name__ == "__main__":
tokenizer = transformers.BertTokenizer("model/chinese_L-12_H-768_A-12/vocab.txt")
text_2 = tokenizer.batch_encode_plus(["你买啊，买了就是成都人", "你来啊，来了就是深圳人"], max_length=20, pad_to_max_length=True)
print(text_2)
model = Model.from_pretrained("bert-base-chinese")
out = model([tf.convert_to_tensor(text_2["input_ids"]),tf.convert_to_tensor(text_2['attention_mask'])])
```
unfortunately，the print on console was different for what I expect when I run this program. I expect that the out of **print(out)** is a tensor with size [2,20,768], but in fact, the print was run three times, outputs is **(<tf.Tensor 'bert/encoder/layer_._11/output/LayerNorm/batchnorm/add_1:0' shape=(3, 5, 768) dtype=float32>, <tf.Tensor 'bert/pooler/dense/Tanh:0' shape=(3, 768) dtype=float32>)** , **(<tf.Tensor 'bert/encoder/layer_._11/output/LayerNorm/batchnorm/add_1:0' shape=(3, 5, 768) dtype=float32>, <tf.Tensor 'bert/pooler/dense/Tanh:0' shape=(3, 768) dtype=float32>)** and **(<tf.Tensor 'bert/encoder/layer_._11/output/LayerNorm/batchnorm/add_1:0' shape=(2, 20, 768) dtype=float32>, <tf.Tensor 'bert/pooler/dense/Tanh:0' shape=(2, 768) dtype=float32>)**
My question are 1) why 2)how could I get the last outputs for following layers
I am looking forward your help, thanks
<!-- You should first ask your question on the forum or SO, and only if
you didn't get an answer ask it here on GitHub. -->
**A link to original question on Stack Overflow**: | 07-17-2020 12:16:39 | 07-17-2020 12:16:39 | We are trying to move these kinds of questions (which are not obvious bugs, but question for customized use cases) more and more to https://discuss.huggingface.co/. Would you mind posting it there again with your sample code:
```python
class Model(transformers.TFBertPreTrainedModel):
def __init__(self,config, *inputs, **kwargs):
super(Model,self).__init__(config, *inputs, **kwargs)
self.bert = transformers.TFBertMainLayer(config,name="bert")
@tf.function
def call(self, inputs, training=None, mask=None,**kwargs):
out = self.bert(inputs)
print(out)
return out
```<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,843 | closed | Added model card for neuraly/bert-base-italian-cased-sentiment | Hello, we are making this pull request to add our Italian sentiment model to your repository of transformers.
Thank you again for hosting the models :) | 07-17-2020 10:23:07 | 07-17-2020 10:23:07 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5843?src=pr&el=h1) Report
> Merging [#5843](https://codecov.io/gh/huggingface/transformers/pull/5843?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/0b6c255a95368163d2b1d37635e5ce5bdd1b9423&el=desc) will **decrease** coverage by `0.39%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5843?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5843 +/- ##
==========================================
- Coverage 78.50% 78.11% -0.40%
==========================================
Files 146 146
Lines 26049 26049
==========================================
- Hits 20450 20347 -103
- Misses 5599 5702 +103
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5843?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5843/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5843/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5843/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.96% <0.00%> (-0.26%)` | :arrow_down: |
| [src/transformers/file\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5843/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.49% <0.00%> (+0.29%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5843/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5843?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5843?src=pr&el=footer). Last update [0b6c255...d6d30df](https://codecov.io/gh/huggingface/transformers/pull/5843?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
transformers | 5,842 | closed | ImportError: cannot import name 'BERT_PRETRAINED_MODEL_ARCHIVE_MAP' from 'transformers' | Hi,
I'm rerunning the code BERT-E2E-ABSA from https://github.com/lixin4ever/BERT-E2E-ABSA, which is based on transformers 2.0.0
I installed the latest transformers 3.0.2 (both from pip and souce) and I have an error that I can not import BERT_PRETRAINED_MODEL_ARCHIVE_MAP.
This simple command will result in errors below:
> python -c "from transformers import BERT_PRETRAINED_MODEL_ARCHIVE_MAP"
> Traceback (most recent call last):
> File "<string>", line 1, in <module>
> ImportError: cannot import name 'BERT_PRETRAINED_MODEL_ARCHIVE_MAP' from 'transformers' (/home/hoa/transformers/src/transformers/__init__.py)
Transformer is well installed as it can do the following:
> python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('I hate you'))"
> [{'label': 'NEGATIVE', 'score': 0.9991129040718079}]
Do you know why ? Did I miss something ?
- `transformers` version: 3.0.2
- Platform: Linux-5.4.0-1019-gcp-x86_64-with-debian-bullseye-sid
- Python version: 3.7.7
- PyTorch version (GPU?): 1.2.0 (True)
- Tensorflow version (GPU?): not installed (NA)
- Using GPU in script?: Yes
- Using distributed or parallel set-up in script?: No
- Other configurations: Ubuntu 20.04, Python 3.7, CUDA 11
Note: I even installed pytorch on cpu laptop and rerun, still had the same issue !
| 07-17-2020 10:01:54 | 07-17-2020 10:01:54 | Yeah, there was a breaking change. The name has been changed to `BERT_PRETRAINED_MODEL_ARCHIVE_LIST`.
See https://github.com/huggingface/transformers/pull/4636.<|||||>ImportError: cannot import name 'BertweetTokenizer' from 'transformers' (/opt/conda/lib/python3.7/site-packages/transformers/__init__.py)
<|||||>ImportError: cannot import name 'SquadExample' from 'transformers' (unknown location)
I get this error when I try to import QuestionAnsweringModel from simpletransformers.question_answering
Any help would be appreciated! |
transformers | 5,841 | closed | [Model card] Bert2Bert | 07-17-2020 09:24:38 | 07-17-2020 09:24:38 | ||
transformers | 5,840 | closed | [WIP - Don't merge][EncoderDecoder] Extend Trainer for Bert2Bert | This PR draft should be used to train a Bert2Bert model.
It's not ready to be merged yet. | 07-17-2020 09:20:09 | 07-17-2020 09:20:09 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5840?src=pr&el=h1) Report
> Merging [#5840](https://codecov.io/gh/huggingface/transformers/pull/5840?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/fa5423b1695cd24856bcff47214172e0f540d924&el=desc) will **decrease** coverage by `1.26%`.
> The diff coverage is `41.17%`.
[](https://codecov.io/gh/huggingface/transformers/pull/5840?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5840 +/- ##
==========================================
- Coverage 77.79% 76.53% -1.27%
==========================================
Files 145 145
Lines 25355 25365 +10
==========================================
- Hits 19726 19414 -312
- Misses 5629 5951 +322
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5840?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/5840/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `37.41% <37.50%> (-0.55%)` | :arrow_down: |
| [src/transformers/training\_args.py](https://codecov.io/gh/huggingface/transformers/pull/5840/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmluZ19hcmdzLnB5) | `77.77% <100.00%> (+0.22%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5840/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.39%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5840/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5840?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5840?src=pr&el=footer). Last update [fa5423b...d9f6d07](https://codecov.io/gh/huggingface/transformers/pull/5840?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,839 | closed | Created model card for my extreme summarization model | 07-17-2020 08:53:22 | 07-17-2020 08:53:22 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5839?src=pr&el=h1) Report
> Merging [#5839](https://codecov.io/gh/huggingface/transformers/pull/5839?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/3d9556a72b6709d5fa09bf7bce7404158c169d21&el=desc) will **decrease** coverage by `0.72%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5839?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5839 +/- ##
==========================================
- Coverage 78.11% 77.38% -0.73%
==========================================
Files 146 146
Lines 26049 26049
==========================================
- Hits 20347 20159 -188
- Misses 5702 5890 +188
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5839?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5839/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.39%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5839/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `43.98% <0.00%> (-49.38%)` | :arrow_down: |
| [src/transformers/file\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5839/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.19% <0.00%> (-0.30%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5839/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.46% <0.00%> (+0.50%)` | :arrow_up: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5839/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `82.31% <0.00%> (+1.28%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5839/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5839/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `95.18% <0.00%> (+74.91%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5839?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5839?src=pr&el=footer). Last update [3d9556a...05a0062](https://codecov.io/gh/huggingface/transformers/pull/5839?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,838 | closed | Created model card for my summarization model | 07-17-2020 08:25:39 | 07-17-2020 08:25:39 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5838?src=pr&el=h1) Report
> Merging [#5838](https://codecov.io/gh/huggingface/transformers/pull/5838?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/3d9556a72b6709d5fa09bf7bce7404158c169d21&el=desc) will **increase** coverage by `0.39%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5838?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5838 +/- ##
==========================================
+ Coverage 78.11% 78.50% +0.39%
==========================================
Files 146 146
Lines 26049 26049
==========================================
+ Hits 20347 20450 +103
+ Misses 5702 5599 -103
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5838?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5838/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `43.98% <0.00%> (-49.38%)` | :arrow_down: |
| [src/transformers/file\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5838/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.19% <0.00%> (-0.30%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5838/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (+0.25%)` | :arrow_up: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5838/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `82.31% <0.00%> (+1.28%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5838/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `95.18% <0.00%> (+74.91%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5838?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5838?src=pr&el=footer). Last update [3d9556a...b6dd51b](https://codecov.io/gh/huggingface/transformers/pull/5838?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>Thank you! |
|
transformers | 5,837 | closed | [seq2seq] MAX_LEN env var for MT commands | 07-17-2020 07:06:39 | 07-17-2020 07:06:39 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5837?src=pr&el=h1) Report
> Merging [#5837](https://codecov.io/gh/huggingface/transformers/pull/5837?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/529850ae7bca0ff388778c3c0d66240834cf56c3&el=desc) will **decrease** coverage by `1.18%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5837?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5837 +/- ##
==========================================
- Coverage 78.48% 77.30% -1.19%
==========================================
Files 146 146
Lines 26200 26049 -151
==========================================
- Hits 20563 20137 -426
- Misses 5637 5912 +275
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5837?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5837/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.39%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5837/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |
| [src/transformers/modeling\_reformer.py](https://codecov.io/gh/huggingface/transformers/pull/5837/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19yZWZvcm1lci5weQ==) | `87.86% <0.00%> (-2.71%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5837/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.46% <0.00%> (+0.25%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5837/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5837/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `95.53% <0.00%> (+69.51%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5837?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5837?src=pr&el=footer). Last update [529850a...ea4f6b0](https://codecov.io/gh/huggingface/transformers/pull/5837?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,836 | closed | [not sure whether to] pin torch<=1.5.1 | 07-17-2020 05:39:02 | 07-17-2020 05:39:02 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5836?src=pr&el=h1) Report
> Merging [#5836](https://codecov.io/gh/huggingface/transformers/pull/5836?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/d088d744adb4e5aa45262a34acab3ae9e81de169&el=desc) will **decrease** coverage by `0.98%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5836?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5836 +/- ##
==========================================
- Coverage 78.10% 77.12% -0.99%
==========================================
Files 146 146
Lines 26047 26047
==========================================
- Hits 20344 20088 -256
- Misses 5703 5959 +256
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5836?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5836/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.39%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5836/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `26.02% <0.00%> (-69.52%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5836/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (+0.25%)` | :arrow_up: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5836/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `82.31% <0.00%> (+1.28%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5836/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5836/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `95.18% <0.00%> (+74.91%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5836?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5836?src=pr&el=footer). Last update [d088d74...dd8448b](https://codecov.io/gh/huggingface/transformers/pull/5836?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,835 | closed | solve Illegal seek in wandb teardown (test suite) | somehow wandb teardown fails in `tests/test_trainer.py::TrainerIntegrationTest::test_trainer_eval_mrpc`:
```
pytest -n 1 tests/test_trainer.py::TrainerIntegrationTest::test_trainer_eval_mrpc
====================================================================== test session starts =======================================================================
platform linux -- Python 3.7.5, pytest-5.4.3, py-1.9.0, pluggy-0.13.1
rootdir: /mnt/nvme1/code/huggingface/transformers-tests-1
plugins: hypothesis-5.5.4, filter-subpackage-0.1.1, arraydiff-0.3, flaky-3.6.1, ipynb-1.1.1.dev0, cov-2.10.0, astropy-header-0.1.2, forked-1.2.0, doctestplus-0.5.0, openfiles-0.4.0, remotedata-0.3.2, xdist-1.32.0
gw0 [1]
FE [100%]
wandb: Waiting for W&B process to finish, PID 19525
============================================================================= ERRORS =============================================================================
_______________________________________________ ERROR at teardown of TrainerIntegrationTest.test_trainer_eval_mrpc _______________________________________________
[gw0] linux -- Python 3.7.5 /home/stas/anaconda3/envs/main/bin/python
self = <contextlib._GeneratorContextManager object at 0x7efc47610690>, type = None, value = None, traceback = None
def __exit__(self, type, value, traceback):
if type is None:
try:
> next(self.gen)
E OSError: [Errno 29] Illegal seek
/home/stas/anaconda3/envs/main/lib/python3.7/contextlib.py:119: OSError
============================================================================ FAILURES ============================================================================
_________________________________________________________ TrainerIntegrationTest.test_trainer_eval_mrpc __________________________________________________________
[gw0] linux -- Python 3.7.5 /home/stas/anaconda3/envs/main/bin/python
self = <contextlib._GeneratorContextManager object at 0x7efc476bbd50>, type = None, value = None, traceback = None
def __exit__(self, type, value, traceback):
if type is None:
try:
> next(self.gen)
E OSError: [Errno 29] Illegal seek
/home/stas/anaconda3/envs/main/lib/python3.7/contextlib.py:119: OSError
==================================================================== short test summary info =====================================================================
ERROR tests/test_trainer.py::TrainerIntegrationTest::test_trainer_eval_mrpc - OSError: [Errno 29] Illegal seek
FAILED tests/test_trainer.py::TrainerIntegrationTest::test_trainer_eval_mrpc - OSError: [Errno 29] Illegal seek
```
This is a possible solution, based on this solution
https://github.com/wandb/client/issues/1138#issuecomment-654943065 | 07-17-2020 05:25:30 | 07-17-2020 05:25:30 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5835?src=pr&el=h1) Report
> Merging [#5835](https://codecov.io/gh/huggingface/transformers/pull/5835?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/0735def8e1200ed45a2c33a075bc1595b12ef56a&el=desc) will **decrease** coverage by `1.65%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5835?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5835 +/- ##
==========================================
- Coverage 80.08% 78.43% -1.66%
==========================================
Files 153 146 -7
Lines 27984 26012 -1972
==========================================
- Hits 22412 20402 -2010
- Misses 5572 5610 +38
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5835?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5835/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `43.98% <0.00%> (-46.86%)` | :arrow_down: |
| [src/transformers/tokenization\_reformer.py](https://codecov.io/gh/huggingface/transformers/pull/5835/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fcmVmb3JtZXIucHk=) | `41.66% <0.00%> (-40.00%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5835/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-32.51%)` | :arrow_down: |
| [src/transformers/trainer\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5835/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyX3V0aWxzLnB5) | `85.71% <0.00%> (-14.29%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_albert.py](https://codecov.io/gh/huggingface/transformers/pull/5835/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9hbGJlcnQucHk=) | `76.47% <0.00%> (-11.93%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_xlm.py](https://codecov.io/gh/huggingface/transformers/pull/5835/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl94bG0ucHk=) | `80.98% <0.00%> (-10.18%)` | :arrow_down: |
| [src/transformers/modeling\_reformer.py](https://codecov.io/gh/huggingface/transformers/pull/5835/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19yZWZvcm1lci5weQ==) | `87.70% <0.00%> (-7.98%)` | :arrow_down: |
| [src/transformers/benchmark/benchmark.py](https://codecov.io/gh/huggingface/transformers/pull/5835/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9iZW5jaG1hcmsvYmVuY2htYXJrLnB5) | `74.01% <0.00%> (-7.88%)` | :arrow_down: |
| [src/transformers/tokenization\_marian.py](https://codecov.io/gh/huggingface/transformers/pull/5835/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fbWFyaWFuLnB5) | `93.80% <0.00%> (-5.37%)` | :arrow_down: |
| [src/transformers/modeling\_auto.py](https://codecov.io/gh/huggingface/transformers/pull/5835/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19hdXRvLnB5) | `74.41% <0.00%> (-4.20%)` | :arrow_down: |
| ... and [81 more](https://codecov.io/gh/huggingface/transformers/pull/5835/diff?src=pr&el=tree-more) | |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5835?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5835?src=pr&el=footer). Last update [0735def...9d566b1](https://codecov.io/gh/huggingface/transformers/pull/5835?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>CI doesn't have wandb installed.<|||||>I get an identical issue with wandb shutdown in other places as well, e.g. when running:
```pytest examples/test_examples.py::ExamplesTests::test_run_glue```
So far the solution seems to be to disable wandb. This works:
```
WANDB_DISABLED=true pytest examples/test_examples.py::ExamplesTests::test_run_glue
```
Unless wandb integration is being tested perhaps disabling it globally will solve this everywhere?
<|||||>pinging @borisdayma in case
<|||||>The wandb cli is being updated and will have a noop mode that we can use for easier testing.
For now `WANDB_DISABLED=true` should work.<|||||>> The wandb cli is being updated and will have a noop mode that we can use for easier testing.
would it be possible to ping this issue when it happened, so that we could resolve this? Thank you very much, @borisdayma <|||||>Sounds good, I'll keep an eye on it!<|||||>Seems to have been resolved upstream (wandb) - I can no longer reproduce this problem, closing |
transformers | 5,834 | closed | Trainer support for iterabledataset | To fix --> #5829 | 07-17-2020 05:21:33 | 07-17-2020 05:21:33 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5834?src=pr&el=h1) Report
> Merging [#5834](https://codecov.io/gh/huggingface/transformers/pull/5834?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/d088d744adb4e5aa45262a34acab3ae9e81de169&el=desc) will **increase** coverage by `0.01%`.
> The diff coverage is `40.00%`.
[](https://codecov.io/gh/huggingface/transformers/pull/5834?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5834 +/- ##
==========================================
+ Coverage 78.10% 78.11% +0.01%
==========================================
Files 146 146
Lines 26047 26053 +6
==========================================
+ Hits 20344 20352 +8
+ Misses 5703 5701 -2
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5834?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/5834/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `38.46% <40.00%> (+0.61%)` | :arrow_up: |
| [src/transformers/training\_args.py](https://codecov.io/gh/huggingface/transformers/pull/5834/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmluZ19hcmdzLnB5) | `80.61% <0.00%> (+3.06%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5834?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5834?src=pr&el=footer). Last update [d088d74...edd876a](https://codecov.io/gh/huggingface/transformers/pull/5834?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>I will add a test for it.
<|||||>@sshleifer --> I have made the changes you recommended and added a test that should fail without this change. Please review and let me know if there is anything else needed.<|||||>OK Will make the changes. Also do you know where to place the dataset class to prevent failure in the tf tests?<|||||>I've made the changes. Please review and feel free to edit as needed and merge<|||||>Should be good to go. I removed the WIP. Thanks.<|||||>LGTM but will wait for review from @sgugger too |
transformers | 5,833 | closed | OpenAI GPT NoPaddingTokenFastTokenizerMatchingTest test fails | # 🐛 Bug
## Information
`tests/test_tokenization_fast.py::NoPaddingTokenFastTokenizerMatchingTest::test_all_tokenizers`
fails
## To reproduce
```
pytest -n 1 tests/test_tokenization_fast.py::NoPaddingTokenFastTokenizerMatchingTest::test_all_tokenizers
================================================================ test session starts =================================================================
platform linux -- Python 3.7.5, pytest-5.4.3, py-1.9.0, pluggy-0.13.1
rootdir: /mnt/nvme1/code/huggingface/transformers-master
plugins: hypothesis-5.5.4, filter-subpackage-0.1.1, arraydiff-0.3, flaky-3.6.1, ipynb-1.1.1.dev0, cov-2.10.0, astropy-header-0.1.2, forked-1.2.0, doctestplus-0.5.0, openfiles-0.4.0, remotedata-0.3.2, xdist-1.32.0
gw0 [1]
F [100%]
====================================================================== FAILURES ======================================================================
____________________________________________ NoPaddingTokenFastTokenizerMatchingTest.test_all_tokenizers _____________________________________________
[gw0] linux -- Python 3.7.5 /home/stas/anaconda3/envs/main/bin/python
self = <tests.test_tokenization_fast.NoPaddingTokenFastTokenizerMatchingTest testMethod=test_all_tokenizers>
def test_all_tokenizers(self):
for tok_case in self.TOKENIZERS_CLASSES:
for pretrained_name in tok_case.python_cls.pretrained_vocab_files_map[tok_case.vocab_key].keys():
# Tokenizer.filter makes it possible to filter which Tokenizer to case based on all the
# information available in Tokenizer (name, rust class, python class, vocab key name)
if tok_case.filter is None or (
tok_case.filter is not None and tok_case.filter(tok_case, pretrained_name)
):
kwargs = dict(t for t in tok_case.kwargs) if tok_case.kwargs else {}
with self.subTest("{} ({})".format(tok_case.name, pretrained_name)):
tokenizer_r = tok_case.rust_cls.from_pretrained(pretrained_name, **kwargs)
tokenizer_p = tok_case.python_cls.from_pretrained(pretrained_name, **kwargs)
> self.fast_align_python(tokenizer_r, tokenizer_p, tok_case, pretrained_name)
tests/test_tokenization_fast.py:62:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
tests/test_tokenization_fast.py:836: in fast_align_python
self.assert_tokenization_python_rust_equals(tokenizer_r, tokenizer_p)
tests/test_tokenization_fast.py:192: in assert_tokenization_python_rust_equals
self.assertSequenceEqual(input_p[key], input_r[key])
E AssertionError: Sequences differ: [616,[111 chars] 0, 40477, 4830, 994, 580, 566, 260, 5958, 260[5290 chars] 239] != [616,[111 chars] 0, 4830, 994, 580, 566, 260, 5958, 260, 1490,[5160 chars] 239]
E
E First differing element 29:
E 40477
E 4830
E
E First sequence contains 12 additional elements.
E First extra element 973:
E 1832
E
E Diff is 8701 characters long. Set self.maxDiff to None to see it.
----------------------------------------------------------------- Captured log call ------------------------------------------------------------------
WARNING transformers.tokenization_utils_base:tokenization_utils_base.py:2086 Token indices sequence length is longer than the specified maximum sequence length for this model (985 > 512). Running this sequence through the model will result in indexing errors
================================================================== warnings summary ==================================================================
/home/stas/anaconda3/envs/main/lib/python3.7/site-packages/tensorflow_core/python/pywrap_tensorflow_internal.py:15
/home/stas/anaconda3/envs/main/lib/python3.7/site-packages/tensorflow_core/python/pywrap_tensorflow_internal.py:15: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
import imp
/home/stas/anaconda3/envs/main/lib/python3.7/site-packages/graphql/type/directives.py:55
/home/stas/anaconda3/envs/main/lib/python3.7/site-packages/graphql/type/directives.py:55: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated since Python 3.3,and in 3.9 it will stop working
assert isinstance(locations, collections.Iterable), 'Must provide locations for directive.'
/home/stas/anaconda3/envs/main/lib/python3.7/site-packages/graphql/type/directives.py:62
/home/stas/anaconda3/envs/main/lib/python3.7/site-packages/graphql/type/directives.py:62: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated since Python 3.3,and in 3.9 it will stop working
assert isinstance(args, collections.Mapping), '{} args must be a dict with argument names as keys.'.format(name)
/home/stas/anaconda3/envs/main/lib/python3.7/site-packages/graphql/type/typemap.py:1
/home/stas/anaconda3/envs/main/lib/python3.7/site-packages/graphql/type/typemap.py:1: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated since Python 3.3,and in 3.9 it will stop working
from collections import OrderedDict, Sequence, defaultdict
-- Docs: https://docs.pytest.org/en/latest/warnings.html
============================================================== short test summary info ===============================================================
FAILED tests/test_tokenization_fast.py::NoPaddingTokenFastTokenizerMatchingTest::test_all_tokenizers - AssertionError: Sequences differ: [616,[111 ...
=========================================================== 1 failed, 4 warnings in 5.44s ============================================================
```
## Expected behavior
not fail ;)
I started digging and it fails only for `OpenAI GPT` tokenizer and passes `GPT2` tests.
If someone could give me a pointer, I'd be happy to investigate more.
I'm not sure whether it's the python or rust implementation that is at fault.
I tested that if I force max_length=512, to remove the warning, the test still fails in the same way.
## Environment info
```
- `transformers` version: 3.0.2
- tokenizers: 0.8.1rc1
- Platform: Linux-4.15.0-109-generic-x86_64-with-debian-buster-sid
- Python version: 3.7.5
- PyTorch version (GPU?): 1.5.1 (True)
- Tensorflow version (GPU?): 2.0.1 (False)
- Using GPU in script?: <fill in>
- Using distributed or parallel set-up in script?: <fill in>
```
| 07-17-2020 04:41:47 | 07-17-2020 04:41:47 | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
<|||||>looks like it has been fixed. |
transformers | 5,832 | closed | [wip] T5 tokenizer should add special tokens | 07-17-2020 04:00:01 | 07-17-2020 04:00:01 | Thanks for taking care of it @sshleifer !<|||||># [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5832?src=pr&el=h1) Report
> Merging [#5832](https://codecov.io/gh/huggingface/transformers/pull/5832?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/615be03f9d961c0c9722fe10e7830e011066772e&el=desc) will **decrease** coverage by `0.18%`.
> The diff coverage is `84.61%`.
[](https://codecov.io/gh/huggingface/transformers/pull/5832?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5832 +/- ##
==========================================
- Coverage 78.66% 78.48% -0.19%
==========================================
Files 146 146
Lines 26200 26213 +13
==========================================
- Hits 20611 20574 -37
- Misses 5589 5639 +50
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5832?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/tokenization\_t5.py](https://codecov.io/gh/huggingface/transformers/pull/5832/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdDUucHk=) | `94.04% <84.61%> (-1.73%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5832/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `26.02% <0.00%> (-69.52%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5832/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.51%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5832/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5832?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5832?src=pr&el=footer). Last update [615be03...04f18d4](https://codecov.io/gh/huggingface/transformers/pull/5832?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,831 | closed | minor doc fixes | correct superclass name and small grammar fixes
then later added a correction in the error message:
It appears to be `BaseTokenizer` from looking at:
`from tokenizers.implementations import BaseTokenizer as BaseTokenizerFast`
and not `Tokenizer` as it currently says.
p.s. Please let me know whether you prefer specific separate PRs or bundled PRs for related things. | 07-17-2020 02:44:05 | 07-17-2020 02:44:05 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5831?src=pr&el=h1) Report
> Merging [#5831](https://codecov.io/gh/huggingface/transformers/pull/5831?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/d088d744adb4e5aa45262a34acab3ae9e81de169&el=desc) will **decrease** coverage by `1.32%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5831?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5831 +/- ##
==========================================
- Coverage 78.10% 76.78% -1.33%
==========================================
Files 146 146
Lines 26047 26047
==========================================
- Hits 20344 20000 -344
- Misses 5703 6047 +344
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5831?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/tokenization\_utils\_fast.py](https://codecov.io/gh/huggingface/transformers/pull/5831/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHNfZmFzdC5weQ==) | `94.20% <ø> (ø)` | |
| [src/transformers/data/data\_collator.py](https://codecov.io/gh/huggingface/transformers/pull/5831/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9kYXRhL2RhdGFfY29sbGF0b3IucHk=) | `19.81% <0.00%> (-79.28%)` | :arrow_down: |
| [...rc/transformers/data/datasets/language\_modeling.py](https://codecov.io/gh/huggingface/transformers/pull/5831/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9kYXRhL2RhdGFzZXRzL2xhbmd1YWdlX21vZGVsaW5nLnB5) | `34.69% <0.00%> (-57.15%)` | :arrow_down: |
| [src/transformers/data/datasets/glue.py](https://codecov.io/gh/huggingface/transformers/pull/5831/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9kYXRhL2RhdGFzZXRzL2dsdWUucHk=) | `50.74% <0.00%> (-35.83%)` | :arrow_down: |
| [src/transformers/trainer\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5831/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyX3V0aWxzLnB5) | `60.00% <0.00%> (-25.72%)` | :arrow_down: |
| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/5831/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `16.51% <0.00%> (-21.34%)` | :arrow_down: |
| [src/transformers/data/processors/glue.py](https://codecov.io/gh/huggingface/transformers/pull/5831/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9kYXRhL3Byb2Nlc3NvcnMvZ2x1ZS5weQ==) | `32.00% <0.00%> (-17.10%)` | :arrow_down: |
| [src/transformers/training\_args.py](https://codecov.io/gh/huggingface/transformers/pull/5831/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmluZ19hcmdzLnB5) | `66.32% <0.00%> (-11.23%)` | :arrow_down: |
| [src/transformers/tokenization\_transfo\_xl.py](https://codecov.io/gh/huggingface/transformers/pull/5831/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdHJhbnNmb194bC5weQ==) | `33.56% <0.00%> (-8.93%)` | :arrow_down: |
| [src/transformers/tokenization\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5831/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fcm9iZXJ0YS5weQ==) | `95.89% <0.00%> (-2.74%)` | :arrow_down: |
| ... and [2 more](https://codecov.io/gh/huggingface/transformers/pull/5831/diff?src=pr&el=tree-more) | |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5831?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5831?src=pr&el=footer). Last update [d088d74...38d6390](https://codecov.io/gh/huggingface/transformers/pull/5831?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>Thanks for the PR. Don't hesitate to tag me on PRs related to doc.
Also note, that I haven't had time to clean up all docstrings yet (here the sphinx syntax is not respected so there is no link to the mentioned classes). I've started with Config and will make my way down to the docs (following the index). |
transformers | 5,830 | closed | [cleanups] make Marian save as Marian | allow `MarianTokenizer.prepare_translation_batch` to ignore kwargs like `src_lang` | 07-17-2020 01:28:30 | 07-17-2020 01:28:30 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5830?src=pr&el=h1) Report
> Merging [#5830](https://codecov.io/gh/huggingface/transformers/pull/5830?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/d088d744adb4e5aa45262a34acab3ae9e81de169&el=desc) will **decrease** coverage by `0.06%`.
> The diff coverage is `100.00%`.
[](https://codecov.io/gh/huggingface/transformers/pull/5830?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5830 +/- ##
==========================================
- Coverage 78.10% 78.04% -0.07%
==========================================
Files 146 146
Lines 26047 26049 +2
==========================================
- Hits 20344 20329 -15
- Misses 5703 5720 +17
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5830?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/\_\_init\_\_.py](https://codecov.io/gh/huggingface/transformers/pull/5830/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9fX2luaXRfXy5weQ==) | `99.24% <100.00%> (ø)` | |
| [src/transformers/modeling\_marian.py](https://codecov.io/gh/huggingface/transformers/pull/5830/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19tYXJpYW4ucHk=) | `90.90% <100.00%> (+2.02%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_t5.py](https://codecov.io/gh/huggingface/transformers/pull/5830/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl90NS5weQ==) | `45.98% <0.00%> (-44.92%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_gpt2.py](https://codecov.io/gh/huggingface/transformers/pull/5830/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9ncHQyLnB5) | `65.42% <0.00%> (-29.91%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5830/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `79.44% <0.00%> (-6.52%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5830/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `86.92% <0.00%> (-1.97%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5830/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `82.31% <0.00%> (+1.28%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5830/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5830/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `95.18% <0.00%> (+74.91%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5830?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5830?src=pr&el=footer). Last update [d088d74...41f5de7](https://codecov.io/gh/huggingface/transformers/pull/5830?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
transformers | 5,829 | closed | ValueError: DataLoader with IterableDataset: expected unspecified sampler option, | # 🐛 Bug
## Information
This is in regards to the `Trainer` object. The `get_train_dataloader` function uses RandomSampler or Distributed sampler, both of which I believe won't work with an [iterable datasets](https://pytorch.org/docs/stable/data.html#iterable-style-datasets) - which are only compatible with InfiniteConstantSampler see [pytorch documentation](https://pytorch.org/docs/stable/_modules/torch/utils/data/dataloader.html#DataLoader)
https://github.com/huggingface/transformers/blob/d088d744adb4e5aa45262a34acab3ae9e81de169/src/transformers/trainer.py#L229-L242
## Steps to reproduce the behavior:
1. Create an Iterable dataset and pass it to trainer object.
Gives the following error:
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<timed eval> in <module>
~/.conda/envs/my_root/lib/python3.6/site-packages/transformers/trainer.py in train(self, model_path)
382 training will resume from the optimizer/scheduler states loaded here.
383 """
--> 384 train_dataloader = self.get_train_dataloader()
385 if self.args.max_steps > 0:
386 t_total = self.args.max_steps
~/.conda/envs/my_root/lib/python3.6/site-packages/transformers/trainer.py in get_train_dataloader(self)
247 sampler=train_sampler,
248 collate_fn=self.data_collator,
--> 249 drop_last=self.args.dataloader_drop_last,
250 )
251
~/.conda/envs/my_root/lib/python3.6/site-packages/torch/utils/data/dataloader.py in __init__(self, dataset, batch_size, shuffle, sampler, batch_sampler, num_workers, collate_fn, pin_memory, drop_last, timeout, worker_init_fn, multiprocessing_context)
177 raise ValueError(
178 "DataLoader with IterableDataset: expected unspecified "
--> 179 "sampler option, but got sampler={}".format(sampler))
180 elif batch_sampler is not None:
181 # See NOTE [ Custom Samplers and IterableDataset ]
ValueError: DataLoader with IterableDataset: expected unspecified sampler option, but got sampler=<torch.utils.data.sampler.RandomSampler object at 0x7fff5803b2b0>
## Expected behavior
Ideally, we should check if the dataset is an iterable dataset and if so, pass no sampler to the dataloader. Or allow the option of passing in our own sampler to the trainer.
## Environment info
- `transformers` version: 3.0.2
- Platform: Linux-4.4.121-92.104-default-x86_64-with-SuSE-12-x86_64
- Python version: 3.6.10
- PyTorch version (GPU?): 1.5.1 (False)
- Tensorflow version (GPU?): not installed (NA)
- Using GPU in script?: NO
- Using distributed or parallel set-up in script?: NO
I can contribute and create a PR to fix this. Let me know. | 07-16-2020 23:03:08 | 07-16-2020 23:03:08 | Thanks for the fix with #5834 :)<|||||>The same issue exists for _get_eval_sampler... I patched the Trainer in the mean time, but it would be good if it were fixed.<|||||>Hi @avercau can you open an issue with your problem? Thank you |
transformers | 5,828 | closed | language tag addition on albert-mongolian | hello!
I just added the language tag to model card. Can you guys check for me? | 07-16-2020 22:10:19 | 07-16-2020 22:10:19 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5828?src=pr&el=h1) Report
> Merging [#5828](https://codecov.io/gh/huggingface/transformers/pull/5828?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/d088d744adb4e5aa45262a34acab3ae9e81de169&el=desc) will **decrease** coverage by `0.83%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5828?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5828 +/- ##
==========================================
- Coverage 78.10% 77.26% -0.84%
==========================================
Files 146 146
Lines 26047 26047
==========================================
- Hits 20344 20126 -218
- Misses 5703 5921 +218
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5828?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5828/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.39%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5828/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `83.95% <0.00%> (-2.01%)` | :arrow_down: |
| [src/transformers/file\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5828/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.49% <0.00%> (+0.29%)` | :arrow_up: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5828/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `82.31% <0.00%> (+1.28%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5828/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `95.18% <0.00%> (+74.91%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5828?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5828?src=pr&el=footer). Last update [d088d74...fa415dc](https://codecov.io/gh/huggingface/transformers/pull/5828?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>Thanks! |
transformers | 5,827 | closed | Reproducibility when using pretrained GPT2 | # 🚀 Feature request
<!-- A clear and concise description of the feature proposal.
Please provide a link to the paper and code in case they exist. -->
Ensure reproducibility when loading pretrained models.
## Motivation
<!-- Please outline the motivation for the proposal. Is your feature request
related to a problem? e.g., I'm always frustrated when [...]. If this is related
to another GitHub issue, please link here too. -->
In order to distribute results in a transparent manner, it is important to ensure reproducibility.
When loading a pretrained model, I cannot get the same weights, probably due to non-initialized weights.
Even when using `set_seed`, I cannot get twice the same models.
## Your contribution
<!-- Is there any way that you could help, e.g. by submitting a PR?
Make sure to read the CONTRIBUTING.MD readme:
https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md -->
I created a [small example](https://colab.research.google.com/gist/borisdayma/efa9ce5e8c7078bf12031b525f21f107/transformers-repeatability.ipynb) to illustrate the issue. | 07-16-2020 20:25:57 | 07-16-2020 20:25:57 | Definitely seems bad. Have you experienced this with models besides GPT2?
I think there might be two issues here:
1) `trainer.utils.set_seed` doesn't work completely: @julien-c
2) I'm not sure whether this warning is expected behavior:
```
Some weights of GPT2LMHeadModel were not initialized from the model checkpoint at gpt2 and are newly initialized: ['h.0.attn.masked_bias', 'h.1.attn.masked_bias', 'h.2.attn.masked_bias', 'h.3.attn.masked_bias', 'h.4.attn.masked_bias', 'h.5.attn.masked_bias', 'h.6.attn.masked_bias', 'h.7.attn.masked_bias', 'h.8.attn.masked_bias', 'h.9.attn.masked_bias', 'h.10.attn.masked_bias', 'h.11.attn.masked_bias', 'lm_head.weight']
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
```<|||||>However, hashes won't be same even for a cloned state dict. Try to run this script
```
from copy import deepcopy
from transformers import GPT2LMHeadModel
import torch
from hashlib import blake2b
def get_hash(file):
return blake2b(open(file, 'rb').read()).hexdigest()
model1 = GPT2LMHeadModel.from_pretrained("gpt2")
sd1 = model1.state_dict()
sd2 = deepcopy(sd1)
with open("sd1", "wb") as f:
torch.save(sd1, f)
with open("sd2", "wb") as f:
torch.save(sd2, f)
assert get_hash("sd1") == get_hash("sd2")
```
I also tried to save just individual weights from state dicts, but in this case hashes were equal. So I'm really confused, maybe
it's an issue with Pytorch saving mechanism.<|||||>@sshleifer - I don't really that this issue is related to the Trainer.
To answer this issue, I don't think there is a bug regarding reproducibility. We have multiple tests that verify:
1) That saving / loading a pretrained GPT2 model produces the exact same results => the model has the exact same weights and logic => the model results are reproducible, see:
https://github.com/huggingface/transformers/blob/9d37c56bab8f7f1f1aa0b65be039516072254e77/tests/test_modeling_common.py#L75
2) That loading a pretrained GPT2 model from the modelhub always produces the same results => ... => the model are reproducible, see: https://github.com/huggingface/transformers/blob/9d37c56bab8f7f1f1aa0b65be039516072254e77/tests/test_modeling_gpt2.py#L352
The second raised issue here are the warnings when loading the pretrained weights. There have been **a lot** of issues about this, e.g.: https://github.com/huggingface/transformers/issues/5800, https://github.com/huggingface/transformers/issues/5348, https://github.com/huggingface/transformers/issues/3553, https://github.com/huggingface/transformers/issues/5814 => These warnings happen because of two reasons:
1) It concerns the embedding matrix, where as the output embeddings are tied to the input embeddings so that the weights don't matter and are therefore not saved
2) It concerns these buffer weights: https://github.com/huggingface/transformers/blob/9d37c56bab8f7f1f1aa0b65be039516072254e77/src/transformers/modeling_gpt2.py#L128 , which are hardcoded values and don't need to be saved either.
Since we are getting a lot of duplicated issues about this it might be worth to spend some time to disable all of them (also pinging @LysandreJik, @sshleifer, and @sgugger here in case you feel like taking a deeper look into the warnings :-)) <|||||>The issue may be more related to transparency than reproducibility.
It would be good to be able to prove which model was used at the start, and to show that 2 models are the same (even more if it's done for audit purposes in a late future).<|||||>Comparing hashes for saved model files does not work, as for models with **same** state dict hashes would differ (at least for GPT2). So maybe it's better to load state dicts into memory and compare tensors for each key.<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,826 | closed | Vocab size mismatch on EncoderDecoder model from_pretrained | # 🐛 Bug
## Information
Model I am using (Bert, XLNet ...): EncoderDecoder with bert-base-multilingual-cased in both
Language I am using the model on (English, Chinese ...): not relevant for the bug
The problem arises when using:
* [ ] the official example scripts: (give details below)
* [x] my own modified scripts: (give details below)
The tasks I am working on is: Not relevant for the bug
## To reproduce
Steps to reproduce the behavior:
I am trying to load a training checkpoint using the `save_pretrained` and `from_pretrained` API with the EncoderDecoder model. `EncoderDecoderModel.from_pretrained` fails to load the model when the configuration is loaded from the previously checkpointed model. I believe it's because it is loading a default vocab size (30522) instead of whatever is defined in the saved config (119547) in my case. To repro this run:
```
from transformers import EncoderDecoderModel, BertTokenizer, BertConfig, EncoderDecoderConfig
# Loading encoder-decoder model and saving it
load_dir = 'bert-base-multilingual-cased'
encoder_config = BertConfig.from_pretrained(load_dir)
decoder_config = BertConfig.from_pretrained(load_dir, is_decoder=True)
print(encoder_config.vocab_size)
print(encoder_config.vocab_size)
tokenizer = BertTokenizer.from_pretrained(load_dir)
model = EncoderDecoderModel.from_encoder_decoder_pretrained(load_dir, load_dir, encoder_config=encoder_config, decoder_config=decoder_config) # initialize Bert2Bert
model.save_pretrained('ok')
# Loading saved model and its configuration
encoder_config = BertConfig.from_pretrained('ok')
decoder_config = BertConfig.from_pretrained('ok')
print(encoder_config.vocab_size)
print(encoder_config.vocab_size)
encoder_decoder_config = EncoderDecoderConfig.from_encoder_decoder_configs(encoder_config, decoder_config)
model2 = EncoderDecoderModel.from_pretrained('ok', config=encoder_decoder_config) # This throws
```
The exception is the following:
```
File "/home/ancruzsa/.local/lib/python3.6/site-packages/transformers/modeling_utils.py", line 781, in from_pretrained
model.__class__.__name__, "\n\t".join(error_msgs)
RuntimeError: Error(s) in loading state_dict for EncoderDecoderModel:
size mismatch for encoder.embeddings.word_embeddings.weight: copying a param with shape torch.Size([119547, 768]) from checkpoint, the shape in current model is torch.Size([30522, 768]).
size mismatch for decoder.bert.embeddings.word_embeddings.weight: copying a param with shape torch.Size([119547, 768]) from checkpoint, the shape in current model is torch.Size([30522, 768]).
size mismatch for decoder.cls.predictions.bias: copying a param with shape torch.Size([119547]) from checkpoint, the shape in current model is torch.Size([30522]).
size mismatch for decoder.cls.predictions.decoder.weight: copying a param with shape torch.Size([119547, 768]) from checkpoint, the shape in current model is torch.Size([30522, 768]).
size mismatch for decoder.cls.predictions.decoder.bias: copying a param with shape torch.Size([119547]) from checkpoint, the shape in current model is torch.Size([30522]).
```
## Expected behavior
`from_pretrained(path)` should load the model without issues and using the provided configuration.
Edit: I was expecting `from_pretrained` with a single path as argument to work as explained in [#4595 comment](https://github.com/huggingface/transformers/issues/4595#issuecomment-638077144). However, it seems like doing `EncoderDecoderModel.from_encoder_decoder_pretrained('ok', 'ok', encoder_config=encoder_config, decoder_config=decoder_config)` does not throw an exception but it gives different results in text generation compared to `EncoderDecoderModel.from_pretrained(path)`. It would be great to confirm if both are supported and load the model weights correctly.
## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 3.0.2
- Platform: Linux
- Python version: 3.6.9
- PyTorch version (GPU?): 1.5.0 / Yes with GPU
- Tensorflow version (GPU?): None
- Using GPU in script?: Yes
- Using distributed or parallel set-up in script?: No
| 07-16-2020 20:22:06 | 07-16-2020 20:22:06 | Hey @afcruzs,
These lines are the problem I think:
```python
# Loading saved model and its configuration
encoder_config = BertConfig.from_pretrained('ok')
decoder_config = BertConfig.from_pretrained('ok')
print(encoder_config.vocab_size)
print(encoder_config.vocab_size)
encoder_decoder_config = EncoderDecoderConfig.from_encoder_decoder_configs(encoder_config, decoder_config)
model2 = EncoderDecoderModel.from_pretrained('ok', config=encoder_decoder_config) # This throws
```
If you replace these lines with
```python
# Loading saved model and its configuration
encoder_decoder_config = EncoderDecoderConfig.from_pretrained("ok")
model2 = EncoderDecoderModel.from_pretrained('ok', config=encoder_decoder_config)
```
no error should be thrown.
This line here:
```python
encoder_config = BertConfig.from_pretrained('ok')
```
saves a EncoderDecoderConfig as a Bert Encoder config which should not be done IMO.<|||||>Thanks @patrickvonplaten! that is indeed much clearer. My actual use case is to load the hf pretrained module with possibly modifying the config, saving with save_pretrained, and then later loading with from_pretrained. So this is my final code:
```
load_dir = 'bert-base-multilingual-cased'
encoder_config = BertConfig.from_pretrained(load_dir)
decoder_config = BertConfig.from_pretrained(load_dir, is_decoder=True)
model = EncoderDecoderModel.from_encoder_decoder_pretrained(load_dir, load_dir, encoder_config=encoder_config, decoder_config=decoder_config)
# Train for some time...
# Save model!
model.save_pretrained('ok')
# Loading saved model and its configuration
encoder_decoder_config = EncoderDecoderConfig.from_pretrained("ok")
model2 = EncoderDecoderModel.from_pretrained('ok', config=encoder_decoder_config)
```
I think it would be a good idea to add similar examples in the docs for clarity. Specially for `EncoderDecoderConfig.from_pretrained("ok")` and `.from_pretrained(load_dir, is_decoder=True)` since as you pointed out, doing so carelessly can lead to load the decoder config as encoder. I'm happy to help with the examples if you agree with them!<|||||>Hey @afcruzs,
I agree very much that the `EncoderDecoderModel` should have better documentation.
My plan was to release a notebook soon that explains in detail how to use the `EncoderDecoderModel` and then also to update the docs.
I won't be able to start with this until 3/08 so feel free to open A PR :-) |
transformers | 5,825 | closed | Add inference widget examples | 07-16-2020 20:00:19 | 07-16-2020 20:00:19 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5825?src=pr&el=h1) Report
> Merging [#5825](https://codecov.io/gh/huggingface/transformers/pull/5825?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/d088d744adb4e5aa45262a34acab3ae9e81de169&el=desc) will **increase** coverage by `0.39%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5825?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5825 +/- ##
==========================================
+ Coverage 78.10% 78.50% +0.39%
==========================================
Files 146 146
Lines 26047 26047
==========================================
+ Hits 20344 20448 +104
+ Misses 5703 5599 -104
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5825?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5825/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `43.98% <0.00%> (-49.38%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5825/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (+0.25%)` | :arrow_up: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5825/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `82.31% <0.00%> (+1.28%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5825/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `95.18% <0.00%> (+74.91%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5825?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5825?src=pr&el=footer). Last update [d088d74...682478c](https://codecov.io/gh/huggingface/transformers/pull/5825?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,824 | closed | New Community NB Add | Signed-off-by: lordtt13 <[email protected]> | 07-16-2020 19:03:05 | 07-16-2020 19:03:05 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5824?src=pr&el=h1) Report
> Merging [#5824](https://codecov.io/gh/huggingface/transformers/pull/5824?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/283500ff9f9041fc027a2ce54b4e1f6337c5abbb&el=desc) will **decrease** coverage by `0.43%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5824?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5824 +/- ##
==========================================
- Coverage 78.50% 78.07% -0.44%
==========================================
Files 146 146
Lines 26047 26047
==========================================
- Hits 20448 20335 -113
- Misses 5599 5712 +113
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5824?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5824/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5824/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `83.70% <0.00%> (-2.51%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5824/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5824/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5824?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5824?src=pr&el=footer). Last update [283500f...811bc81](https://codecov.io/gh/huggingface/transformers/pull/5824?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
transformers | 5,823 | closed | Add model card for dv-wave | Dhivehi / Maldives language model - includes links to notebooks for training and comparing performance on a given news classification task | 07-16-2020 19:02:23 | 07-16-2020 19:02:23 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5823?src=pr&el=h1) Report
> Merging [#5823](https://codecov.io/gh/huggingface/transformers/pull/5823?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/c45d7a707d56096b7ed48deaa3ba186fd7c306d4&el=desc) will **increase** coverage by `0.34%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5823?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5823 +/- ##
==========================================
+ Coverage 78.16% 78.50% +0.34%
==========================================
Files 146 146
Lines 26047 26047
==========================================
+ Hits 20359 20448 +89
+ Misses 5688 5599 -89
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5823?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5823/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |
| [src/transformers/generation\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5823/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3V0aWxzLnB5) | `96.82% <0.00%> (-0.29%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5823/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `82.31% <0.00%> (+1.28%)` | :arrow_up: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5823/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (+2.25%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5823/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `95.18% <0.00%> (+74.91%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5823?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5823?src=pr&el=footer). Last update [c45d7a7...46073af](https://codecov.io/gh/huggingface/transformers/pull/5823?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>Great! Link to Dhivehi for those who'd like to learn more about this language: https://en.wikipedia.org/wiki/Maldivian_language |
transformers | 5,822 | closed | [seq2seq] test that finetune takes < 7 seconds | Could also check memory and loss.
| 07-16-2020 18:19:26 | 07-16-2020 18:19:26 | |
transformers | 5,821 | closed | Create README.md | 07-16-2020 18:03:42 | 07-16-2020 18:03:42 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5821?src=pr&el=h1) Report
> Merging [#5821](https://codecov.io/gh/huggingface/transformers/pull/5821?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/c45d7a707d56096b7ed48deaa3ba186fd7c306d4&el=desc) will **decrease** coverage by `0.85%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5821?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5821 +/- ##
==========================================
- Coverage 78.16% 77.30% -0.86%
==========================================
Files 146 146
Lines 26047 26047
==========================================
- Hits 20359 20135 -224
- Misses 5688 5912 +224
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5821?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5821/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.39%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5821/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |
| [src/transformers/generation\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5821/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3V0aWxzLnB5) | `96.82% <0.00%> (-0.29%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5821/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `82.31% <0.00%> (+1.28%)` | :arrow_up: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5821/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.46% <0.00%> (+2.50%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5821/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5821/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `95.18% <0.00%> (+74.91%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5821?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5821?src=pr&el=footer). Last update [c45d7a7...397ba6e](https://codecov.io/gh/huggingface/transformers/pull/5821?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,820 | closed | Migrate Marian models names to ISO-639-3 where possible | Proposed renamings:
ROMANCE-> roa
CELTIC -> cel
NORWAY -> no (check for conflicts w existing bilingual norwegian models.)
remove ch_group
@julien-c for NORTH_EU, why is the code ["multilingual"](https://github.com/huggingface/moon-landing/blob/d50efc4725a7546fb8159626470f3babacd44911/server/lib/ModelInfo.ts#L134-L170)
Should it be the same for all groups that do not map cleanly to an iso-639-3 group?
I will do this via mv, not cp, unless anyone objects.
cc @julien-c
Unresloved:
how will people know if a model requires a target language code?
we will still have many cases where model name is not a language code. | 07-16-2020 17:58:35 | 07-16-2020 17:58:35 | > for NORTH_EU, why is the code "multilingual"
Couldn't find a better qualifier and I think there were a lot of languages
> how will people know if a model requires a target language code?
How do they know now? Maybe add it to the model card? or you could detect if `code.includes('_')`?
> we will still have many cases where model name is not a language code.
Not that many, right?<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,819 | closed | [seq2seq] pack_dataset.py rewrites dataset in max_tokens format | Inspired by fairseq max_tokens_dataset.
This works on disk, by regenerating the data directory, so it can work with or without @Pradhy729 's PR.
It is especially useful for avoiding padding computation in a multigpu setup, where SortishSampler doesn't work.
It will also play nice with TPU by allowing us to always of the same `input_ids` shape, rather than trimming batches to avoid padding computation.
Empirically, it reduces epoch time for mbart finetuning on 8GPU from 2.5hrs to 1.5 hrs. (You can increase batch size to 8 from 4), by cutting the number of examples by a factor of 2, and spending a bit more time on each example.
It also doesn't lead to that much more truncation than the previous approach, although I haven't quantified this.
### Example
```
wget https://s3.amazonaws.com/datasets.huggingface.co/translation/wmt_en_ro_128.tgz
tar -czvf wmt_en_ro_128.tgz
```
Creates a directory called `wmt_en_ro_packed/` with fewer examples
### Usage
```bash
python pack_dataset.py --tok_name facebook/mbart-large-en-ro --max_seq_len 128 --data_dir wmt_en_ro --save_path wmt_en_ro_packed_128
```
Output:
```bash
100%|██████████████████████████████████████████| 1999/1999 [00:01<00:00, 1533.19it/s]
packed val split from 1999 examples -> 683.
100%|██████████████████████████████████████████| 1999/1999 [00:01<00:00, 1590.80it/s]
packed test split from 1999 examples -> 652.
``` | 07-16-2020 17:18:16 | 07-16-2020 17:18:16 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5819?src=pr&el=h1) Report
> Merging [#5819](https://codecov.io/gh/huggingface/transformers/pull/5819?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/63761614eb7a5f96e089b77d5469ea31a7177e16&el=desc) will **decrease** coverage by `1.62%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5819?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5819 +/- ##
==========================================
- Coverage 78.59% 76.96% -1.63%
==========================================
Files 146 146
Lines 26047 26047
==========================================
- Hits 20471 20047 -424
- Misses 5576 6000 +424
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5819?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/data/data\_collator.py](https://codecov.io/gh/huggingface/transformers/pull/5819/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9kYXRhL2RhdGFfY29sbGF0b3IucHk=) | `19.81% <0.00%> (-79.28%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5819/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |
| [...rc/transformers/data/datasets/language\_modeling.py](https://codecov.io/gh/huggingface/transformers/pull/5819/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9kYXRhL2RhdGFzZXRzL2xhbmd1YWdlX21vZGVsaW5nLnB5) | `34.69% <0.00%> (-57.15%)` | :arrow_down: |
| [src/transformers/data/datasets/glue.py](https://codecov.io/gh/huggingface/transformers/pull/5819/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9kYXRhL2RhdGFzZXRzL2dsdWUucHk=) | `50.74% <0.00%> (-35.83%)` | :arrow_down: |
| [src/transformers/trainer\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5819/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyX3V0aWxzLnB5) | `60.00% <0.00%> (-25.72%)` | :arrow_down: |
| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/5819/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `16.51% <0.00%> (-21.34%)` | :arrow_down: |
| [src/transformers/data/processors/glue.py](https://codecov.io/gh/huggingface/transformers/pull/5819/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9kYXRhL3Byb2Nlc3NvcnMvZ2x1ZS5weQ==) | `32.00% <0.00%> (-17.10%)` | :arrow_down: |
| [src/transformers/training\_args.py](https://codecov.io/gh/huggingface/transformers/pull/5819/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmluZ19hcmdzLnB5) | `66.32% <0.00%> (-11.23%)` | :arrow_down: |
| [src/transformers/tokenization\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5819/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fcm9iZXJ0YS5weQ==) | `95.89% <0.00%> (-2.74%)` | :arrow_down: |
| [src/transformers/tokenization\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5819/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25feGxuZXQucHk=) | `88.28% <0.00%> (-1.81%)` | :arrow_down: |
| ... and [5 more](https://codecov.io/gh/huggingface/transformers/pull/5819/diff?src=pr&el=tree-more) | |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5819?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5819?src=pr&el=footer). Last update [6376161...144b667](https://codecov.io/gh/huggingface/transformers/pull/5819?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
transformers | 5,818 | closed | [seq2seq] Don't copy self.source in sortishsampler | 07-16-2020 16:39:09 | 07-16-2020 16:39:09 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5818?src=pr&el=h1) Report
> Merging [#5818](https://codecov.io/gh/huggingface/transformers/pull/5818?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/63761614eb7a5f96e089b77d5469ea31a7177e16&el=desc) will **decrease** coverage by `1.35%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5818?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5818 +/- ##
==========================================
- Coverage 78.59% 77.24% -1.36%
==========================================
Files 146 146
Lines 26047 26047
==========================================
- Hits 20471 20119 -352
- Misses 5576 5928 +352
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5818?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5818/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.39%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5818/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |
| [src/transformers/file\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5818/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.19% <0.00%> (-0.30%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5818/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (+0.50%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_gpt2.py](https://codecov.io/gh/huggingface/transformers/pull/5818/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9ncHQyLnB5) | `95.32% <0.00%> (+31.77%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5818/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5818?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5818?src=pr&el=footer). Last update [6376161...f3a250e](https://codecov.io/gh/huggingface/transformers/pull/5818?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>I don't think you touch this file but, cc @nateraw |
|
transformers | 5,817 | closed | error with transformers 2.9.1 but not with 2.3.0, same code, why? | # 🐛 Bug
## Information
Model I am using (Bert, XLNet ...): Bert
Language I am using the model on (English, Chinese ...): Chinese
The problem arises when using:
* [ ] the official example scripts: (give details below)
* [x] my own modified scripts: (give details below)
The tasks I am working on is:
* [ ] an official GLUE/SQUaD task: (give the name)
* [x] my own task or dataset: (give details below)
## To reproduce
Steps to reproduce the behavior:
My problem is `return encoded` Error with transformers-2.9.1! but with transformers-2.3.0 was fine! Which way is Right? Very Confused!
My code as following:
```
# -*- coding: utf-8 -*-
import torch
from transformers import BertTokenizer, BertModel
from torch.utils.data import Dataset, DataLoader
from functools import partial
import logging
logging.basicConfig(level=logging.INFO)
bert_path = "/Users/kiwi/Desktop/chinese_wwm_ext"
model = BertModel.from_pretrained(bert_path)
tokenizer = BertTokenizer.from_pretrained(bert_path)
def tok_collate(batch_data):
batch_sentence = [x[0] for x in batch_data]
encoded = tokenizer.batch_encode_plus(
batch_sentence,
add_special_tokens=True,
return_tensors='pt',
pad_to_max_length=True)
#return encoded['input_ids'], encoded['token_type_ids'], encoded['attention_mask']
# (tensor([[ 101, 704, 1066, 704, 1925, 2600, 741, 6381, 510, 1744, 2157, 712,
# 2375, 510, 704, 1925, 1092, 1999, 712, 2375, 739, 6818, 2398, 8108,
# 3189, 683, 7305, 6626, 3959, 1266, 4689, 3636, 3727, 2356, 5440, 2175,
# 4554, 2658, 7344, 2971, 2339, 868, 102]]), tensor([[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
# 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]), tensor([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
# 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]]))
return encoded
#
# File "/Users/kiwi/anaconda/python.app/Contents/lib/python3.6/site-packages/transformers/tokenization_utils.py", line 203, in __getattr__
# return self.data[item]
# KeyError: '__getstate__'
def data_loader(data):
dl = DataLoader(data, batch_size=8, shuffle=False, collate_fn=partial(tok_collate),
num_workers=2)
for batch_data in dl:
print(batch_data)
data = [('中共中央总书记、国家主席、中央军委主席习近平10日专门赴湖北省武汉市考察疫情防控工作', 1)]
data_loader(data)
```
<!-- If you have code snippets, error messages, stack traces please provide them here as well.
Important! Use code tags to correctly format your code. See https://help.github.com/en/github/writing-on-github/creating-and-highlighting-code-blocks#syntax-highlighting
Do not use screenshots, as they are hard to read and (more importantly) don't allow others to copy-and-paste your code.-->
## Expected behavior
<!-- A clear and concise description of what you would expect to happen. -->
```
encoded = {'input_ids': tensor([[ 101, 704, 1066, 704, 1925, 2600, 741, 6381, 510, 1744, 2157, 712,
2375, 510, 704, 1925, 1092, 1999, 712, 2375, 739, 6818, 2398, 8108,
3189, 683, 7305, 6626, 3959, 1266, 4689, 3636, 3727, 2356, 5440, 2175,
4554, 2658, 7344, 2971, 2339, 868, 102]]), 'token_type_ids': tensor([[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]), 'attention_mask': tensor([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]])}
```
## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version:2.9.1
- Platform:
- Python version:3.6+
- PyTorch version (GPU?):
- Tensorflow version (GPU?):
- Using GPU in script?:
- Using distributed or parallel set-up in script?:
| 07-16-2020 16:24:46 | 07-16-2020 16:24:46 | Hi! Indeed, I can reproduce. This seems to be an edge-case we were not testing for in late v2+ versions. It does run correctly on v2.3.0, as you've shown, and on recent versions (v3+) as well.
After looking a bit deeper into it, it seems to have happened because of the introduction of `BatchEncoding`, in version v2.9.0. It was later patched in v3.0.0 so the blacklisted versions for using a parallelization mechanism (here the dataloader) with `batch_encode_plus` would be the versions between `v2.9.0` and `v3.0.0`. That would be `v2.9.x`, `v2.10.x` and `v2.11.x`.
Hope this helps, and sorry for the inconvenience. |
transformers | 5,816 | closed | Additional layers to BERT | # ❓ Questions & Help
<!-- The GitHub issue tracker is primarly intended for bugs, feature requests,
new models and benchmarks, and migration questions. For all other questions,
we direct you to the Hugging Face forum: https://discuss.huggingface.co/ .
You can also try Stack Overflow (SO) where a whole community of PyTorch and
Tensorflow enthusiast can help you out. In this case, make sure to tag your
question with the right deep learning framework as well as the
huggingface-transformers tag:
https://stackoverflow.com/questions/tagged/huggingface-transformers
-->
## Details
<!-- Description of your issue -->
I'm currently fine-tuning BERTForSequenceClassification model for a classification task and I wanted to know if there are ways to add additional layers before the final classification layer?
<!-- You should first ask your question on the forum or SO, and only if
you didn't get an answer ask it here on GitHub. -->
| 07-16-2020 16:05:30 | 07-16-2020 16:05:30 | Hi @psureshmagadi17 , you can add additional layers easily, take a loot the source code for `BERTForSequenceClassification`, you can take that code as it is and add the additional layers before the final classifier. <|||||>> Hi @psureshmagadi17 , you can add additional layers easily, take a loot the source code for `BERTForSequenceClassification`, you can take that code as it is and add the additional layers before the final classifier.
Hi @patil-suraj , thank you for your response. Did you mean that we can just alter the code in main class? If yes, do you have an example? <|||||>Hi @psureshmagadi17, if your goal is to add layers to a pretrained model only for fine-tuning BERTForSequenceClassification I think the best option is to modify the BertForSequenceClassification Module.
If you want to add attention layers, make sure to use the sequence_output of the BertModel Module and not the pooled_output in the forward function, then use a BertPooler layer before the classifier.<|||||>> Hi @psureshmagadi17, if your goal is to add layers to a pretrained model only for fine-tuning BERTForSequenceClassification I think the best option is to modify the BertForSequenceClassification Module.
>
> If you want to add attention layers, make sure to use the sequence_output of the BertModel Module and not the pooled_output in the forward function, then use a BertPooler layer before the classifier.
Hi @nassim-yagoub - thank you for the response! I'm fairly new to this process i.e., modify the network structure. Do you have an example or discussion that I can follow to help me through this process? <|||||>A small example:
import torch.nn as nn
from transformers import BertModel
class CustomBERTModel(nn.Module):
def __init__(self):
super(CustomBERTModel, self).__init__()
self.bert = BertModel.from_pretrained("bert-base-uncased")
# add your additional layers here, for example a dropout layer followed by a linear classification head
self.dropout = nn.Dropout(0.3)
self.out = nn.Linear(768, 2)
def forward(self, ids, mask, token_type_ids):
sequence_output, pooled_output = self.bert(
ids,
attention_mask=mask,
token_type_ids=token_type_ids
)
# we apply dropout to the sequence output, tensor has shape (batch_size, sequence_length, 768)
sequence_output = self.dropout(sequence_output)
# next, we apply the linear layer. The linear layer (which applies a linear transformation)
# takes as input the hidden states of all tokens (so seq_len times a vector of size 768, each corresponding to
# a single token in the input sequence) and outputs 2 numbers (scores, or logits) for every token
# so the logits are of shape (batch_size, sequence_length, 2)
logits = self.out(sequence_output)
return logits<|||||>> A small example:
>
> ```
> import torch.nn as nn
> from transformers import BertModel
>
> class CustomBERTModel(nn.Module):
> def __init__(self):
> super(CustomBERTModel, self).__init__()
> self.bert = BertModel.from_pretrained("bert-base-uncased")
> # add your additional layers here, for example a dropout layer followed by a linear classification head
> self.dropout = nn.Dropout(0.3)
> self.out = nn.Linear(768, 2)
>
> def forward(self, ids, mask, token_type_ids):
> sequence_output, pooled_output = self.bert(
> ids,
> attention_mask=mask,
> token_type_ids=token_type_ids
> )
>
> # we apply dropout to the sequence output, tensor has shape (batch_size, sequence_length, 768)
> sequence_output = self.dropout(sequence_output)
>
> # next, we apply the linear layer. The linear layer (which applies a linear transformation)
> # takes as input the hidden states of all tokens (so seq_len times a vector of size 768, each corresponding to
> # a single token in the input sequence) and outputs 2 numbers (scores, or logits) for every token
> # so the logits are of shape (batch_size, sequence_length, 2)
> logits = self.out(sequence_output)
>
> return logits
> ```
Thank you, @NielsRogge<|||||>For example if you want to add the same layers used in Bert, you may want to modify the Module this way (with new_layers_config being the same than the original config, except for the number of layers):
```python
class BertForSequenceClassification(BertPreTrainedModel):
def __init__(self, config, new_layers_config):
super().__init__(config)
self.num_labels = config.num_labels
self.bert = BertModel(config)
self.new_layers = BertEncoder(new_layers_config)
self.pooler = BertPooler(config)
self.dropout = nn.Dropout(config.hidden_dropout_prob)
self.classifier = nn.Linear(config.hidden_size, config.num_labels)
self.init_weights()
def forward(
self,
input_ids=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
head_mask=None,
inputs_embeds=None,
labels=None,
output_attentions=None,
output_hidden_states=None,
):
outputs = self.bert(
input_ids,
attention_mask=attention_mask,
token_type_ids=token_type_ids,
position_ids=position_ids,
head_mask=head_mask,
inputs_embeds=inputs_embeds,
output_attentions=output_attentions,
output_hidden_states=output_hidden_states,
)
sequence_output = outputs[0]
new_layers_output = self.new_layers(sequence_output)[0]
pooled_output = self.pooler(new_layers_output)
pooled_output = self.dropout(pooled_output)
logits = self.classifier(pooled_output)
outputs = (logits,) + outputs[2:] # add hidden states and attention if they are here
if labels is not None:
if self.num_labels == 1:
# We are doing regression
loss_fct = MSELoss()
loss = loss_fct(logits.view(-1), labels.view(-1))
else:
loss_fct = CrossEntropyLoss()
loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))
outputs = (loss,) + outputs
return outputs # (loss), logits, (hidden_states), (attentions)
```
We added a BertEncoder and a BertPooler to the base implementation.
You can also retreive the hidden_states and attention of the new layers if you want to, I did not do it here.<|||||>Thanks @nassim-yagoub !<|||||>@nassim-yagoub - I had another question : are the weights for BERTForSequenceClassification Model layers frozen by default?<|||||>The weights are not frozen by default when you load them, however you can manually freeze them with `.requires_grad = False`<|||||>> The weights are not frozen by default when you load them, however you can manually freeze them with `.requires_grad = False`
Thank you @nassim-yagoub!<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
<|||||>> A small example:
>
> ```
> import torch.nn as nn
> from transformers import BertModel
>
> class CustomBERTModel(nn.Module):
> def __init__(self):
> super(CustomBERTModel, self).__init__()
> self.bert = BertModel.from_pretrained("bert-base-uncased")
> # add your additional layers here, for example a dropout layer followed by a linear classification head
> self.dropout = nn.Dropout(0.3)
> self.out = nn.Linear(768, 2)
>
> def forward(self, ids, mask, token_type_ids):
> sequence_output, pooled_output = self.bert(
> ids,
> attention_mask=mask,
> token_type_ids=token_type_ids
> )
>
> # we apply dropout to the sequence output, tensor has shape (batch_size, sequence_length, 768)
> sequence_output = self.dropout(sequence_output)
>
> # next, we apply the linear layer. The linear layer (which applies a linear transformation)
> # takes as input the hidden states of all tokens (so seq_len times a vector of size 768, each corresponding to
> # a single token in the input sequence) and outputs 2 numbers (scores, or logits) for every token
> # so the logits are of shape (batch_size, sequence_length, 2)
> logits = self.out(sequence_output)
>
> return logits
> ```
I wonder is there a way to do this modification while still using the "trainer" just like in the example codes of run_ner.py?
My second question is about the return values of self.bert, so basically "sequence_output" is the context vector for the sentence, but what does pooled_output mean? <|||||>> A small example:
>
> ```
> import torch.nn as nn
> from transformers import BertModel
>
> class CustomBERTModel(nn.Module):
> def __init__(self):
> super(CustomBERTModel, self).__init__()
> self.bert = BertModel.from_pretrained("bert-base-uncased")
> # add your additional layers here, for example a dropout layer followed by a linear classification head
> self.dropout = nn.Dropout(0.3)
> self.out = nn.Linear(768, 2)
>
> def forward(self, ids, mask, token_type_ids):
> sequence_output, pooled_output = self.bert(
> ids,
> attention_mask=mask,
> token_type_ids=token_type_ids
> )
>
> # we apply dropout to the sequence output, tensor has shape (batch_size, sequence_length, 768)
> sequence_output = self.dropout(sequence_output)
>
> # next, we apply the linear layer. The linear layer (which applies a linear transformation)
> # takes as input the hidden states of all tokens (so seq_len times a vector of size 768, each corresponding to
> # a single token in the input sequence) and outputs 2 numbers (scores, or logits) for every token
> # so the logits are of shape (batch_size, sequence_length, 2)
> logits = self.out(sequence_output)
>
> return logits
> ```
If possible it would be great if you can provide the training loop for the same |
transformers | 5,815 | closed | cast_bool_to_primitive breaks TensorFlow graph support. | # 🐛 Bug
## To reproduce
```python
import transformers
bert = tf.function(transformers.TFBertForMaskedLM.from_pretrained('bert-base-uncased'))
for i in range(2):
(_, hidden_state) = bert(tf.constant([[10,11,12]]), output_hidden_states=True)
print(f'computed {i}')
```
Errors with
```
ValueError: not enough values to unpack (expected 2, got 1)
```
## Expected behavior
```
computed 1
computed 2
```
Same result as if `tf.function` was not used.
## Environment info
Example environment : https://colab.research.google.com/gist/AndreasMadsen/593df94a3319dee58bba33a26efedeb3/untitled6.ipynb
- `transformers` version: 3.0.2
- Platform: Linux-4.19.104+-x86_64-with-Ubuntu-18.04-bionic
- Python version: 3.6.9
- PyTorch version (GPU?): 1.5.1+cu101 (False)
- Tensorflow version (GPU?): 2.2.0 (False)
- Using GPU in script?: <fill in>
- Using distributed or parallel set-up in script?: <fill in>
-----
## Details
The bug happens due to `cast_bool_to_primitive`, that was introduced in https://github.com/huggingface/transformers/commit/6e603cb7892b49a2cbbc10ba859759f92c3fb7a6. Before that, it was possible to get the `hidden_states` from Bert in TensorFlow graph/function mode.
Generally speaking, casting TensorFlow tensors to primitives is not a good practice, as it only works in eager mode. It is also completely unnecessary in this case, as using `if bool_tensor_scalar:` works perfectly fine.
```python
def print_bool(x):
if x:
print('True')
else:
print('False')
print_bool_graph = tf.function(print_bool)
print('eager:')
print_bool(True) # Prints True
print_bool(False) # Prints False
print_bool(tf.constant(True)) # Prints True
print_bool(tf.constant(False)) # Prints False
print('')
print('graph:')
print_bool_graph(True) # Prints True
print_bool_graph(False) # Prints False
print_bool_graph(tf.constant(True)) # Prints True
print_bool_graph(tf.constant(False)) # Prints False
```
I can see there are some cases where defaults are used. The right way to handle that is to implement the default handling upstream in the first `call()` method. A lesser way would be to implement it as:
```python
def cast_bool_to_primitive(x, default_value=False):
if x is None:
return default_value
return x
``` | 07-16-2020 16:03:17 | 07-16-2020 16:03:17 | Hey @AndreasMadsen,
Thanks a lot for the detailed error description! I added this function, so I will try to take a look as soon as possible. I will be off for the next two weeks though (but will not this issue down). If by change @jplu you find some time, feel free to take a look :-) <|||||>Should be fixed in the PR https://github.com/huggingface/transformers/pull/5468<|||||>Hi @jplu, I'm sorry, but I doubt #5468 will fix the issue. Fundamentally speaking casting to primitives is not a good practice in TensorFlow, as it invalidates the use of `@tf.function` and is generally unnecessary as described above. Casting to primitives is, in my experience, just never the correct solution in TensorFlow.
I do think #5468 mitigates the issue, which is maybe where the confusion is coming from. This is because, the models will now correctly default to the `config` object when `output_hidden_states=True` **is not** specified as an input. In those cases object property is never cast to a tensor to begin with, therefore the `@tf.function` graph will be statically compiled to always output the `hidden_states`, as intended.
However, the behavior is different when `output_hidden_states=True` is specified as an input, as it will be cast to a Tensor when it becomes part of the `inputs` argument in `call()`. After that, it is not possible to convert back to a primitive, as that invalidates `@tf.function`.
If you insist on keeping it as a primitive, the best solution might be to specify it as an aux-input, similar to `training` and `mask` in a `keras.layers.Layer`, as they don't get converted the same way. I'm not familiar enough with the Keras internals to know the details here, and I think it might also be incompatible with `compute_output_shape` etc.
_BTW, in the keras RNN layers, hidden_state is only specified in the constructor, properly because it can get a bit messy having to specify it in the `inputs`, but I don't see anything fundamentally wrong with specifying it in `inputs`._<|||||>The PR fix your issue at least for the piece of code you are providing, here the result I have:
```
>>> import transformers
>>> import tensorflow as tf
>>> bert = tf.function(transformers.TFBertForMaskedLM.from_pretrained('bert-base-uncased'))
>>> for i in range(2):
... (_, hidden_state) = bert(tf.constant([[10,11,12]]), output_hidden_states=True)
... print(f'computed {i}')
...
computed 0
computed 1
>>>
```<|||||>I believe the solution of @AndreasMadsen is correct -- `output_hidden_states` and also `output_attentions` should be passed as named arguments of the `call` method of `TFBertEncoder.call` -- that way they are not converted to Tensors and can be just constants.<|||||>@jplu Sorry for the misunderstanding. The test now works because `output_hidden_states` is now an auxiliary-input, thus is stays a primitive, thus the casting is no longer involved. However, casting to primitives is still not good practice in TensorFlow, so it doesn't take much to break it.
I read your code more thoroughly, and have the following two failure cases for you.
**Case 1:**
```python
bert = tf.function(transformers.TFBertForMaskedLM.from_pretrained('bert-base-uncased'))
outputs = bert(tf.constant([[10,11,12]]), output_attentions=True)
assert len(outputs) == 2
```
Fails with an IndexError. Essentially because casting in a `tf.function` can not be done. Edit: I think it is an inconsistent casting, in some functions the `output_attentions` is a primitive and the "casting" works by being a no-op, it other functions the casting can't be done as it is in a `tf.function`.
**Case 2:**
```python
bert = tf.function(transformers.TFBertForMaskedLM.from_pretrained('bert-base-uncased'))
outputs = bert(tf.constant([[10,11,12]]), output_hidden_states=tf.constant(True))
assert len(outputs) == 2
```
outputs only one output (two is exected), because `tf.constant` can not be casted in a `tf.function`.
---
However, I would like that, instead of working around these issues, you read the documentation on AutoGraph.
I really think there is a misunderstanding here, about what TensorFlow AutoGraph can do for you and why casting to primitives is really not necessary at all. I would suggest you read https://www.tensorflow.org/guide/function#conditionals and also check out the hidden docs, such as https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/autograph/g3doc/reference/control_flow.md#effects-of-the-tracing-process which explains it in more details.
What @foxik says is true, but I think depending on the auxiliary-inputs just avoids the misunderstanding. Truly, casting to primitives is just not a good idea.<|||||>Ok, thanks for the hints, I will review that part! I should be able to start working on it from tomorrow or Thursday.<|||||>I have started to work on this today. I basically just removed the casting to primitive function and updated the code accordingly. Nevertheless it doesn't work when booleans are tensors (your Case 2) because AutoGraph do not let you return an output of different size on each branch. A simple example to simply the use case can be:
```
import tensorflow as tf
@tf.function
def output(output1, output2):
first_conditional_output = ()
second_conditional_output = ()
for i in range(2):
if output1:
first_conditional_output = first_conditional_output + (i,)
if output2:
second_conditional_output = second_conditional_output + (i,)
outputs = (0,)
if output1:
outputs = outputs + (first_conditional_output,)
if output2:
outputs = outputs + (second_conditional_output,)
return outputs
```
This piece of code works fine as long as the parameters are primitives, but when using `tf.constant(True)` for at least one of the two makes it fail with:
```
/opt/anaconda3/envs/transformers/lib/python3.7/site-packages/tensorflow/python/ops/cond_v2.py:633 _make_indexed_slices_indices_types_match
assert len(set(outs_per_branch)) == 1, outs_per_branch
AssertionError: [1, 0]
```
I currently struggling on this, and try to find a workaround but if it is not possible to fix this, I think we will just skip this functionality.<|||||>@jplu TensorFlow guys internally use `tf.cond` to handle this case, for example here: https://github.com/tensorflow/tensorflow/blob/2b96f3662bd776e277f86997659e61046b56c315/tensorflow/python/framework/smart_cond.py#L27-L59 . Keras layers like Dropout have another special case for `Variables` here (the `smart_module.smart_cond` is the above method): https://github.com/tensorflow/tensorflow/blob/2b96f3662bd776e277f86997659e61046b56c315/tensorflow/python/keras/utils/tf_utils.py#L42-L65<|||||>I already came across this, but I don't see how it could solve this issue. Each if in this piece of code is automatically translated to a `tf.cond`. The problem is that there is no else and then the branches `true_fn` and `false_fn` will have a different output size from each other. This is not allowed and this is the problem :(
Unless I'm missing something in what you said.
I'm still trying to figure out how to find a workaround anyway.<|||||>@jplu I thought that `tf.cond` can return arbitrary results from the `true_fn` and `false_fn` (contrary to Autograph), but you are right, it is not allowed during graph construction -- my bad.
Personally I think it is enough for the `output_hidden_size` to be a Python constant (and fail otherwise). It is in fact not obvious what type should the output have in case computation graph is used and the `output_hidden_size` is undefined Tensor value.
So I think you were right with skipping the functionality ;-)<|||||>@jplu Good point, regarding the variable output-signature. I think it is perfectly acceptable to assert for a primitive input, at least for now.
Alternatively, the solution would be to return an empty tensor when `tf.constant([False])` is used. Such an approach could look like this:
```python
import tensorflow as tf
@tf.function
def output(output1, output2):
first_conditional_output = tf.TensorArray(tf.int64, size=0, dynamic_size=True, clear_after_read=True)
second_conditional_output = tf.TensorArray(tf.int64, size=0, dynamic_size=True, clear_after_read=True)
for i in range(2):
if output1:
first_conditional_output = first_conditional_output.write(i, i)
if output2:
second_conditional_output = second_conditional_output.write(i, i)
outputs = (0,)
if isinstance(output1, tf.Tensor) or output1:
outputs = outputs + (first_conditional_output.stack(),)
if isinstance(output2, tf.Tensor) or output2:
outputs = outputs + (second_conditional_output.stack(),)
return outputs
```
edit: PS: there is more documentation here: https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/autograph/g3doc/reference/index.md<|||||>@AndreasMadsen Thanks!! I basically already came across the same solution (which is I think the only one :smile:), but the issue here is that we will always have three outputs, which is not really what is wanted. I have to talk to the whole team to see if they are ok with this or not, If yes, we will update the code base accordingly otherwise we will skip this.<|||||>I'm trying to figure out if using `tf.boolean_mask` could work.<|||||>Hi @jplu, I will leave it up to you to decide what is "wanted". But you should consider the usage pattern when unpacking the output:
**with always three outputs**:
```python
@tf.function
def usage(hasOutput1, hasOutput2):
(one, output1, output2) = output(hasOutput1, hasOutput2)
tf.print(one)
if hasOutput1:
tf.print(output1)
if hasOutput2:
tf.print(output2)
```
**with always variable outputs**:
```python
@tf.function
def usage(hasOutput1, hasOutput2):
output1 = tf.zeros((0,))
output2 = tf.zeros((0,))
if hasOutput1 and hasOutput2:
(one, output1, output2) = output(hasOutput1, hasOutput2)
elif hasOutput1:
(one, output1) = output(hasOutput1, hasOutput2)
elif hasOutput2:
(one, output2) = output(hasOutput1, hasOutput2)
else:
(one, ) = output(hasOutput1, hasOutput2)
tf.print(one)
if hasOutput1:
tf.print(output1)
if hasOutput2:
tf.print(output2)
```
<|||||>You are totally right @AndreasMadsen!! Now everything should work like expected in the PR at the condition to use primitive booleans. If we all decide to fix the output size to 3 I will open a new PR for this.
Nevertheless, I've just spotted another problem with the usage of `TensorArray`, instead to have a tuple that looks like : `(batch_size, num_tokens, hidden_states_size) * num_layers` we get a tensor that looks like `(num_layers, batch_size, num_tokens, hidden_states_size)` which cause several other issues for later.<|||||>> Nevertheless, I've just spotted another problem with the usage of TensorArray, instead to have a tuple that looks like : (batch_size, num_tokens, hidden_states_size) * num_layers we get a tensor that looks like (num_layers, batch_size, num_tokens, hidden_states_size) which cause several other issues for later.
To avoid a breaking change, you could do it as a tuple of empty tensors. For the next major version, I would suggest it become one big tensor. You can swap `transpose`/swap the first and last axis, to make it mostly compatible, with indexing, unpack, etc..<|||||>Exactly!! This should be a good way to go if the decision of having always 3 outputs is taken.<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,814 | closed | How to download original weights of gpt2 | # ❓ Questions & Help
Hi, I have been trying to reproduce losses from [this paper](https://cdn.openai.com/better-language-models/language_models_are_unsupervised_multitask_learners.pdf), but I realized that my losses are lower than the ones from the paper. Also, I noticed a warning - Some weights of GPT2LMHeadModel were not initialized from the model checkpoint at gpt2-medium and are newly initialized: ... Does it mean that the original weights of GPT-2 was changed? If so, how can I download the original weights?
For instance my PPL for PTB is 35.20 while the PPL in the paper is 65.85. | 07-16-2020 14:50:30 | 07-16-2020 14:50:30 | Hey @nuradilK,
1) The GPT2 weight initialization is a known warning bug we should remove, but which does not affect model performance,
see https://github.com/huggingface/transformers/issues/5800 e.g.
2) It's very hard for us to track down low performance of specific user code (which data to use, data processing, ...). Do you mind posting a question like "How to reproduce GPT2 PPL results" on https://discuss.huggingface.co/ since it is not really a bug but more a high level question. <|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,813 | closed | Create README.md | 07-16-2020 14:06:11 | 07-16-2020 14:06:11 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5813?src=pr&el=h1) Report
> Merging [#5813](https://codecov.io/gh/huggingface/transformers/pull/5813?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/89a78be51f1c2afd263da66ec76d3297432f0c2a&el=desc) will **not change** coverage.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5813?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5813 +/- ##
=======================================
Coverage 78.19% 78.19%
=======================================
Files 146 146
Lines 26047 26047
=======================================
Hits 20367 20367
Misses 5680 5680
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5813?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5813/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.96% <0.00%> (-0.26%)` | :arrow_down: |
| [src/transformers/file\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5813/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.49% <0.00%> (+0.29%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5813?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5813?src=pr&el=footer). Last update [89a78be...1965bb0](https://codecov.io/gh/huggingface/transformers/pull/5813?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
|
transformers | 5,812 | closed | Update README.md | Fix missig "-" in meta data | 07-16-2020 13:59:04 | 07-16-2020 13:59:04 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5812?src=pr&el=h1) Report
> Merging [#5812](https://codecov.io/gh/huggingface/transformers/pull/5812?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/89a78be51f1c2afd263da66ec76d3297432f0c2a&el=desc) will **increase** coverage by `0.31%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5812?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5812 +/- ##
==========================================
+ Coverage 78.19% 78.50% +0.31%
==========================================
Files 146 146
Lines 26047 26047
==========================================
+ Hits 20367 20448 +81
+ Misses 5680 5599 -81
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5812?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5812/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5812/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (ø)` | |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5812/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `82.31% <0.00%> (+1.28%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5812/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `95.18% <0.00%> (+74.91%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5812?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5812?src=pr&el=footer). Last update [89a78be...a5f39ab](https://codecov.io/gh/huggingface/transformers/pull/5812?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
|
transformers | 5,811 | closed | [Longformer] fix longformer slow-down | A drastic slow-down of Longformer after the PR: https://github.com/huggingface/transformers/pull/5219 was deteced by @HHousen (thanks a lot!) here: https://github.com/huggingface/transformers/issues/4406#issuecomment-658483050 .
After digging a bit into the code it can be seen that the line:
`is_global_attn = any(is_index_global_attn.flatten())` is the reason of the drastic slow-down.
Running the following benchmark on master:
```
python examples/benchmarking/run_benchmark.py --models allenai/longformer-base-4096 --no_memory --sequence_length 512 1024
```
yields the following result:
```
==================== INFERENCE - SPEED - RESULT ====================
--------------------------------------------------------------------------------
Model Name Batch Size Seq Length Time in s
--------------------------------------------------------------------------------
allenai/longformer-base-4096 8 512 0.647
allenai/longformer-base-4096 8 1024 1.284
--------------------------------------------------------------------------------
==================== ENVIRONMENT INFORMATION ====================
- transformers_version: 3.0.2
- framework: PyTorch
- use_torchscript: False
- framework_version: 1.5.0
- python_version: 3.7.7
- system: Linux
- cpu: x86_64
- architecture: 64bit
- date: 2020-07-16
- time: 14:07:02.719531
- fp16: False
- use_multiprocessing: True
- only_pretrain_model: False
- cpu_ram_mb: 32089
- use_gpu: True
- num_gpus: 1
- gpu: TITAN RTX
- gpu_ram_mb: 24217
- gpu_power_watts: 280.0
- gpu_performance_state: 0
- use_tpu: False
```
Now on this branch the results are as follows:
```
==================== INFERENCE - SPEED - RESULT ====================
--------------------------------------------------------------------------------
Model Name Batch Size Seq Length Time in s
--------------------------------------------------------------------------------
allenai/longformer-base-4096 8 512 0.141
allenai/longformer-base-4096 8 1024 0.271
--------------------------------------------------------------------------------
==================== ENVIRONMENT INFORMATION ====================
- transformers_version: 3.0.2
- framework: PyTorch
- use_torchscript: False
- framework_version: 1.5.0
- python_version: 3.7.7
- system: Linux
- cpu: x86_64
- architecture: 64bit
- date: 2020-07-16
- time: 14:07:57.623378
- fp16: False
- use_multiprocessing: True
- only_pretrain_model: False
- cpu_ram_mb: 32089
- use_gpu: True
- num_gpus: 1
- gpu: TITAN RTX
- gpu_ram_mb: 24217
- gpu_power_watts: 280.0
- gpu_performance_state: 0
- use_tpu: False
```
So this simple line is responsible for a slow-down of factor 4.
Moral of the story: Never use `any()` on a PyTorch tensor => always use tensor.any(). I wonder if this is actually a known problem of PyTorch/Python. It might be a good to check our code if we have more statements like this.
Another lesson for me is that one should always run the benchmarking before and after doing such a big refactoring as in https://github.com/huggingface/transformers/pull/5219 .
It's very simple to run the benchmark script for a model and takes usually only a couple of seconds. Ideally we should have performance regression tests to automatically detect such slow downs.
Pinging @thomwolf @mfuntowicz @sshleifer @LysandreJik @ibeltagy - think this is quite interesting. | 07-16-2020 13:49:56 | 07-16-2020 13:49:56 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5811?src=pr&el=h1) Report
> Merging [#5811](https://codecov.io/gh/huggingface/transformers/pull/5811?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/89a78be51f1c2afd263da66ec76d3297432f0c2a&el=desc) will **decrease** coverage by `0.80%`.
> The diff coverage is `100.00%`.
[](https://codecov.io/gh/huggingface/transformers/pull/5811?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5811 +/- ##
==========================================
- Coverage 78.19% 77.38% -0.81%
==========================================
Files 146 146
Lines 26047 26047
==========================================
- Hits 20367 20157 -210
- Misses 5680 5890 +210
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5811?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/modeling\_longformer.py](https://codecov.io/gh/huggingface/transformers/pull/5811/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19sb25nZm9ybWVyLnB5) | `89.21% <100.00%> (ø)` | |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5811/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.39%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5811/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (ø)` | |
| [src/transformers/generation\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5811/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3V0aWxzLnB5) | `97.11% <0.00%> (+0.28%)` | :arrow_up: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5811/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `82.31% <0.00%> (+1.28%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5811/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `95.18% <0.00%> (+74.91%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5811?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5811?src=pr&el=footer). Last update [89a78be...25c6419](https://codecov.io/gh/huggingface/transformers/pull/5811?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>Great find. +1 to tests! |
transformers | 5,810 | closed | Add missing arguments for BertWordPieceTokenizer | Hi:)
It seems that in `__init__` of `BertTokenizerFast`, arguments of `pad_token` and `mask_token` of `BertWordPieceTokenizer` are missing.
I've added them in this commit. If these missed arguments are intended, please let me know:)
Thank you!
```python
class BertTokenizerFast(PreTrainedTokenizerFast):
def __init__(
self,
vocab_file,
do_lower_case=True,
unk_token="[UNK]",
sep_token="[SEP]",
pad_token="[PAD]",
cls_token="[CLS]",
mask_token="[MASK]",
clean_text=True,
tokenize_chinese_chars=True,
strip_accents=None,
wordpieces_prefix="##",
**kwargs
):
super().__init__(
BertWordPieceTokenizer(
vocab_file=vocab_file,
unk_token=unk_token,
sep_token=sep_token,
cls_token=cls_token,
clean_text=clean_text,
handle_chinese_chars=tokenize_chinese_chars,
strip_accents=strip_accents,
lowercase=do_lower_case,
wordpieces_prefix=wordpieces_prefix,
),
unk_token=unk_token,
sep_token=sep_token,
pad_token=pad_token,
cls_token=cls_token,
mask_token=mask_token,
**kwargs,
)
``` | 07-16-2020 13:46:30 | 07-16-2020 13:46:30 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5810?src=pr&el=h1) Report
> Merging [#5810](https://codecov.io/gh/huggingface/transformers/pull/5810?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/89a78be51f1c2afd263da66ec76d3297432f0c2a&el=desc) will **decrease** coverage by `0.88%`.
> The diff coverage is `n/a`.
[](https://codecov.io/gh/huggingface/transformers/pull/5810?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5810 +/- ##
==========================================
- Coverage 78.19% 77.30% -0.89%
==========================================
Files 146 146
Lines 26047 26047
==========================================
- Hits 20367 20136 -231
- Misses 5680 5911 +231
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5810?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/tokenization\_bert.py](https://codecov.io/gh/huggingface/transformers/pull/5810/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fYmVydC5weQ==) | `91.32% <ø> (ø)` | |
| [src/transformers/modeling\_tf\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5810/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.39%)` | :arrow_down: |
| [src/transformers/modeling\_tf\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5810/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5810/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.46% <0.00%> (+0.25%)` | :arrow_up: |
| [src/transformers/file\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5810/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.49% <0.00%> (+0.29%)` | :arrow_up: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5810/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `82.31% <0.00%> (+1.28%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5810/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5810/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `95.18% <0.00%> (+74.91%)` | :arrow_up: |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5810?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5810?src=pr&el=footer). Last update [89a78be...d7b3627](https://codecov.io/gh/huggingface/transformers/pull/5810?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>@LysandreJik
Can you merge this PR? Thank you:) |
transformers | 5,809 | closed | TypeError: forward() got an unexpected keyword argument 'head_mask' | I'm getting the above error while training for EncoderDecoder Model for longformer.
Following is the code snippet from the training loop, followed by the model definition.
` output = model(input_ids = b_input_ids, attention_mask = b_input_masks, decoder_input_ids = b_decoder_input_ids, decoder_attention_mask = b_decoder_input_masks )
`
`model = EncoderDecoderModel.from_encoder_decoder_pretrained('allenai/longformer-base-4096','allenai/longformer-base-4096')
`
The crazy thing is I've not even defined 'head_mask' and left it to take its default value of 'none'.
Following the complete error.
`---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-18-68f3d2859ad0> in <module>()
32 optimizer.zero_grad()
33
---> 34 output = model(input_ids = b_input_ids, attention_mask = b_input_masks, head_mask = None,decoder_input_ids = b_decoder_input_ids, decoder_attention_mask = b_decoder_input_masks )
35 loss = output[0]
36
2 frames
/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py in __call__(self, *input, **kwargs)
548 result = self._slow_forward(*input, **kwargs)
549 else:
--> 550 result = self.forward(*input, **kwargs)
551 for hook in self._forward_hooks.values():
552 hook_result = hook(self, input, result)
TypeError: forward() got an unexpected keyword argument 'head_mask'` | 07-16-2020 13:09:06 | 07-16-2020 13:09:06 | The `EncoderDecoderModel` does not work with longformer yet. Closing this in favor of https://github.com/huggingface/transformers/issues/4225 . |
transformers | 5,808 | closed | [Benchmark] Fix models without `architectures` param in config | This PR fixes a bug, when running benchmarks for models that have `config.architectures = None`, *e.g.* `longformer-base-4096`. | 07-16-2020 13:03:39 | 07-16-2020 13:03:39 | |
transformers | 5,807 | closed | While running finetune.py in seq2seq examples on a custom dataset, I am getting the following error. | ### This is what I'm getting:
Can I get some help? Thank you
Epoch 1: 34% 3410/10000 [1:43:10<3:19:22, 1.82s/it, loss=3.052, v_num=6]Traceback (most recent call last):
File "finetune.py", line 344, in <module>
main(args)
File "finetune.py", line 322, in main
logger=logger,
File "/content/drive/My Drive/Colab Notebooks/transformers/examples/lightning_base.py", line 339, in generic_train
trainer.fit(model)
File "/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/trainer.py", line 918, in fit
self.single_gpu_train(model)
File "/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/distrib_parts.py", line 176, in single_gpu_train
self.run_pretrain_routine(model)
File "/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/trainer.py", line 1093, in run_pretrain_routine
self.train()
File "/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/training_loop.py", line 375, in train
self.run_training_epoch()
File "/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/training_loop.py", line 458, in run_training_epoch
_outputs = self.run_training_batch(batch, batch_idx)
File "/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/training_loop.py", line 634, in run_training_batch
loss, batch_output = optimizer_closure()
File "/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/training_loop.py", line 598, in optimizer_closure
output_dict = self.training_forward(split_batch, batch_idx, opt_idx, self.hiddens)
File "/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/training_loop.py", line 773, in training_forward
output = self.model.training_step(*args)
File "finetune.py", line 131, in training_step
loss_tensors = self._step(batch)
File "finetune.py", line 126, in _step
outputs = self(source_ids, attention_mask=source_mask, decoder_input_ids=y_ids, labels=lm_labels,)
File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 550, in __call__
result = self.forward(*input, **kwargs)
File "finetune.py", line 112, in forward
return self.model(input_ids, **kwargs)
File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 550, in __call__
result = self.forward(*input, **kwargs)
File "/usr/local/lib/python3.6/dist-packages/transformers/modeling_t5.py", line 1175, in forward
return_tuple=return_tuple,
File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 550, in __call__
result = self.forward(*input, **kwargs)
File "/usr/local/lib/python3.6/dist-packages/transformers/modeling_t5.py", line 692, in forward
input_ids = input_ids.view(-1, input_shape[-1])
RuntimeError: cannot reshape tensor of 0 elements into shape [-1, 0] because the unspecified dimension size -1 can be any value and is ambiguous
Epoch 1: 34%|███▍ | 3410/10000 [1:43:10<3:19:23, 1.82s/it, loss=3.052, ### v_num=6]
| 07-16-2020 12:58:56 | 07-16-2020 12:58:56 | Hi @laibamehnaz , what is your batch size ?
Here's my wild guess. I think one of your batch has empty lines. The batches are trimmed using the below function to remove excessive pad tokens
```python3
def trim_batch(
input_ids, pad_token_id, attention_mask=None,
):
"""Remove columns that are populated exclusively by pad_token_id"""
keep_column_mask = input_ids.ne(pad_token_id).any(dim=0)
if attention_mask is None:
return input_ids[:, keep_column_mask]
else:
return (input_ids[:, keep_column_mask], attention_mask[:, keep_column_mask])
```
if your batch has empty lines then after tokenizing it'll contain only pad tokens, and when its passed to trim batch it'll remove all those pad tokens and the batch will be empty i.e of shape [batch_size, 0]. This could the reason for the `RuntimeError`
you can verify this with
```python3
tok = T5Tokenizer.from_pretrained("t5-base")
text_batch = ["", ""]
enc = tok(text_batch, max_length=64, padding="max_length", return_tensors="pt")
print(enc['input_ids'].shape)
ids = trim_batch(enc['input_ids'], tok.pad_token_id)
input_shape = ids.size()
ids = ids.view(-1, input_shape[-1])
```
it'll give the same `RuntimeError`.
So I think you should check your dataset for empty lines, that should solve this issue <|||||>Hi @patil-suraj ,
Thank you so much for your quick response. I have checked my dataset already for the same issue, and it didn't show any empty lines, and that's why I was so confused. But sure, I will check again. Also, my batch_size=1. <|||||>If your bs is 1 then most probably blank line is the reason <|||||>@laibamehnaz can you just iterate through the dataloader with bs 1 and see when the shape comes out to be [bs, 0] this should allow you to pinpoint the error<|||||>Hi @patil-suraj, I tried what you suggested, but I do not get the same error. My smallest sentence is 1 word long, and there are no empty lines. <|||||>Interesting, can you post the word here, maybe tokenizer is skipping it then. One more thing you can try is do forward pass only with that one word example and see if it gives the error<|||||>Also pinging @sshleifer <|||||>Also, there is [this line](https://github.com/huggingface/transformers/blob/ab0b3fa15c159034a55ee95ac39b7a6ba4527c4a/examples/seq2seq/finetune.py#L133) in finetune.py, which suggests that if your decoder_input_ids are length 0 and you're not adding special tokens (as was the case previously for t5), the labels would get truncated to shape (1, 0).<|||||>https://github.com/huggingface/transformers/blob/ab0b3fa15c159034a55ee95ac39b7a6ba4527c4a/examples/seq2seq/finetune.py#L133<|||||>Right, so one word example is encoded as single id and due to shifting and no special token, shape becomes (1,0). So one more reason to automatically add `</s>` with t5 tokenizer <|||||>@laibamehnaz so immediate solution for this issue is to add `</s>` at the end of every target example text<|||||>Alright, thank you so much. Will get back after trying it out. <|||||>Works fine now. Thank you @patil-suraj.<|||||>hi @laibamehnaz , there's another bug in finetune.py when using T5. See issue #5987 and PR #5994<|||||>Fix is easy, just check the change in the PR<|||||>Thanks a lot!<|||||>That PR is merged now, you can now just clone the master or pull if you've already cloned<|||||>I did notice that. Thanks a lot!<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,806 | closed | BERT-viewer is broken for russian? | # 🐛 Bug
Hi I'm looking at BERT demos for russian, for example:
https://huggingface.co/DeepPavlov/rubert-base-cased?text=%D0%AF+%5BMASK%5D+%D1%81%D0%B0%D0%BB%D0%B0%D1%82+%D0%BD%D0%B0+%D0%BE%D0%B1%D0%B5%D0%B4.
English version for comparison:
https://huggingface.co/bert-base-cased?text=I+%5BMASK%5D+salad+for+lunch.
In the first link all token scores are 0.0 and they are absolute rubbish. Is everything fine with the viewer or the model? | 07-16-2020 12:11:11 | 07-16-2020 12:11:11 | I don't see anything obviously wrong on our side, so you should ask the model authors (@deeppavlov and other members of @deepmipt) – maybe they uploaded the model without a trained final LM prediction layer, for instance.
Off topic, but e.g. by contrast this translation model works very well for russian: https://huggingface.co/Helsinki-NLP/opus-mt-ru-en?text=%D0%A3+%D0%BC%D0%B5%D0%BD%D1%8F+%D0%B1%D1%8B%D0%BB+%D1%81%D0%B0%D0%BB%D0%B0%D1%82+%D0%BD%D0%B0+%D0%BE%D0%B1%D0%B5%D0%B4<|||||>Hi!
We (DeepPavlov) might have a problem with Russian model that was uploaded to HugginFace and we need to check it.
Everything should be okay with MLM head if you use checkpoints from here:
http://docs.deeppavlov.ai/en/master/features/pretrained_vectors.html#downloads
<|||||>This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,805 | closed | [Fix] Checkpoint saving and I/O for XLNetQASimple | **1. [Fix] Create directories for model checkpoints**
Saving model checkpoints without making directories would lead to error.
(i.e. The output directory will not exist using the currenct script.)
The [older version](https://github.com/huggingface/transformers/blob/3f5ccb183e3cfa755dea2dd2afd9abbf1a0f93b8/examples/run_squad.py#L193) of the part of this script works fine.
**2. [Fix] Fix wrong input/output structures for XLNet**
Running XLNet using the current script would lead to this error below.
> TypeError: forward() got an unexpected keyword argument 'cls_index'
The reason is that the [AutoModelForQuestionAnswering](https://github.com/huggingface/transformers/blob/ba2400189b2242620868096ae49babf93bd9ce00/examples/question-answering/run_squad.py#L735) would initialize XLNetForQuestionAnsweringSimple, while the I/O structures are written for XLNetForQuestionAnswering. | 07-16-2020 11:51:16 | 07-16-2020 11:51:16 | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,804 | closed | Add MPNet | 07-16-2020 11:41:39 | 07-16-2020 11:41:39 | @StillKeepTry Don't forget to ping me if you need another round of review!
By the way, I haven't seen any refactorization for code like this:
```python
from .modeling_bert import (
BertEmbeddings,
BertPreTrainedModel,
BertIntermediate,
BertOutput,
BertSelfAttention,
BertSelfOutput,
BertAttention,
BertModel
)
```
I recommend you copy the code like `BertEmbeddings` and paste them into your own modeling file and rename it to `MPNetEmebedding`.<|||||>Would you please add tests for MPNet? You can do it following other models.<|||||>Also there are several CI fails, please fix it and we can move to the next step. Thanks for your great contribution!<|||||>Hey @StillKeepTry - there seems to have been a problem with a git rebase. We sadly cannot review 905 changed files. Can you open a new PR that only has the necessary changes? <|||||>> Hey @StillKeepTry - there seems to have been a problem with a git rebase. We sadly cannot review 905 changed files. Can you open a new PR that only has the necessary changes?
ok<|||||>The new PR is submitted at [https://github.com/huggingface/transformers/pull/8804](https://github.com/huggingface/transformers/pull/8804)
@patrickvonplaten |
|
transformers | 5,803 | closed | Hosted Inference API: Error loading tokenizer Can't load config | Hi,
[https://huggingface.co/sampathkethineedi/industry-classification-api](https://huggingface.co/sampathkethineedi/industry-classification-api)
I uploaded my classification model fine tuned on BERT. There is no issue running the model from the hub and using the ‘sentiment-analysis’ pipeline. But there seems to be some problem when it comes to Hosted Inference API.
```
Error loading tokenizer Can't load config for 'sampathkethineedi/industry-classification-api classification'. Make sure that: - 'sampathkethineedi/industry-classification-api' is a correct model identifier listed on 'https://huggingface.co/models' - or 'sampathkethineedi/industry-classification-api' is the correct path to a directory containing a config.json file OSError("Can't load config for 'sampathkethineedi/industry-classification-api'. Make sure that:\n\n- 'sampathkethineedi/industry-classification-api' is a correct model identifier listed on 'https://huggingface.co/models'\n\n- or 'sampathkethineedi/industry-classification-api' is the correct path to a directory containing a config.json file\n\n")
```
Can someone help me with this?
Thanks! | 07-16-2020 11:13:52 | 07-16-2020 11:13:52 | I can load the pipeline locally so I am not sure what's happening here, indeed. Will post here when we know more.<|||||>Hey @julien-c, thanks for the quick response!
I'm assuming the inference API is using the `transformers-cli serve` ?
When I do the same locally, I'm able to get results at **localhost:8888/forward** using `POST -d {"inputs": ["< Text >"]}`
But when I try it on the https://api-inference.huggingface.co/models/sampathkethineedi/industry-classification-api, it doesn't work.
The Hosted Inference API has the same structure as `transformers-cli serve` or is it any different?<|||||>[https://huggingface.co/sampathkethineedi/industry-classification-api](https://huggingface.co/sampathkethineedi/industry-classification-api)
I just checked and the API is working. Would love to know what was causing the problem. Closing the issue.<|||||>@sampathkethineedi can i get dataset u trained on |
transformers | 5,802 | closed | [WIP - Benchmark] Add generate function | This PR adds the `generate` function to Benchmarks. You can try it out by running:
```
python examples/benchmarking/run_benchmark.py --models gpt2 --batch_sizes 1 --sequence_lengths 20 --no_inference --generate
```
## TODO
- [ ] Add for TF as well | 07-16-2020 10:31:01 | 07-16-2020 10:31:01 | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
|
transformers | 5,801 | closed | [Benchmark] fix benchmark non standard model | This PR fixes a typo in Benchmark, that allows to load specifc architectures and not just the base model. | 07-16-2020 10:02:42 | 07-16-2020 10:02:42 | # [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5801?src=pr&el=h1) Report
> Merging [#5801](https://codecov.io/gh/huggingface/transformers/pull/5801?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/8ce610bc96ace8be6b514c0f2c3fb0f76eb7ee15&el=desc) will **increase** coverage by `0.08%`.
> The diff coverage is `75.00%`.
[](https://codecov.io/gh/huggingface/transformers/pull/5801?src=pr&el=tree)
```diff
@@ Coverage Diff @@
## master #5801 +/- ##
==========================================
+ Coverage 77.33% 77.42% +0.08%
==========================================
Files 146 146
Lines 26047 26047
==========================================
+ Hits 20143 20166 +23
+ Misses 5904 5881 -23
```
| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5801?src=pr&el=tree) | Coverage Δ | |
|---|---|---|
| [src/transformers/benchmark/benchmark\_tf.py](https://codecov.io/gh/huggingface/transformers/pull/5801/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9iZW5jaG1hcmsvYmVuY2htYXJrX3RmLnB5) | `65.03% <50.00%> (+3.49%)` | :arrow_up: |
| [src/transformers/benchmark/benchmark.py](https://codecov.io/gh/huggingface/transformers/pull/5801/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9iZW5jaG1hcmsvYmVuY2htYXJrLnB5) | `81.10% <100.00%> (+7.08%)` | :arrow_up: |
| [src/transformers/modeling\_tf\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5801/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |
| [src/transformers/tokenization\_xlm.py](https://codecov.io/gh/huggingface/transformers/pull/5801/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25feGxtLnB5) | `16.26% <0.00%> (-66.67%)` | :arrow_down: |
| [src/transformers/data/processors/squad.py](https://codecov.io/gh/huggingface/transformers/pull/5801/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9kYXRhL3Byb2Nlc3NvcnMvc3F1YWQucHk=) | `28.34% <0.00%> (-6.55%)` | :arrow_down: |
| [src/transformers/generation\_tf\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5801/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `83.70% <0.00%> (-2.76%)` | :arrow_down: |
| [src/transformers/modeling\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5801/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |
| [src/transformers/tokenization\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5801/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHMucHk=) | `89.95% <0.00%> (-0.81%)` | :arrow_down: |
| ... and [1 more](https://codecov.io/gh/huggingface/transformers/pull/5801/diff?src=pr&el=tree-more) | |
------
[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5801?src=pr&el=continue).
> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)
> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`
> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5801?src=pr&el=footer). Last update [8ce610b...dbf09d1](https://codecov.io/gh/huggingface/transformers/pull/5801?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).
<|||||>Pinging @LysandreJik for notification.<|||||>Nice! |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.