Datasets:

---

DanielPFlorian

/

Transformers-Github-Issues

like 0

[]

[]

[]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5119?src=pr&el=h1) Report\n> Merging [#5119](https://codecov.io/gh/huggingface/transformers/pull/5119?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/355954ffca798bb81d9db8886e30ce10f11e8a40&el=desc) will **decrease** coverage by `0.00%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5119?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5119 +/- ##\n==========================================\n- Coverage 77.28% 77.28% -0.01% \n==========================================\n Files 133 133 \n Lines 22134 22134 \n==========================================\n- Hits 17107 17106 -1 \n- Misses 5027 5028 +1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5119?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5119/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `75.00% <0.00%> (-0.41%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5119/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.26% <0.00%> (-0.16%)` | :arrow_down: |\n| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/5119/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `39.62% <0.00%> (+0.23%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5119?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5119?src=pr&el=footer). Last update [355954f...09ff62a](https://codecov.io/gh/huggingface/transformers/pull/5119?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[ { "login": "patrickvonplaten", "id": 23423619, "node_id": "MDQ6VXNlcjIzNDIzNjE5", "avatar_url": "https://avatars.githubusercontent.com/u/23423619?v=4", "gravatar_id": "", "url": "https://api.github.com/users/patrickvonplaten", "html_url": "https://github.com/patrickvonplaten", "followers_url": "https://api.github.com/users/patrickvonplaten/followers", "following_url": "https://api.github.com/users/patrickvonplaten/following{/other_user}", "gists_url": "https://api.github.com/users/patrickvonplaten/gists{/gist_id}", "starred_url": "https://api.github.com/users/patrickvonplaten/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/patrickvonplaten/subscriptions", "organizations_url": "https://api.github.com/users/patrickvonplaten/orgs", "repos_url": "https://api.github.com/users/patrickvonplaten/repos", "events_url": "https://api.github.com/users/patrickvonplaten/events{/privacy}", "received_events_url": "https://api.github.com/users/patrickvonplaten/received_events", "type": "User", "site_admin": false } ]

[ "@patrickvonplaten, do you want to take a look?", "Hey @chrisdoyleIE ,\r\n\r\nUsing your code example above (I corrected some typos and missing imports), I am not able to reproduce the error. If a longer input is need to produce this error, please provide all necessary code to reproduce this error. Ideally, I should be able to copy paste the code into a console and get the same error as you :-)", "Hey @patrickvonplaten , \r\n\r\nI'll do some digging and see if I can't reproduce it myself such that it's easily paste-able and then I can share this code (I currently have a few custom packages calling eachother which is hairier than i'd like and not trivial to insert into an issue).", "I found a similar error when doing summarization, and just wanted to follow up on this. I have been stuck on this for a little bit now and I just wanted to check if there was a simple user-end solution to this, maybe incorrect arguments, etc.\r\nThis is a simplified notebook with the error: https://colab.research.google.com/drive/1Fj74x2NDJbhsty-oXzfhOCO185T80zw3?usp=sharing", "Thanks for the notebook @MathewPerez! Will checkit now ", "Awesome, I can reproduce the error - will look at a fix now :-) ", "The problem is that `max_length` is not bigger than `cur_len` so that model will not produce any text. \r\nThis will fix the problem:\r\n\r\n```python\r\noutputs = model.generate(input_ids=input_ids, num_beams=3, max_length=75)\r\n```" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null }, { "id": 1843244711, "node_id": "MDU6TGFiZWwxODQzMjQ0NzEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/New%20model", "name": "New model", "color": "fbca04", "default": false, "description": "" } ]

[]

[ "Can we update to add this into transformers", "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.\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5116?src=pr&el=h1) Report\n> Merging [#5116](https://codecov.io/gh/huggingface/transformers/pull/5116?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/355954ffca798bb81d9db8886e30ce10f11e8a40&el=desc) will **increase** coverage by `0.00%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5116?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5116 +/- ##\n=======================================\n Coverage 77.28% 77.29% \n=======================================\n Files 133 133 \n Lines 22134 22135 +1 \n=======================================\n+ Hits 17107 17109 +2 \n+ Misses 5027 5026 -1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5116?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5116/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.44% <100.00%> (+0.01%)` | :arrow_up: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5116/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `75.00% <0.00%> (-0.41%)` | :arrow_down: |\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5116/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `91.06% <0.00%> (+0.12%)` | :arrow_up: |\n| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/5116/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `39.62% <0.00%> (+0.23%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5116?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5116?src=pr&el=footer). Last update [355954f...fa3e6dc](https://codecov.io/gh/huggingface/transformers/pull/5116?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 2139563322, "node_id": "MDU6TGFiZWwyMTM5NTYzMzIy", "url": "https://api.github.com/repos/huggingface/transformers/labels/cleanup", "name": "cleanup", "color": "e7fc49", "default": false, "description": "" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5115?src=pr&el=h1) Report\n> Merging [#5115](https://codecov.io/gh/huggingface/transformers/pull/5115?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/355954ffca798bb81d9db8886e30ce10f11e8a40&el=desc) will **increase** coverage by `0.01%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5115?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5115 +/- ##\n==========================================\n+ Coverage 77.28% 77.30% +0.01% \n==========================================\n Files 133 133 \n Lines 22134 22130 -4 \n==========================================\n+ Hits 17107 17108 +1 \n+ Misses 5027 5022 -5 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5115?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5115/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.85% <100.00%> (+0.42%)` | :arrow_up: |\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5115/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `91.04% <100.00%> (+0.09%)` | :arrow_up: |\n| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/5115/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `39.62% <0.00%> (+0.23%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5115?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5115?src=pr&el=footer). Last update [355954f...5d77afc](https://codecov.io/gh/huggingface/transformers/pull/5115?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Great!" ]

[]

[ { "login": "sgugger", "id": 35901082, "node_id": "MDQ6VXNlcjM1OTAxMDgy", "avatar_url": "https://avatars.githubusercontent.com/u/35901082?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sgugger", "html_url": "https://github.com/sgugger", "followers_url": "https://api.github.com/users/sgugger/followers", "following_url": "https://api.github.com/users/sgugger/following{/other_user}", "gists_url": "https://api.github.com/users/sgugger/gists{/gist_id}", "starred_url": "https://api.github.com/users/sgugger/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/sgugger/subscriptions", "organizations_url": "https://api.github.com/users/sgugger/orgs", "repos_url": "https://api.github.com/users/sgugger/repos", "events_url": "https://api.github.com/users/sgugger/events{/privacy}", "received_events_url": "https://api.github.com/users/sgugger/received_events", "type": "User", "site_admin": false } ]

[]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "Did you try to train with half precision using the apex/amp package?", "No, I didn't. Anyway, this was only evaluating the pretrained model.", "I now have installed nvidia apex and tried training a new model with fp16. It ran out of memory after around 10% of the evaluation loop. I didn't expect it, since the model shouldn't need more memory in the middle of evaluation, with all batches having the full sequence length (in my case, 16k).\r\n\r\nIs there maybe something to optimize GPU memory usage?", "Hmm, lemme check...\r\nWhat is the sequence length and batch size you use excatly? Also, you can reduce `num_hashes` to save memory.", "I'm using this training setup with the `Trainer` from `huggingface`:\r\n```\r\n axial_pos_emb_dim = 128, 384\r\n hidden_dim = sum(axial_pos_emb_dim)\r\n\r\n axial_pos_max = 64, 256 # the product of this is the maximum length\r\n max_length = axial_pos_max[0] * axial_pos_max[1]\r\n\r\n hidden_dropout = 0.2\r\n attn_dropout = 0.1\r\n ff_dim = 2 * hidden_dim\r\n num_heads = 8\r\n dim_per_head = hidden_dim // num_heads\r\n num_layers = 6\r\n layers = ['local', 'lsh'] * (num_layers // 2)\r\n chunk_size = 0\r\n bucket_size = 64\r\n num_hashes = 2\r\n vocab_size = 258\r\n\r\n config = ReformerConfig(\r\n dim_per_head, layers, chunk_size_feed_forward=chunk_size,\r\n axial_pos_embds_dim=axial_pos_emb_dim,\r\n axial_pos_shape=axial_pos_max,\r\n max_position_embeddings=max_length,\r\n eos_token_id=1, feed_forward_size=ff_dim,\r\n hidden_dropout_prob=hidden_dropout, hidden_size=hidden_dim,\r\n lsh_attention_probs_dropout_prob=attn_dropout,\r\n local_attention_probs_dropout_prob=attn_dropout,\r\n num_attention_heads=num_heads, num_buckets=None,\r\n pad_token_id=0, lsh_attn_chunk_length=bucket_size,\r\n num_hashes=num_hashes, vocab_size=vocab_size)\r\n model = ReformerModelWithLMHead(config)\r\n\r\n training_args = TrainingArguments(\r\n 'model', do_train=True, do_eval=True,\r\n do_predict=False, evaluate_during_training=True,\r\n gradient_accumulation_steps=1,\r\n learning_rate=0.001,\r\n logging_dir='model/tensorboard',\r\n logging_steps=5,\r\n save_steps=1000,\r\n per_device_train_batch_size=1,\r\n per_device_eval_batch_size=1,\r\n fp16=True)\r\n```\r\n\r\nUnless I'm doing something wrong, it's a batch size of 1 (both for training and evaluating) and a sequence length of 16384 (64 * 256).", "Ok, so I found that the main culprit was that the `Trainer` was storing all model predictions in GPU memory during evaluation at https://github.com/huggingface/transformers/blob/c01480bba3b2f0bd8516679476235f4701c21b3b/src/transformers/trainer.py#L775\r\n\r\nPassing `prediction_loss_only=False` avoided that. By the way, I believe this should be the default value in the `Trainer`, and that the `cat` operation could use cpu tensors, in case the validation dataset is big. ", "Out of curiosity, how big is your validation dataset/how large is the in-memory size of `preds` in that prediction loop?", "@julien-c this happened with the enwik8 test set, 5M characters.", "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.\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "The notebook is not using the collation function from `transformers` but the `T2TDataCollator` it defines. Apart from stopping subclassing `DataCollator` (as it's not a class anymore on master) this shouldn't be a problem.", "Hi, @antoniomastro1996 , if you are using transformers from master branch then there are few changes.\r\nAs @sgugger said, `DataCollator` is not a class anymore, so just don't subclass. Also as datacollator is now a `callable`, rename the `batch_collate` method to `__call__`.\r\n\r\nOr you can use version 2.11.0 and it'll work as it is. I'll update the notebook once the new version is released. Let me know if you run into any other issues.", "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.\n" ]

[]

[]

[ "Is your `batch` on the correct device too?", "Thank you, I moved `batch` to the GPU and it works.\r\n`batch = tokenizer.prepare_translation_batch(sentences).to('cuda')`", "More generally: `batch.to(model.device)`." ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5110?src=pr&el=h1) Report\n> Merging [#5110](https://codecov.io/gh/huggingface/transformers/pull/5110?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/355954ffca798bb81d9db8886e30ce10f11e8a40&el=desc) will **increase** coverage by `0.03%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5110?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5110 +/- ##\n==========================================\n+ Coverage 77.28% 77.32% +0.03% \n==========================================\n Files 133 133 \n Lines 22134 22164 +30 \n==========================================\n+ Hits 17107 17139 +32 \n+ Misses 5027 5025 -2 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5110?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/\\_\\_init\\_\\_.py](https://codecov.io/gh/huggingface/transformers/pull/5110/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9fX2luaXRfXy5weQ==) | `99.16% <ø> (ø)` | |\n| [src/transformers/modeling\\_auto.py](https://codecov.io/gh/huggingface/transformers/pull/5110/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19hdXRvLnB5) | `70.76% <ø> (ø)` | |\n| [src/transformers/modeling\\_xlm.py](https://codecov.io/gh/huggingface/transformers/pull/5110/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ194bG0ucHk=) | `90.23% <100.00%> (+0.98%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5110/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.42% <0.00%> (ø)` | |\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5110/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `91.06% <0.00%> (+0.12%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5110?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5110?src=pr&el=footer). Last update [355954f...439c7d0](https://codecov.io/gh/huggingface/transformers/pull/5110?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "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.\n", "In the end this was added by #5614, so closing here." ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[ { "login": "julien-c", "id": 326577, "node_id": "MDQ6VXNlcjMyNjU3Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/326577?v=4", "gravatar_id": "", "url": "https://api.github.com/users/julien-c", "html_url": "https://github.com/julien-c", "followers_url": "https://api.github.com/users/julien-c/followers", "following_url": "https://api.github.com/users/julien-c/following{/other_user}", "gists_url": "https://api.github.com/users/julien-c/gists{/gist_id}", "starred_url": "https://api.github.com/users/julien-c/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/julien-c/subscriptions", "organizations_url": "https://api.github.com/users/julien-c/orgs", "repos_url": "https://api.github.com/users/julien-c/repos", "events_url": "https://api.github.com/users/julien-c/events{/privacy}", "received_events_url": "https://api.github.com/users/julien-c/received_events", "type": "User", "site_admin": false }, { "login": "mfuntowicz", "id": 2241520, "node_id": "MDQ6VXNlcjIyNDE1MjA=", "avatar_url": "https://avatars.githubusercontent.com/u/2241520?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mfuntowicz", "html_url": "https://github.com/mfuntowicz", "followers_url": "https://api.github.com/users/mfuntowicz/followers", "following_url": "https://api.github.com/users/mfuntowicz/following{/other_user}", "gists_url": "https://api.github.com/users/mfuntowicz/gists{/gist_id}", "starred_url": "https://api.github.com/users/mfuntowicz/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/mfuntowicz/subscriptions", "organizations_url": "https://api.github.com/users/mfuntowicz/orgs", "repos_url": "https://api.github.com/users/mfuntowicz/repos", "events_url": "https://api.github.com/users/mfuntowicz/events{/privacy}", "received_events_url": "https://api.github.com/users/mfuntowicz/received_events", "type": "User", "site_admin": false }, { "login": "sshleifer", "id": 6045025, "node_id": "MDQ6VXNlcjYwNDUwMjU=", "avatar_url": "https://avatars.githubusercontent.com/u/6045025?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sshleifer", "html_url": "https://github.com/sshleifer", "followers_url": "https://api.github.com/users/sshleifer/followers", "following_url": "https://api.github.com/users/sshleifer/following{/other_user}", "gists_url": "https://api.github.com/users/sshleifer/gists{/gist_id}", "starred_url": "https://api.github.com/users/sshleifer/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/sshleifer/subscriptions", "organizations_url": "https://api.github.com/users/sshleifer/orgs", "repos_url": "https://api.github.com/users/sshleifer/repos", "events_url": "https://api.github.com/users/sshleifer/events{/privacy}", "received_events_url": "https://api.github.com/users/sshleifer/received_events", "type": "User", "site_admin": false }, { "login": "patrickvonplaten", "id": 23423619, "node_id": "MDQ6VXNlcjIzNDIzNjE5", "avatar_url": "https://avatars.githubusercontent.com/u/23423619?v=4", "gravatar_id": "", "url": "https://api.github.com/users/patrickvonplaten", "html_url": "https://github.com/patrickvonplaten", "followers_url": "https://api.github.com/users/patrickvonplaten/followers", "following_url": "https://api.github.com/users/patrickvonplaten/following{/other_user}", "gists_url": "https://api.github.com/users/patrickvonplaten/gists{/gist_id}", "starred_url": "https://api.github.com/users/patrickvonplaten/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/patrickvonplaten/subscriptions", "organizations_url": "https://api.github.com/users/patrickvonplaten/orgs", "repos_url": "https://api.github.com/users/patrickvonplaten/repos", "events_url": "https://api.github.com/users/patrickvonplaten/events{/privacy}", "received_events_url": "https://api.github.com/users/patrickvonplaten/received_events", "type": "User", "site_admin": false } ]

[ "I can't replicate (I understand that's the point :)) \r\n\r\nThe traceback from your link is:\r\n\r\n```python\r\nOSError: Can't load weights for 'sshleifer/tiny-distilroberta-base'. Make sure that:\r\nE \r\nE - 'sshleifer/tiny-distilroberta-base' is a correct model identifier listed on 'https://huggingface.co/models'\r\nE \r\nE - or 'sshleifer/tiny-distilroberta-base' is the correct path to a directory containing a file named one of pytorch_model.bin, tf_model.h5, model.ckpt.\r\n```\r\n\r\nContents are fine, it seems:\r\n\r\n\r\n\r\n\r\n@julien-c have you seen similar intermittent S3 failures?", "From what I experience, this happens not just for the same model, but for different ones. Very hard to traceback what is going on there IMO.", "Yes, I had a similar circleci failure trying to get the bert tokenizer yesterday.\r\n", "I have found a clue. I hypothesize that some models have user posted metadata that somehow ducks them up. \r\n\r\nI have never seen flaky bart-large failure, and there is no metadata:\r\n\r\n```bash\r\n\r\naws s3api head-object --bucket models.huggingface.co --key bert/facebook/bart-large/config.json\r\n\r\n=> bytes\t1264\tapplication/json\t\"faf4c3c00764dc47a3a10a63a004dcc3\"\tFri, 24 Apr 2020 15:58:48 GMT\tJhIFsOvvLtrLn0vJjGNN6ZhJGUlbXBEP\r\n\r\n```\r\n\r\nwhereas there are `Helsinki-NLP/opus-mt-en-ROMANCE` failures, and metadata:\r\n\r\n```bash\r\naws s3api head-object --bucket models.huggingface.co --key bert/Helsinki-NLP/opus-mt-en-ROMANCE/config.json\r\n\r\n=> bytes\t1113\ttext/plain\t\"13ca8d49ee7f02a26f935cb4a60e6557\"\tTue, 12 May 2020 22:39:10 GMT\tjAI94kJ_exk0tG6z0Yr.Rea4_j0g02Ih\r\nMETADATA\tatime:1589321950/ctime:1589321949/gid:1007/gname:shleifer/md5:13ca8d49ee7f02a26f935cb4a60e6557/mode:33188/mtime:1589321949/uid:1006/uname:shleifer\r\n```\r\n\r\n\r\n\r\nNow I need to figure out how to get rid of the metadata to test my hypothesis, if anyone knows how.\r\n\r\n**Update:**\r\n```bash\r\nbytes\t474\tapplication/json\t\"622babbd58848ec69a2433ba0c6edab3\"\tTue, 12 May 2020 01:26:15 GMT\tUKLWouonOr8duRlrHqH4.iNZEF5bmOQY\r\nMETADATA\tatime:1589246769/ctime:1589246726/gid:20/gname:staff/md5:622babbd58848ec69a2433ba0c6edab3/mode:33188/mtime:1589246726/uid:501/uname:shleifer\r\n```\r\nalso shows metadata, and there are flaky failures with that model.\r\n", "Should be easy, I can guide you through the S3 doc if needed.\r\n\r\nI doubt that's the reason though. What about intermittent connectivity issues? They seem more likely.", "It could certainly be connectivity. I manually deleted the metadata for all contents of `Helsinki-NLP/opus-mt-en-ROMANCE/` and `sshleifer/tiny-distilroberta-base`, so if there are more failures there we will know that my theory is wrong. I think the metadata is caused by using `s3cmd` instead of `awscli`.", "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.\n" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5108?src=pr&el=h1) Report\n> Merging [#5108](https://codecov.io/gh/huggingface/transformers/pull/5108?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/355954ffca798bb81d9db8886e30ce10f11e8a40&el=desc) will **decrease** coverage by `0.08%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5108?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5108 +/- ##\n==========================================\n- Coverage 77.28% 77.20% -0.09% \n==========================================\n Files 133 133 \n Lines 22134 22134 \n==========================================\n- Hits 17107 17088 -19 \n- Misses 5027 5046 +19 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5108?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5108/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ194bG5ldC5weQ==) | `76.35% <0.00%> (-2.30%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5108/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `86.49% <0.00%> (-0.94%)` | :arrow_down: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5108/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `75.00% <0.00%> (-0.41%)` | :arrow_down: |\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5108/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `91.06% <0.00%> (+0.12%)` | :arrow_up: |\n| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/5108/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `39.62% <0.00%> (+0.23%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5108?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5108?src=pr&el=footer). Last update [355954f...6330b4e](https://codecov.io/gh/huggingface/transformers/pull/5108?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "> @julien-c check out that dataset meta tag is right\r\n\r\nIt seems everythigs is right" ]

[]

[]

[ "Hi! If you want to initialize a new `RobertaForSequenceClassification` model you can do so as such:\r\n\r\n```py\r\nfrom transformers import RobertaForSequenceClassification, RobertaConfig\r\n\r\nconfig = RobertaConfig(\r\n # put the args you need, like the vocab size\r\n vocab_size=100\r\n)\r\n\r\nmodel = RobertaForSequenceClassification(config)\r\n```", "I had the same problem when i try to finetune RoBERTa model with my own tokenizer as @raj5287 did. I have already updated all packages (pytorch, transformers)\r\n\r\nCode:\r\n\r\n> tokenizer = RobertaTokenizerFast.from_pretrained(tokenizer_path)\r\n> model = RobertaForMaskedLM.from_pretrained(self.config)\r\n> dataset = LineByLineTextDataset(\r\n> tokenizer=tokenizer,\r\n> file_path=input_text_path,\r\n> block_size=512,\r\n> )\r\n> \r\n> data_collator = DataCollatorForLanguageModeling(\r\n> tokenizer=tokenizer, mlm=True, mlm_probability=0.15\r\n> )\r\n> training_args = TrainingArguments(\r\n> output_dir=output_path,\r\n> overwrite_output_dir=True,\r\n> num_train_epochs=3,\r\n> per_gpu_train_batch_size=8,\r\n> save_steps=50_000,\r\n> save_total_limit=2,\r\n> )\r\n> \r\n> trainer = Trainer(\r\n> model=model,\r\n> args=training_args,\r\n> data_collator=data_collator,\r\n> train_dataset=dataset,\r\n> prediction_loss_only=True\r\n> )\r\n> trainer.train()\r\n> trainer.save_model(output_path)\r\n> \r\n\r\nOutput:\r\n\r\nFile \"Fine_tuning_tokenizer.py\", line 106, in <module>\r\n main(sys.argv[1:])\r\n File \"Fine_tuning_tokenizer.py\", line 94, in main\r\n modelRoberta.fine_tune_LM(inputfile, outputDir,tokenizer_path=outputDir)\r\n File \"/sentiment-embeddings/projeto-modularizar/src/uff/ic/mell/sentimentembedding/modelos/modelo_roberta.py\", line 58, in fine_tune_LM\r\n trainer.train()\r\n File \"/home/apps/anaconda3/lib/python3.7/site-packages/transformers/trainer.py\", line 549, in train\r\n tr_loss += self.training_step(model, inputs)\r\n File \"/home/apps/anaconda3/lib/python3.7/site-packages/transformers/trainer.py\", line 762, in training_step\r\n return loss.item()\r\nRuntimeError: CUDA error: an illegal memory access was encountered", "> I had the same problem when i try to finetune RoBERTa model with my own tokenizer as @raj5287 did. I have already updated all packages (pytorch, transformers)\r\n> \r\n> Code:\r\n> \r\n> > tokenizer = RobertaTokenizerFast.from_pretrained(tokenizer_path)\r\n> > model = RobertaForMaskedLM.from_pretrained(self.config)\r\n> > dataset = LineByLineTextDataset(\r\n> > tokenizer=tokenizer,\r\n> > file_path=input_text_path,\r\n> > block_size=512,\r\n> > )\r\n> > ```\r\n> > data_collator = DataCollatorForLanguageModeling(\r\n> > tokenizer=tokenizer, mlm=True, mlm_probability=0.15\r\n> > )\r\n> > training_args = TrainingArguments(\r\n> > output_dir=output_path,\r\n> > overwrite_output_dir=True,\r\n> > num_train_epochs=3,\r\n> > per_gpu_train_batch_size=8,\r\n> > save_steps=50_000,\r\n> > save_total_limit=2,\r\n> > )\r\n> > \r\n> > trainer = Trainer(\r\n> > model=model,\r\n> > args=training_args,\r\n> > data_collator=data_collator,\r\n> > train_dataset=dataset,\r\n> > prediction_loss_only=True\r\n> > )\r\n> > trainer.train()\r\n> > trainer.save_model(output_path)\r\n> > ```\r\n> \r\n> Output:\r\n> \r\n> File \"Fine_tuning_tokenizer.py\", line 106, in \r\n> main(sys.argv[1:])\r\n> File \"Fine_tuning_tokenizer.py\", line 94, in main\r\n> modelRoberta.fine_tune_LM(inputfile, outputDir,tokenizer_path=outputDir)\r\n> File \"/sentiment-embeddings/projeto-modularizar/src/uff/ic/mell/sentimentembedding/modelos/modelo_roberta.py\", line 58, in fine_tune_LM\r\n> trainer.train()\r\n> File \"/home/apps/anaconda3/lib/python3.7/site-packages/transformers/trainer.py\", line 549, in train\r\n> tr_loss += self.training_step(model, inputs)\r\n> File \"/home/apps/anaconda3/lib/python3.7/site-packages/transformers/trainer.py\", line 762, in training_step\r\n> return loss.item()\r\n> RuntimeError: CUDA error: an illegal memory access was encountered\r\n\r\nIs it ok to use \"DataCollatorForLanguageModeling\" in a classification task? ", "> I had the same problem when i try to finetune RoBERTa model with my own tokenizer as @raj5287 did. I have already updated all packages (pytorch, transformers)\r\n> \r\n> Code:\r\n> \r\n> > tokenizer = RobertaTokenizerFast.from_pretrained(tokenizer_path)\r\n> > model = RobertaForMaskedLM.from_pretrained(self.config)\r\n> > dataset = LineByLineTextDataset(\r\n> > tokenizer=tokenizer,\r\n> > file_path=input_text_path,\r\n> > block_size=512,\r\n> > )\r\n> > ```\r\n> > data_collator = DataCollatorForLanguageModeling(\r\n> > tokenizer=tokenizer, mlm=True, mlm_probability=0.15\r\n> > )\r\n> > training_args = TrainingArguments(\r\n> > output_dir=output_path,\r\n> > overwrite_output_dir=True,\r\n> > num_train_epochs=3,\r\n> > per_gpu_train_batch_size=8,\r\n> > save_steps=50_000,\r\n> > save_total_limit=2,\r\n> > )\r\n> > \r\n> > trainer = Trainer(\r\n> > model=model,\r\n> > args=training_args,\r\n> > data_collator=data_collator,\r\n> > train_dataset=dataset,\r\n> > prediction_loss_only=True\r\n> > )\r\n> > trainer.train()\r\n> > trainer.save_model(output_path)\r\n> > ```\r\n> \r\n> Output:\r\n> \r\n> File \"Fine_tuning_tokenizer.py\", line 106, in\r\n> main(sys.argv[1:])\r\n> File \"Fine_tuning_tokenizer.py\", line 94, in main\r\n> modelRoberta.fine_tune_LM(inputfile, outputDir,tokenizer_path=outputDir)\r\n> File \"/sentiment-embeddings/projeto-modularizar/src/uff/ic/mell/sentimentembedding/modelos/modelo_roberta.py\", line 58, in fine_tune_LM\r\n> trainer.train()\r\n> File \"/home/apps/anaconda3/lib/python3.7/site-packages/transformers/trainer.py\", line 549, in train\r\n> tr_loss += self.training_step(model, inputs)\r\n> File \"/home/apps/anaconda3/lib/python3.7/site-packages/transformers/trainer.py\", line 762, in training_step\r\n> return loss.item()\r\n> RuntimeError: CUDA error: an illegal memory access was encountered\r\n\r\nHey @SergioBarretoJr ,I guess one of the causes is when the model encounters an unseen token. and no we cannot use DataCollatorForLanguageModeling for classification task @Chandrak1907. will have to loop through epochs for training and fine tuning.", "Thks guys @raj5287 @Chandrak1907 ! I solved this problems! When I trained my BPE tokenizer, I defined my vocabulary size of 52.000 and I used RoBERTa config from hugging face that has a different vocabulary size. This caused this error.", "@raj5287 , @SergioBarretoJr Can you share github link for your code? I could not find proper reference on huggingface.co. Thanks.", "@SergioBarretoJr I am also facing the same issue, can you please help me?", "> @SergioBarretoJr I am also facing the same issue, can you please help me?\r\n\r\n@amandalmia14 Have you ever checked if you set vocab size equal in tokenizer and RoBERTa.config file? \r\n\r\n def train_tokenizer(self,file_path,outDir):\r\n # Initialize a tokenizer\r\n tokenizer = ByteLevelBPETokenizer()\r\n # Customize training\r\n tokenizer.train(files=file_path, **vocab_size=52_000**, min_frequency=2, special_tokens=[\r\n \"<s>\",\r\n \"<pad>\",\r\n \"</s>\",\r\n \"<unk>\",\r\n \"<mask>\",\r\n ])\r\n self.tokenizer=tokenizer\r\n tokenizer.save(outDir)\r\n\r\n\r\nThis was how I solver my issue. Can I help with something more?", "@SergioBarretoJr Thanks for the snippet, I am able to train the custom language model with the help of [this](https://colab.research.google.com/github/huggingface/blog/blob/master/notebooks/01_how_to_train.ipynb). However, after creating the custom Roberta Model, I want to use it for creating a multi-label sentence classifier. \r\n\r\nCan you help me with that snippet of code? Thanks " ]

[]

[]

[ "Not yet -> on the roadmap :-) ", "any updates here?" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "I solved it editing trainer.py, in this line https://github.com/huggingface/transformers/blob/master/src/transformers/trainer.py#L765 . I added start_positions and end_positions to that \"possible labels names\" list, and it worked. Review the bug, please.", "Thanks @alexvaca0 !\r\n\r\ncc @julien-c , @sgugger ", "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.\n", "> I solved it editing trainer.py, in this line https://github.com/huggingface/transformers/blob/master/src/transformers/trainer.py#L765 . I added start_positions and end_positions to that \"possible labels names\" list, and it worked. Review the bug, please.\r\n\r\nI am having the same issue. Can you elaborate on how you solved this problem? I could not locate a possible label list and start/end positions?", "#7191 should solve the problem described here.", "@sgugger How can we calculate validation and evaluation loss for question answering finetuning pipeline (run_qa.py)" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5103?src=pr&el=h1) Report\n> Merging [#5103](https://codecov.io/gh/huggingface/transformers/pull/5103?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/417e492f1e832c0b93512600d3385aa4c8a887c9&el=desc) will **decrease** coverage by `0.89%`.\n> The diff coverage is `87.26%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5103?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5103 +/- ##\n==========================================\n- Coverage 77.98% 77.09% -0.90% \n==========================================\n Files 138 138 \n Lines 23786 23836 +50 \n==========================================\n- Hits 18550 18376 -174 \n- Misses 5236 5460 +224 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5103?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/tokenization\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5103/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHMucHk=) | `91.48% <78.78%> (-3.34%)` | :arrow_down: |\n| [src/transformers/tokenization\\_utils\\_fast.py](https://codecov.io/gh/huggingface/transformers/pull/5103/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHNfZmFzdC5weQ==) | `91.97% <85.71%> (-0.63%)` | :arrow_down: |\n| [src/transformers/tokenization\\_utils\\_base.py](https://codecov.io/gh/huggingface/transformers/pull/5103/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHNfYmFzZS5weQ==) | `90.86% <91.93%> (-0.21%)` | :arrow_down: |\n| [src/transformers/tokenization\\_bert.py](https://codecov.io/gh/huggingface/transformers/pull/5103/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fYmVydC5weQ==) | `89.95% <100.00%> (-0.51%)` | :arrow_down: |\n| [src/transformers/tokenization\\_gpt2.py](https://codecov.io/gh/huggingface/transformers/pull/5103/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fZ3B0Mi5weQ==) | `97.12% <100.00%> (+0.04%)` | :arrow_up: |\n| [src/transformers/tokenization\\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5103/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fcm9iZXJ0YS5weQ==) | `94.52% <100.00%> (ø)` | |\n| [src/transformers/tokenization\\_transfo\\_xl.py](https://codecov.io/gh/huggingface/transformers/pull/5103/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdHJhbnNmb194bC5weQ==) | `42.48% <100.00%> (+0.13%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5103/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `19.92% <0.00%> (-75.00%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5103/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `85.86% <0.00%> (-0.15%)` | :arrow_down: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5103?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5103?src=pr&el=footer). Last update [417e492...ca291b5](https://codecov.io/gh/huggingface/transformers/pull/5103?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "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.\n" ]

[]

[]

[ "Hi @ghsama `lm_labels` is changed to `labels` after 2.11.0. So if you are using version <=2.11.0 use `lm_labels` and if you're using master branch then use `labels`", "Yes ! i just figure it out . Thank you @patil-suraj :D " ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "This might provide more information: https://github.com/conda-forge/transformers-feedstock/issues/3\r\n\r\nIn short, the conda-forge update is blocked by `sentencepiece` not being available through conda.", "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.\n", "The feedstock has now been updated for Linux and is available. Other OSes are still waiting on sentencepiece.", "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.\n", "This should be fixed by #8073 ", "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.\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5099?src=pr&el=h1) Report\n> Merging [#5099](https://codecov.io/gh/huggingface/transformers/pull/5099?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/efeb75b8054cc299698cf8bc09f395ada2660745&el=desc) will **increase** coverage by `0.04%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5099?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5099 +/- ##\n==========================================\n+ Coverage 77.24% 77.29% +0.04% \n==========================================\n Files 133 133 \n Lines 22134 22134 \n==========================================\n+ Hits 17097 17108 +11 \n+ Misses 5037 5026 -11 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5099?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/optimization\\_tf.py](https://codecov.io/gh/huggingface/transformers/pull/5099/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9vcHRpbWl6YXRpb25fdGYucHk=) | `57.27% <ø> (ø)` | |\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5099/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `91.06% <0.00%> (+0.12%)` | :arrow_up: |\n| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/5099/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `39.62% <0.00%> (+0.23%)` | :arrow_up: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5099/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `75.40% <0.00%> (+0.40%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5099/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.26% <0.00%> (+1.24%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5099?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5099?src=pr&el=footer). Last update [efeb75b...d99033f](https://codecov.io/gh/huggingface/transformers/pull/5099?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Thanks @Colanim! Good catch!! Unfortunately it is a duplicate of #4940 :smile: and should be merged soon :)", "Haha didn't see x)" ]

[]

[]

[]

[]

[]

[ "Hi, as seen with @sshleifer, you can train for summarization using the [summarization script](https://github.com/huggingface/transformers/tree/master/examples/summarization). The models supported right now are all BART variants and t5-small. More to come!" ]

[]

[]

[ "Yes", "> Yes\r\n\r\nThat' s awesome!Can you give a code to show? I'm grateful!", "So from the paper: https://arxiv.org/pdf/1910.13461.pdf, you can see that Bart is trained on denoising input sequences in almost any possible way. \r\n\r\nOne way could be for `BartForConditionalGeneration`:\r\n\r\n```python \r\nfrom transformers import BartTokenizer, BartForConditionalGeneration, BartConfig\r\n\r\ntok = BartTokenizer.from_pretrained(\"facebook/bart-large\")\r\nmodel = BartForConditionalGeneration(BartConfig())\r\n\r\ninput_string = \"My dog is <mask> </s>\"\r\ndecoder_input_string = \"<s> My dog is cute\"\r\nlabels_string = \"My dog is cute </s>\"\r\n\r\ninput_ids = tok(input_string, add_special_tokens=False, return_tensors=\"pt\").input_ids\r\ndecoder_input_ids =tok(decoder_input_string, add_special_tokens=False, return_tensors=\"pt\").input_ids\r\nlabels = tok(labels_string, add_special_tokens=False, return_tensors=\"pt\").input_ids\r\n \r\nloss = model(input_ids=input_ids, decoder_input_ids=decoder_input_ids, labels=labels)[0]\r\n```", "Pinging @sshleifer to make sure I did not forget anything", "> Pinging @sshleifer to make sure I did not forget anything\r\n\r\nActually, I was going to ask. how train a model from zero to one. For example, I want to train a Chinese bart model.", "Here's a working example for this, including batching:\r\n\r\n```\r\nfrom transformers import BartTokenizer, BartForConditionalGeneration, BartConfig\r\n\r\ntok = BartTokenizer.from_pretrained(\"facebook/bart-large\")\r\nmodel = BartForConditionalGeneration(BartConfig())\r\n\r\ninput_batch = [\"My dog is <mask></s>\", \"It loves to play in the <mask></s>\"]\r\ndecoder_input_batch = [\"<s>My dog is cute\", \"<s>It loves to play in the park\"]\r\nlabels_batch = [\"My dog is cute</s>\", \"It loves to play in the park</s>\"]\r\n\r\ninput_ids = tok.batch_encode_plus(input_batch, add_special_tokens=False, return_tensors=\"pt\", padding=True).input_ids\r\ndecoder_input_ids = tok.batch_encode_plus(decoder_input_batch, add_special_tokens=False, return_tensors=\"pt\", padding=True).input_ids\r\nlabels = tok.batch_encode_plus(labels_batch, add_special_tokens=False, return_tensors=\"pt\", padding=True).input_ids\r\n\r\nloss = model(input_ids=input_ids, decoder_input_ids=decoder_input_ids, labels=labels)[0]\r\n```\r\n\r\n`>>>` `tensor(10.9981, device='cuda:0', grad_fn=<NllLossBackward>)`", "> Here's a working example for this, including batching:\r\n> \r\n> ```\r\n> from transformers import BartTokenizer, BartForConditionalGeneration, BartConfig\r\n> \r\n> tok = BartTokenizer.from_pretrained(\"facebook/bart-large\")\r\n> model = BartForConditionalGeneration(BartConfig())\r\n> \r\n> input_batch = [\"My dog is <mask></s>\", \"It loves to play in the <mask></s>\"]\r\n> decoder_input_batch = [\"<s>My dog is cute\", \"<s>It loves to play in the park\"]\r\n> labels_batch = [\"My dog is cute</s>\", \"It loves to play in the park</s>\"]\r\n> \r\n> input_ids = tok.batch_encode_plus(input_batch, add_special_tokens=False, return_tensors=\"pt\", padding=True).input_ids\r\n> decoder_input_ids = tok.batch_encode_plus(decoder_input_batch, add_special_tokens=False, return_tensors=\"pt\", padding=True).input_ids\r\n> labels = tok.batch_encode_plus(labels_batch, add_special_tokens=False, return_tensors=\"pt\", padding=True).input_ids\r\n> \r\n> loss = model(input_ids=input_ids, decoder_input_ids=decoder_input_ids, labels=labels)[0]\r\n> ```\r\n> \r\n> `>>>` `tensor(10.9981, device='cuda:0', grad_fn=<NllLossBackward>)`\r\n\r\ninput_batch = [\"My dog is <mask></s>\", \"It loves to play in the <mask></s>\"]\r\ndecoder_input_batch = [\"<s>My dog is cute\", \"<s>It loves to play in the park\"]\r\nlabels_batch = [\"My dog is cute</s>\", \"It loves to play in the park</s>\"]\r\n\r\n\r\nIf I have a text document, each line of a paragraph, how do I rewrite the data input on it? Thanks!", "@tomhosking the paper indicates that it uses both sentence permutation (loss is propagated from all tokens instead of only masked tokens) and infilling (include only one mask token for multiple consecutive masks). would this be a correct input?\r\n\r\ninput_batch = [\"\\<s>It is \\<mask\\> retriever. My dog is \\<mask\\>\\</s>\", \"\\<s>There \\<mask\\> in SF. It loves to play in the \\<mask\\>\\</s>\"]\r\ndecoder_input_batch = [\"\\</s>\\<s>My dog is cute. It is a golden retriever\", \"\\</s>\\<s>It loves to play in the park. There are many parks in SF.\"]\r\nlabels_batch = [\"\\<s>My dog is cute. It is a golden retriever\\</s>\", \"\\<s>It loves to play in the park. There are many parks in SF.\\</s>\"]\r\n\r\n(Note: decoder_input_batch starts with \\</s>\\<s> due to shift_tokens_right #7961)", "Sorry for the intrusion, but I think your values are almost correct @swethmandava, except for the masking absence\r\n\r\n```python\r\ninput_batch = [\"<s>It <mask> retriever. My <mask> cute </s>\", ... ]\r\ndecoder_input_batch = [\"</s><s>My dog is cute. It is a golden retriever\", ...]\r\nlabels_batch = [\"<s>My dog is cute. It is a golden retriever</s>\", ...]\r\n```\r\n\r\nBTW: This `</s>` token at the beginning of decode's input is kind of weird to me, but it's inherited from the fairseq original code. If you wanna train the model from scratch with random weights I think you can go without this... or maybe this trick is important for convergence, we never know :grin:", "Will only 15% mask in the encoder input cause some kind of leakage? The language model in the decoder cannot learn correctly", "If anyone wants to train their MBART model then feel free to use this. \r\nhttps://github.com/prajdabre/yanmtt\r\n\r\nContributions are welcome!", "> Sorry for the intrusion, but I think your values are almost correct @swethmandava, except for the masking absence\r\n> \r\n> ```python\r\n> input_batch = [\"<s>It <mask> retriever. My <mask> cute </s>\", ... ]\r\n> decoder_input_batch = [\"</s><s>My dog is cute. It is a golden retriever\", ...]\r\n> labels_batch = [\"<s>My dog is cute. It is a golden retriever</s>\", ...]\r\n> ```\r\n> \r\n> BTW: This `</s>` token at the beginning of decode's input is kind of weird to me, but it's inherited from the fairseq original code. If you wanna train the model from scratch with random weights I think you can go without this... or maybe this trick is important for convergence, we never know 😁\r\n\r\nI have a non-natural language dataset where I haven't actually been including `<s>` and `</s>` since they don't add any value (and need to be removed later anyway). To work with that, should I insert a pad token at the start of the `decoder_input` representation (and truncate to max_length)?", "> So from the paper: https://arxiv.org/pdf/1910.13461.pdf, you can see that Bart is trained on denoising input sequences in almost any possible way.\r\n> \r\n> One way could be for `BartForConditionalGeneration`:\r\n> \r\n> ```python\r\n> from transformers import BartTokenizer, BartForConditionalGeneration, BartConfig\r\n> \r\n> tok = BartTokenizer.from_pretrained(\"facebook/bart-large\")\r\n> model = BartForConditionalGeneration(BartConfig())\r\n> \r\n> input_string = \"My dog is <mask> </s>\"\r\n> decoder_input_string = \"<s> My dog is cute\"\r\n> labels_string = \"My dog is cute </s>\"\r\n> \r\n> input_ids = tok(input_string, add_special_tokens=False, return_tensors=\"pt\").input_ids\r\n> decoder_input_ids =tok(decoder_input_string, add_special_tokens=False, return_tensors=\"pt\").input_ids\r\n> labels = tok(labels_string, add_special_tokens=False, return_tensors=\"pt\").input_ids\r\n> \r\n> loss = model(input_ids=input_ids, decoder_input_ids=decoder_input_ids, labels=labels)[0]\r\n> ```\r\n\r\nHi, do you have a script to build the training dataset of BART pertain, thanks", "@patrickvonplaten @sshleifer Did anyone ever come around to creating a notebook/script for BART pretraining? (In a linked issue you mentioned it was on the to-do list.)\r\n\r\nThe core difficulty is having a canonical implementation for the data preprocessing (BART is more than just token masking, I believe: e.g.,span masking, shuffling). But a full pretrain pipeline here or in fairseq is also sorely missing. ", "Sadly not :-/ We now have on for Flax in #18297 - could you try to copy-paste the preprocessing logic into a PyTorch one maybe? ", "@patrickvonplaten I've been porting the fairseq implementation to a PyTorch dataloader format. I found that the Flax implementation in HF lacks adding noise for 0-length spans and has some slightly diverging implementation so it was more straightforward to start from the fairseq implementation. I am now especially testing the data processing to get it as close as possible to fairseq's implementation (although it is my believe that [there's a bug in their code](https://github.com/facebookresearch/fairseq/issues/4695)).\r\n\r\nI would like to add a full pytorch example for DLM training of BART in the coming days/weeks but I could use some code reviews in doing that to feel more comfortable. Would that be possible?", "Sure, happy to take a look! ", "Hi\r\n\r\nI remember posting this a year ago but I've written an entire toolkit for this purpose. Feel free to use it. https://github.com/prajdabre/yanmtt\r\n\r\nI've also created a simple notebook for the same (scroll to the pretraining part): https://colab.research.google.com/drive/1ovlA_h0ggblawqR-yCgRs3uRjxFJ8K0l?usp=sharing\r\n\r\n", "Hi Raj, thank you for this. I had come across it but your script seems to have a lot of additional things going on so that it is hard to extract the basics. I also found that you implement word/span masking but not the other things like adding noise or randomly swap a masked token for a random token, so not _completely_ like the original implementation (but correct me if I'm wrong!) .\r\n\r\nI think your library can be very useful to be used as a separate library, thanks! In addition I'll try add a PR in `transformers` for an succinct example to use within transformers with the `Trainer`, with data processing close the `fairseq` implementation.", "Hi,\r\n\r\nMy focus was more on mbart and mt5 which looked only at span masking and reordering. I'm not sure if token replacement will have that big of an impact but can be easily implemented in 1 line. To my understanding, span masking is responsible for majority of the gains. The notebook contains a more watered down version of the masking method in my toolkit. You could consider that version and build on top of it easily.", "Hey guys, I would want to know how to pre-training BART model from scratch. Anyone who know about this? BART, pegasus or other text summarization models are okay for me." ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null }, { "id": 1843244711, "node_id": "MDU6TGFiZWwxODQzMjQ0NzEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/New%20model", "name": "New model", "color": "fbca04", "default": false, "description": "" } ]

[]

[ "This is very interesting!", "This has been proposed before as a separate issue but no action was taken. Hence, I thought I'll start implementing some of the multi-modal models one by one.", "Please let @liunian-harold-li and me know if you need any help. We can also provide the pre-trained models. ", "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.\n" ]

[]

[]

[ "Hi! I don't think it's possible to have two different hardware perform *exactly* the same. A 1e-7 precision is already very high. Out of curiosity, why do you need such a high precision?\r\n\r\nNevertheless, this is more of a pytorch-specific question rather than transformers-related. I'll close the issue here but feel free to link to an issue on the PyTorch forums if you do open one there!", "@LysandreJik I haven't seen this issue with models not trained with transformers (though I probably just haven't looked hard enough).\r\n\r\nCan you give me more info on why this is the case, and maybe point me to some relevant resources?\r\n\r\nAs to your second question-- we need a high precision because we're searching for adversarial examples that maximize model misprediction in our library [TextAttack](https://github.com/QData/TextAttack). This precision error caused a different search outcome with a CPU and GPU. So, for some pair of sequences $a$ and $b$, the model on the CPU predicted $a$ 'more correctly' than it predicted $b$, and the model on the GPU predicted $b$ more correctly than $a$. Our automated tests caught this issue. \r\n\r\nDo you have a suggestion on how to fix it? Should we just truncate model scores to 7 decimal places? That feels like a crude fix." ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5093?src=pr&el=h1) Report\n> Merging [#5093](https://codecov.io/gh/huggingface/transformers/pull/5093?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/90c833870c78bb3d5d807a9a3e6a40d24bf2302b&el=desc) will **not change** coverage.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5093?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5093 +/- ##\n=======================================\n Coverage 77.28% 77.28% \n=======================================\n Files 133 133 \n Lines 22134 22134 \n=======================================\n Hits 17107 17107 \n Misses 5027 5027 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5093?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5093/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.26% <0.00%> (-0.16%)` | :arrow_down: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5093/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `75.40% <0.00%> (+0.40%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5093?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5093?src=pr&el=footer). Last update [90c8338...eae4841](https://codecov.io/gh/huggingface/transformers/pull/5093?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 2039044877, "node_id": "MDU6TGFiZWwyMDM5MDQ0ODc3", "url": "https://api.github.com/repos/huggingface/transformers/labels/marian", "name": "marian", "color": "30cc95", "default": false, "description": "" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5092?src=pr&el=h1) Report\n> Merging [#5092](https://codecov.io/gh/huggingface/transformers/pull/5092?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/20fa82898495f516b221115fc3ef9ec8ebf50b1e&el=desc) will **increase** coverage by `0.07%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5092?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5092 +/- ##\n==========================================\n+ Coverage 77.21% 77.29% +0.07% \n==========================================\n Files 133 133 \n Lines 22134 22134 \n==========================================\n+ Hits 17091 17108 +17 \n+ Misses 5043 5026 -17 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5092?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/tokenization\\_marian.py](https://codecov.io/gh/huggingface/transformers/pull/5092/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fbWFyaWFuLnB5) | `93.75% <100.00%> (+0.89%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5092/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.42% <0.00%> (+0.15%)` | :arrow_up: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5092/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `75.40% <0.00%> (+0.40%)` | :arrow_up: |\n| [src/transformers/modeling\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5092/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ194bG5ldC5weQ==) | `78.65% <0.00%> (+2.29%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5092?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5092?src=pr&el=footer). Last update [20fa828...e30eaf5](https://codecov.io/gh/huggingface/transformers/pull/5092?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "What tokenizer are you using? Is it one already trained? If so, which? If not, on what did you train your tokenizer? What is your code? What is your output?\r\n\r\nIf possible, please fill the template. It's much easier to help you if you provide all the information we need.", "1. Trained using a pretrained biobert for NER task:\r\n`tokenizer = BertTokenizer.from_pretrained(\"monologg/biobert_v1.0_pubmed_pmc\")`\r\n\r\n2. encode_plus method\r\n`encoded_dict = tokenizer.encode_plus(\r\n sent_str, # Sentence to encode.\r\n add_special_tokens = True, \r\n max_length = 75, \r\n pad_to_max_length = True,\r\n return_attention_mask = True, \r\n return_tensors = 'pt', \r\n )`\r\n3. Dataset of \r\n> linnaeus-IOB: \r\nincludes in one line `>= label O`\r\n\r\n3. Current output\r\nthe line of `>=` is split into two symbols without hashes\r\nI get \r\n`> label O`\r\n`= label O`", "Same for the bio dataset BC4CHEMD-IOBES, it has (R), and the tokenizer split them into three tokens without hashes", "Right, I fail to understand why you think this is wrongly tokenized? This tokenizer does not have the token `>=` in its vocabulary.\r\n\r\nYou can check with:\r\n\r\n```py\r\n\">\" in tokenizer.get_vocab() # Returns True\r\n\">=\" in tokenizer.get_vocab() # returns False\r\n```", "@LysandreJik yes sir, thank you\r\n\r\nBut I'm trying to understand why `encode_plus` did not add hashes `##` before `=` while tokenization\r\n\r\nI was expecting to see \r\n`>`\r\n`##=`\r\nso they would relate to the same token\r\nbut that was not the case\r\n\r\nIs my question clear?\r\n", "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.\n" ]

[]

[]

[]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[ { "login": "patrickvonplaten", "id": 23423619, "node_id": "MDQ6VXNlcjIzNDIzNjE5", "avatar_url": "https://avatars.githubusercontent.com/u/23423619?v=4", "gravatar_id": "", "url": "https://api.github.com/users/patrickvonplaten", "html_url": "https://github.com/patrickvonplaten", "followers_url": "https://api.github.com/users/patrickvonplaten/followers", "following_url": "https://api.github.com/users/patrickvonplaten/following{/other_user}", "gists_url": "https://api.github.com/users/patrickvonplaten/gists{/gist_id}", "starred_url": "https://api.github.com/users/patrickvonplaten/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/patrickvonplaten/subscriptions", "organizations_url": "https://api.github.com/users/patrickvonplaten/orgs", "repos_url": "https://api.github.com/users/patrickvonplaten/repos", "events_url": "https://api.github.com/users/patrickvonplaten/events{/privacy}", "received_events_url": "https://api.github.com/users/patrickvonplaten/received_events", "type": "User", "site_admin": false } ]

[ "Not yet sadly - it's on my ToDo list. Hope to be able to work on it soon", "I am working on a script for T5 based upon the current run_language_modeling.py, maybe I can share that once I am done and someone can confirm if it works as expected?", "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.\n", "Hi, I'm working in the same task. [Here](https://github.com/huggingface/transformers/issues/7451) you can see my code if it helps!" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null }, { "id": 1843244711, "node_id": "MDU6TGFiZWwxODQzMjQ0NzEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/New%20model", "name": "New model", "color": "fbca04", "default": false, "description": "" } ]

[]

[ "I'd like a google colab of it ", "Hey @minimaxir! Here's a [colab](https://colab.research.google.com/github/apeguero1/image-gpt/blob/master/Transformers_Image_GPT.ipynb) which loads the weights into a subclass of `GPT2LMHeadModel` and demonstrates unconditional image generation and conditional image completion. \r\n\r\nSome differences I've found between Image-GPT and GPT2 which are reflected in the subclass. \r\n\r\n1) Image-GPT layer normalization doesn't subtract off the mean\r\n2) different activations used in the MLP\r\n3) In Image-GPT, the input and output embeddings are not tied\r\n4) Image-GPT has an extra learned \"sos\" token embedding which is concatenated at the beginning of the sequence\r\n5) The GPT2 `[n_embd, 3*n_embd]` dimensional linear layer, `c_attn`, which produces queries, keys, and values is instead split into 3 separate linear layers each with dimension `[n_head, n_embd/n_head, n_embd]` in Image-GPT (this only affects how to load the weights and not the actual model).\r\n6) In Image-GPT, the `conv1d` module doesn't have a bias term\r\n\r\nSo what's our next step to add this to the repo?", "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.\n", "@apeguero1 we have an \"Adding a new model\" checklist at https://github.com/huggingface/transformers/tree/master/templates/adding_a_new_model", "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.\n" ]

[]

[]

[ "Hi, the T5 tokenizer is a SentencePiece tokenizer, and that's the way SentencePiece works. This underscore means that it's the start of a word. When it's not the start of a word, it's not prepended by anything.\r\n\r\nI think you're thinking of the # symbol because you're used to the BertTokenizer (wordpiece). SentencePiece works a bit differently!" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5086?src=pr&el=h1) Report\n> Merging [#5086](https://codecov.io/gh/huggingface/transformers/pull/5086?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/70bc3ead4f0b08e8cadd1805ada2a22f0c302399&el=desc) will **increase** coverage by `0.00%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5086?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5086 +/- ##\n=======================================\n Coverage 77.26% 77.27% \n=======================================\n Files 133 133 \n Lines 22146 22149 +3 \n=======================================\n+ Hits 17111 17115 +4 \n+ Misses 5035 5034 -1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5086?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/pipelines.py](https://codecov.io/gh/huggingface/transformers/pull/5086/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9waXBlbGluZXMucHk=) | `76.41% <100.00%> (+0.13%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5086/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.57% <0.00%> (+0.15%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5086?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5086?src=pr&el=footer). Last update [70bc3ea...ca3ad69](https://codecov.io/gh/huggingface/transformers/pull/5086?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "I think all Pipelines have this problem, not just the `summarization` pipeline, no? \r\n\r\nThe only time the `task` string is ever really used is in the `class Pipeline(_ScikitCompat):` to call the correct task specific config parameters. \r\n\r\nI would actually propose to change the `task` string logic here a bit so that we give every pipeline class a static string variable `task` and don't pass it in the init. For summarization, e.g.:\r\n\r\n```python\r\nclass SummarizationPipeline(Pipeline):\r\n\r\n task = \"summarization\"\r\n\r\n def __call__(...):\r\n ...\r\n```\r\n\r\nThen in the `Pipeline` class we could have the following logic:\r\n\r\n```python\r\nclass Pipeline:\r\n\r\n task = None\r\n\r\n def __init__(...):\r\n ...\r\n assert self.task in SUPPORTED_TASKS.values(), f\"{self.task} does not exist\"\r\n```\r\n\r\nIMO, \"task\" is not really part of the instantiated object, but more of the class itself. Also, I think a model should always only have one default configuration per task. I think T5 is already an exception in that it can handle multiple task and I don't really see why T5 for example would need two different default configs for summarization. \r\nWe can always overwrite the config parameters when calling the model, so I don't think we restrict ourselves too much with this design and having multiple default configs in the config file would quickly make them unreadable.\r\n\r\nWhat do you think @julien-c ?", "I'm not following everything here @patrickvonplaten :)\r\n\r\nMerging this like that for now but feel free to refine in the future" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5085?src=pr&el=h1) Report\n> Merging [#5085](https://codecov.io/gh/huggingface/transformers/pull/5085?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/7291ea0bff57a017e71b1ea8ec01ff19da298bf0&el=desc) will **decrease** coverage by `0.01%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5085?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5085 +/- ##\n==========================================\n- Coverage 77.24% 77.22% -0.02% \n==========================================\n Files 133 133 \n Lines 22146 22146 \n==========================================\n- Hits 17107 17103 -4 \n- Misses 5039 5043 +4 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5085?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5085/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `85.86% <0.00%> (-0.78%)` | :arrow_down: |\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5085/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `90.61% <0.00%> (-0.12%)` | :arrow_down: |\n| [src/transformers/modeling\\_transfo\\_xl.py](https://codecov.io/gh/huggingface/transformers/pull/5085/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190cmFuc2ZvX3hsLnB5) | `78.81% <0.00%> (+0.19%)` | :arrow_up: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5085/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `75.40% <0.00%> (+0.40%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5085?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5085?src=pr&el=footer). Last update [7291ea0...0092c55](https://codecov.io/gh/huggingface/transformers/pull/5085?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Hi, I agree with the typo change, but there's no need for the `from_tf` flag. `bert-base-cased` is available in both TensorFlow and PyTorch, so there's no need for the flag.", "@LysandreJik you're right! I was getting an error unless I explicitly passed `from_tf=True` yesterday, but not today. Since I can't reproduce it, I'll remove that change. " ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5084?src=pr&el=h1) Report\n> Merging [#5084](https://codecov.io/gh/huggingface/transformers/pull/5084?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/7291ea0bff57a017e71b1ea8ec01ff19da298bf0&el=desc) will **increase** coverage by `0.02%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5084?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5084 +/- ##\n==========================================\n+ Coverage 77.24% 77.26% +0.02% \n==========================================\n Files 133 133 \n Lines 22146 22146 \n==========================================\n+ Hits 17107 17112 +5 \n+ Misses 5039 5034 -5 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5084?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5084/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `90.61% <0.00%> (-0.12%)` | :arrow_down: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5084/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `75.40% <0.00%> (+0.40%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5084/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.42% <0.00%> (+0.77%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5084?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5084?src=pr&el=footer). Last update [7291ea0...e9e1ea6](https://codecov.io/gh/huggingface/transformers/pull/5084?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5083?src=pr&el=h1) Report\n> Merging [#5083](https://codecov.io/gh/huggingface/transformers/pull/5083?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/7291ea0bff57a017e71b1ea8ec01ff19da298bf0&el=desc) will **decrease** coverage by `0.39%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5083?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5083 +/- ##\n==========================================\n- Coverage 77.24% 76.84% -0.40% \n==========================================\n Files 133 133 \n Lines 22146 22146 \n==========================================\n- Hits 17107 17019 -88 \n- Misses 5039 5127 +88 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5083?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_xlm.py](https://codecov.io/gh/huggingface/transformers/pull/5083/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl94bG0ucHk=) | `56.09% <0.00%> (-19.76%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5083/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `85.71% <0.00%> (-0.94%)` | :arrow_down: |\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5083/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `90.61% <0.00%> (-0.12%)` | :arrow_down: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5083?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5083?src=pr&el=footer). Last update [7291ea0...b485684](https://codecov.io/gh/huggingface/transformers/pull/5083?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5082?src=pr&el=h1) Report\n> Merging [#5082](https://codecov.io/gh/huggingface/transformers/pull/5082?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/7291ea0bff57a017e71b1ea8ec01ff19da298bf0&el=desc) will **increase** coverage by `0.00%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5082?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5082 +/- ##\n=======================================\n Coverage 77.24% 77.25% \n=======================================\n Files 133 133 \n Lines 22146 22146 \n=======================================\n+ Hits 17107 17108 +1 \n+ Misses 5039 5038 -1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5082?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5082/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `90.61% <0.00%> (-0.12%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5082/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `86.95% <0.00%> (+0.31%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5082?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5082?src=pr&el=footer). Last update [7291ea0...cc54022](https://codecov.io/gh/huggingface/transformers/pull/5082?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "LGTM" ]

[]

[]

[ "@orestisfl thanks for raising this, I also was scratching my head as the same tokenizer.save(\"EsperBERTo\") worked for me a few days ago and not anymore. I could save the tokenizer using tokenizer.save(\"EsperBERTo/vocab.txt\"), but then I can't load it. If I try to load I get:\r\n```\r\nTypeError: sep_token not found in the vocabulary\r\n```\r\nI'm using BertWordPieceTokenizer (not ByteLevelBPETokenizer used in your example) \r\n\r\nIt worked with tokenizers version 0.7.0, I just checked - I got version 0.8.0rc1 currently installed.\r\nI'll downgrade to 0.7.0 for now. \r\n", "Was this BC intended @n1t0?", "Yes, `tokenizers` `0.8.0` introduces the full tokenizer serialization, whereas before it saved the \"model\" only (vocab.json + merges.txt for BPE). So the save method should be used like that: `.save(\"tokenizer.json\")` and it saves the entire tokenizer to a JSON file.\r\nWe need to update the Notebook to use this new serialization method, but in the meantime, the only thing needed to make it work exactly like before is to replace:\r\n```python\r\n!mkdir EsperBERTo\r\ntokenizer.save(\"EsperBERTo\")\r\n```\r\nby\r\n```python\r\n!mkdir EsperBERTo\r\ntokenizer.save_model(\"EsperBERTo\")\r\n```", "mind updating it before we forget? Thanks!", "Sure, updated it with the quick change I mentioned. Will do a better update later.", "Hey there, thanks for the quick fix!\r\nThe notebook now crashes for me during training, however:\r\n```\r\n---------------------------------------------------------------------------\r\nAttributeError Traceback (most recent call last)\r\n<ipython-input-19-0c647bc3a8b8> in <module>()\r\n----> 1 get_ipython().run_cell_magic('time', '', 'trainer.train()')\r\n\r\n11 frames\r\n<decorator-gen-60> in time(self, line, cell, local_ns)\r\n\r\n<timed eval> in <module>()\r\n\r\n/usr/local/lib/python3.6/dist-packages/transformers/data/data_collator.py in <listcomp>(.0)\r\n 112 probability_matrix = torch.full(labels.shape, self.mlm_probability)\r\n 113 special_tokens_mask = [\r\n--> 114 self.tokenizer.get_special_tokens_mask(val, already_has_special_tokens=True) for val in labels.tolist()\r\n 115 ]\r\n 116 probability_matrix.masked_fill_(torch.tensor(special_tokens_mask, dtype=torch.bool), value=0.0)\r\n\r\nAttributeError: 'RobertaTokenizerFast' object has no attribute 'get_special_tokens_mask'\r\n```\r\n\r\nLet me know if I should make a separate issue ", "This one is for me (this method was actually not working as intended under the hood for Fast-tokenizers...)", "@thomwolf - just to confirm, I tried the change you made and it fixes a problem for me. Thanks!\r\n```\r\nAttributeError: 'BertTokenizerFast' object has no attribute 'get_special_tokens_mask'\r\n```" ]

[]

[]

[ "Great thanks @patil-suraj ", "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5080?src=pr&el=h1) Report\n> Merging [#5080](https://codecov.io/gh/huggingface/transformers/pull/5080?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/ebab096e864a619717a497089d864d10e21bc536&el=desc) will **increase** coverage by `0.00%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5080?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5080 +/- ##\n=======================================\n Coverage 77.26% 77.27% \n=======================================\n Files 128 128 \n Lines 21854 21854 \n=======================================\n+ Hits 16886 16887 +1 \n+ Misses 4968 4967 -1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5080?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5080/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `73.38% <0.00%> (-0.41%)` | :arrow_down: |\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5080/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `90.61% <0.00%> (+0.11%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5080/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.42% <0.00%> (+0.15%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5080?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5080?src=pr&el=footer). Last update [ebab096...2f8a93c](https://codecov.io/gh/huggingface/transformers/pull/5080?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "Hi, @abhisheknovoic this might help you https://huggingface.co/transformers/model_doc/t5.html#training\r\ncheck the Unsupervised denoising training section", "@patil-suraj , do you mean this class? - T5ForConditionalGeneration \r\n\r\nAlso, at the top of the page, there is the following code:\r\n\r\n```input_ids = tokenizer.encode('The <extra_id_1> walks in <extra_id_2> park', return_tensors='pt')\r\nlm_labels = tokenizer.encode('<extra_id_1> cute dog <extra_id_2> the <extra_id_3> </s>', return_tensors='pt')\r\n# the forward function automatically creates the correct decoder_input_ids\r\nmodel(input_ids=input_ids, lm_labels=lm_labels)\r\n```\r\n\r\nAny idea which class is the model instantiated from? I could not find any class with lm_labels parameter.\r\n\r\nThanks", "Yes, it's `T5ForConditionalGeneration`, and `lm_lables` is now changed to `labels`.\r\n\r\nPinging @patrickvonplaten for more details.", "@patil-suraj , I tried the following code which throws an error. Any idea why? Thanks\r\n\r\n```In [32]: from transformers import T5Tokenizer, T5ForConditionalGeneration, T5Config\r\nIn [32]: from transformers import T5Tokenizer, T5ForConditionalGeneration, T5Config\r\n\r\nIn [33]: input_ids = tokenizer.encode('The <extra_id_1> walks in <extra_id_2> park', return_tensors='pt')\r\n\r\nIn [34]: labels = tokenizer.encode('<extra_id_1> cute dog <extra_id_2> the <extra_id_3> </s>', return_tensors='pt')\r\n\r\nIn [35]: config = T5Config()\r\n\r\nIn [36]: model = T5ForConditionalGeneration(config=config)\r\n\r\nIn [37]: model(input_ids=input_ids, lm_labels=labels)\r\n---------------------------------------------------------------------------\r\nAssertionError Traceback (most recent call last)\r\n<ipython-input-37-6717b0ecfbf5> in <module>\r\n----> 1 model(input_ids=input_ids, lm_labels=labels)\r\n\r\n/usr/local/lib/python3.7/site-packages/torch/nn/modules/module.py in __call__(self, *input, **kwargs)\r\n 530 result = self._slow_forward(*input, **kwargs)\r\n 531 else:\r\n--> 532 result = self.forward(*input, **kwargs)\r\n 533 for hook in self._forward_hooks.values():\r\n 534 hook_result = hook(self, input, result)\r\n\r\n/usr/local/lib/python3.7/site-packages/transformers/modeling_t5.py in forward(self, input_ids, attention_mask, encoder_outputs, decoder_input_ids, decoder_attention_mask, decoder_past_key_value_states, use_cache, lm_labels, inputs_embeds, decoder_inputs_embeds, head_mask)\r\n 1068 if lm_labels is not None and decoder_input_ids is None and decoder_inputs_embeds is None:\r\n 1069 # get decoder inputs from shifting lm labels to the right\r\n-> 1070 decoder_input_ids = self._shift_right(lm_labels)\r\n 1071\r\n 1072 # If decoding with past key value states, only the last tokens\r\n\r\n/usr/local/lib/python3.7/site-packages/transformers/modeling_t5.py in _shift_right(self, input_ids)\r\n 609 assert (\r\n 610 decoder_start_token_id is not None\r\n--> 611 ), \"self.model.config.decoder_start_token_id has to be defined. In T5 it is usually set to the pad_token_id. See T5 docs for more information\"\r\n 612\r\n 613 # shift inputs to the right\r\n\r\nAssertionError: self.model.config.decoder_start_token_id has to be defined. In T5 it is usually set to the pad_token_id. See T5 docs for more information\r\n```\r\n\r\nMy versions are \r\n```\r\ntransformers==2.11.0\r\ntokenizers==0.7.0\r\n```", "If you are using 2.11.0 then use `lm_labels` and if you are using master then use `labels`", "@patil-suraj , thanks. I have installed the master version. It still complains with the same error. It seems like I need to specify something for the decoder_start_token_id. ", "Ok, I got it working. I initialized config like follows:\r\n\r\n```\r\nconfig = T5Config(decoder_start_token_id=tokenizer.convert_tokens_to_ids(['<pad>'])[0])\r\n```", "@patil-suraj , however, if we use the master branch, it seems like the tokenizers are broken. The T5 tokenizer doesn't tokenize the sentinel tokens correctly.", "> @patil-suraj , do you mean this class? - T5ForConditionalGeneration\r\n> \r\n> Also, at the top of the page, there is the following code:\r\n> \r\n> ```\r\n> lm_labels = tokenizer.encode('<extra_id_1> cute dog <extra_id_2> the <extra_id_3> </s>', return_tensors='pt')\r\n> # the forward function automatically creates the correct decoder_input_ids\r\n> model(input_ids=input_ids, lm_labels=lm_labels)\r\n> ```\r\n> \r\n> Any idea which class is the model instantiated from? I could not find any class with lm_labels parameter.\r\n> \r\n> Thanks\r\n\r\nFeel free to also open a PR to correct `lm_labels` to `labels` in the comment :-) ", "Just saw that @patil-suraj already did this - awesome thanks :-) \r\n\r\n@abhisheknovoic regarding the T5 tokenizer, can you post some code here that shows that T5 tokenization is broken (would be great if we can easily reproduce the error)", "@patrickvonplaten it would be nice if we also add seq-2-seq (t5, bart) model pre-training examples in official examples \r\n\r\ncc @sshleifer ", "Definitely!", "Not sure if this should be a separate issue or not, but I am having difficulty training my own T5 tokenizer. When training a BPE tokenizer using the amazing huggingface tokenizer library and attempting to load it via\r\n\r\n```python\r\ntokenizer = T5Tokenizer.from_pretrained('./tokenizer')\r\n```\r\nI get the following error:\r\n```\r\nOSError: Model name './tokenizer/' was not found in tokenizers model name list (t5-small, t5-base, t5-large, t5-3b, t5-11b). We assumed './tokenizer/' was a path, a model identifier, or url to a directory containing vocabulary files named ['spiece.model'] but couldn't find such vocabulary files at this path or url.\r\n```\r\n\r\nI attempted to train a sentencepiece model instead using the, again amazing, huggingface tokenizer library, I get the same error because the `tokenizer.save` method does not actual generate the `spiece.model` file.\r\n\r\nAm I doing something wrong?\r\n\r\nTranformers version: 2.11.0\r\nTokenizers version: 0.7.0\r\n\r\nHere is a colab to reproduce the error: https://colab.research.google.com/drive/1WX1Q2Ze9k0SxFMLLv1aFgVGBFMEVTyDe?usp=sharing", "@mfuntowicz @n1t0 - maybe you can help here", "> Definitely!\r\n\r\nThe pre-training scripts would really help.original mesh transformer is very complicated to understand.", "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.\n", "We've released [nanoT5](https://github.com/PiotrNawrot/nanoT5) that reproduces T5-model (similar to BART) pre-training in PyTorch (not Flax). \r\n\r\nYou can take a look! \r\n\r\nAny suggestions are more than welcome." ]

[ { "id": 1834052574, "node_id": "MDU6TGFiZWwxODM0MDUyNTc0", "url": "https://api.github.com/repos/huggingface/transformers/labels/Ex:%20Sequence%20Classification", "name": "Ex: Sequence Classification", "color": "46FFCF", "default": false, "description": "" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5078?src=pr&el=h1) Report\n> Merging [#5078](https://codecov.io/gh/huggingface/transformers/pull/5078?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/e4aaa4580515446cd5a2972ab42fec0b95819c84&el=desc) will **increase** coverage by `0.00%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5078?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5078 +/- ##\n=======================================\n Coverage 77.26% 77.26% \n=======================================\n Files 133 133 \n Lines 22146 22146 \n=======================================\n+ Hits 17110 17111 +1 \n+ Misses 5036 5035 -1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5078?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5078/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `75.40% <0.00%> (+0.40%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5078?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5078?src=pr&el=footer). Last update [e4aaa45...8e0cf02](https://codecov.io/gh/huggingface/transformers/pull/5078?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Hi @LysandreJik , my worry is that PABEE is not a standard inference method. It cannot deal with batch inference. Also, it can only support classification & regression (no tagging, no summarization, etc.). \r\n\r\nFrom another aspect, as a researcher, when I try to do a little tweak with ALBERT, I won't be happy if someone adds some new stuff into the model and it'll add unnecessary burdens to the researchers. They'll definitely hate it. So IMO, it's better to implement them separately. Also, as @sshleifer suggested, I refactored the model to inherit `AlbertTransformer` and `AlbertModel`, etc.", "> Hi @LysandreJik , my worry is that PABEE is not a standard inference method. It cannot deal with batch inference. Also, it can only support classification & regression (no tagging, no summarization, etc.).\r\n> \r\n> From another aspect, as a researcher, when I try to do a little tweak with ALBERT, I won't be happy if someone adds some new stuff into the model and it'll add unnecessary burdens to the researchers. They'll definitely hate it. So IMO, it's better to implement them separately. Also, as @sshleifer suggested, I refactored the model to inherit `AlbertTransformer` and `AlbertModel`, etc.\r\n\r\nSince I refactored the code with inheritance, I figure it is okay to use `adaptive_forward` since I won't have to overwrite the standard `forward` (which would be confusing since it's hard to tell which part is modified for the users). Also, it's better to preserve the standard `forward` so we can easily compare `adaptive_forward` to the standard `forward`.\r\n\r\n@sshleifer I copy most stuff from the original ALBERT & BERT modeling code so I think it also does not make sense if I refactor the parts I copied.\r\n\r\nRe. trainer, maybe we can refactor the code later? I think it's quite optional here but it requires a lot of changes on `run_glue_with_pabee.py`. Some part of me wonders if it is worthy.", "The code you copied has gotten cleaned up since you copied it, hence the suggestions.", "> LGTM pending suggestions, test.\n\nOkay, I’ll add a test" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "@Nighthyst thanks for summarizing these issues as I also ran into them.\r\nI was digging on this last weekend and I think maybe this could help:\r\n\r\nhttps://github.com/huggingface/tokenizers/pull/200\r\n\r\n\r\n>> Provide some more mappings on the Encoding in order to easily identify words after tokenization.\r\n\r\n>> It also exposes a method encode_tokenized on the BaseTokenizer to allow skipping the usual Normalizer and PreTokenizer.\r\nThis is especially useful for NER like datasets, where the pre-tokenization has already been done, and we want to attribute labels to pre-tokenized words.\r\n", "Thanks for bringing this up. I can work on this on a separate PR after merging the PR that resolves the prior issue #4816.", "some interesting finding:\r\n\r\nUsing a fast tokenizer solves the `[UNK]` issue. using one of your provided examples:\r\n\r\n```python\r\nmodel = AutoModelForTokenClassification.from_pretrained(\"dbmdz/bert-large-cased-finetuned-conll03-english\")\r\ntokenizer = AutoTokenizer.from_pretrained(\"bert-base-cased\", use_fast=True)\r\nnlp = TokenClassificationPipeline(model=model,\r\n tokenizer=tokenizer,\r\n grouped_entities=False)\r\n\r\nt=\"Product sales to the PSA Peugeot Citroën group totaled € 1 , 893 . 6 million in 2012 , down 8 . 1 % on a reported basis and 10 . 4 % on a like - for - like basis .\"\r\n\r\nnlp(t)\r\n```\r\n\r\n```\r\n[{'word': 'PS', 'score': 0.9961145520210266, 'entity': 'I-ORG', 'index': 5},\r\n {'word': '##A', 'score': 0.9905584454536438, 'entity': 'I-ORG', 'index': 6},\r\n {'word': 'P', 'score': 0.997616708278656, 'entity': 'I-ORG', 'index': 7},\r\n {'word': '##eu', 'score': 0.9741767644882202, 'entity': 'I-ORG', 'index': 8},\r\n {'word': '##ge', 'score': 0.9928027391433716, 'entity': 'I-ORG', 'index': 9},\r\n {'word': '##ot', 'score': 0.9900722503662109, 'entity': 'I-ORG', 'index': 10},\r\n {'word': 'C', 'score': 0.9574489593505859, 'entity': 'I-ORG', 'index': 11},\r\n {'word': '##it', 'score': 0.824583113193512, 'entity': 'I-ORG', 'index': 12},\r\n {'word': '##ro', 'score': 0.7597800493240356, 'entity': 'I-ORG', 'index': 13},\r\n {'word': '##A', 'score': 0.953075647354126, 'entity': 'I-ORG', 'index': 14},\r\n {'word': '«', 'score': 0.6135829091072083, 'entity': 'I-ORG', 'index': 15}]\r\n```", "@Nighthyst @dav009 Can you guys check if the above issues still persist after the recent PR merged (#4987)?", "Hello @enzoampil, \r\n\r\nI updated transformers with master, with the command: \r\n\r\n`pip install --upgrade git+https://github.com/huggingface/transformers.git`\r\n\r\nThen I tried your tests and mine:\r\n\r\n```Python\r\nfrom transformers import pipeline\r\nNER_MODEL = \"mrm8488/bert-spanish-cased-finetuned-ner\"\r\nnlp_ner = pipeline(\"ner\", model=NER_MODEL,\r\n grouped_entities=True,\r\n tokenizer=(NER_MODEL, {\"use_fast\": False}))\r\n\r\nt = \"\"\"Consuelo Araújo Noguera, ministra de cultura del presidente Andrés Pastrana (1998.2002) fue asesinada por las Farc luego de haber permanecido secuestrada por algunos meses.\"\"\"\r\nnlp_ner(t)\r\n```\r\n\r\nI have the expected output : \r\n\r\n```\r\n[{'entity_group': 'B-PER',\r\n 'score': 0.9710702555520194,\r\n 'word': 'Consuelo Araújo Noguera'},\r\n {'entity_group': 'B-PER',\r\n 'score': 0.9997273534536362,\r\n 'word': 'Andrés Pastrana'},\r\n {'entity_group': 'B-ORG', 'score': 0.8589079678058624, 'word': 'Farc'}]\r\n```\r\n\r\nAnd for your other test : \r\n\r\n```Python\r\nnlp = pipeline('ner', grouped_entities=False)\r\nnlp(\"Enzo works at the the UN\")\r\n```\r\n\r\nOutput : \r\n\r\n```\r\n[{'word': 'En', 'score': 0.9968166351318359, 'entity': 'I-PER', 'index': 1},\r\n {'word': '##zo', 'score': 0.9957635998725891, 'entity': 'I-PER', 'index': 2},\r\n {'word': 'UN', 'score': 0.9986497163772583, 'entity': 'I-ORG', 'index': 7}]\r\n```\r\n\r\nAnd,\r\n\r\n```Python\r\nnlp2 = pipeline('ner', grouped_entities=True)\r\nnlp2(\"Enzo works at the the UN\")\r\n```\r\n\r\nOutput : \r\n\r\n```\r\n{'entity_group': 'I-PER', 'score': 0.9962901175022125, 'word': 'Enzo'},\r\n {'entity_group': 'I-ORG', 'score': 0.9986497163772583, 'word': 'UN'}]\r\n```\r\n\r\nHowever with my test : \r\n\r\n```Python\r\nimport torch\r\nfrom transformers import AutoModelForTokenClassification, AutoTokenizer\r\nfrom transformers import TokenClassificationPipeline\r\n\r\nmodel = AutoModelForTokenClassification.from_pretrained(\"dbmdz/bert-large-cased-finetuned-conll03-english\")\r\ntokenizer = AutoTokenizer.from_pretrained(\"bert-base-cased\")\r\n\r\nnlp_not_grouped = TokenClassificationPipeline(\r\n model=model,\r\n tokenizer=tokenizer,\r\n grouped_entities=False\r\n)\r\n\r\nnlp_grouped = TokenClassificationPipeline(\r\n model=model,\r\n tokenizer=tokenizer,\r\n grouped_entities=True\r\n)\r\n\r\nseq1 = \"Hugging Face Inc. is a company based in New York City. Its headquarters are in DUMBO, therefore very\" \\\r\n \"close to the Manhattan Bridge.\"\r\n\r\nseq2 = \"In addition , the Blabla Group has completed the acquisition of ISO / TS16949 certification .\"\r\n\r\nseq3 = \"Product sales to the PSA Peugeot Citroën group totaled € 1 , 893 . 6 million in 2012 , down 8 . 1 % \"\\\r\n\"on a reported basis and 10 . 4 % on a like - for - like basis .\"\r\n\r\nseq4 = \"To prepare as best as possible the decisions falling under its responsibilities , Faurecia ’ s Board of\"\\\r\n\" Directors has set up three committees : c Audit Committee ; c Strategy Committee ; c Appointments and Compensation\"\\\r\n\" Committee .\"\r\n\r\nsequences = [seq1, seq2, seq3, seq4]\r\n\r\nfor i, seq in enumerate(sequences):\r\n ngrouped, grouped = nlp_not_grouped(seq), nlp_grouped(seq)\r\n print(f\"===================== sentence n°{i+1}\")\r\n print(\"---Sentence---\")\r\n print(seq)\r\n print(\"---Not grouped entities---\")\r\n for ngent in ngrouped:\r\n print(ngent)\r\n print(\"---Grouped entities---\")\r\n for gent in grouped:\r\n print(gent)\r\n```\r\n\r\nI have this : \r\n\r\n```\r\n===================== sentence n°1\r\n---Sentence---\r\nHugging Face Inc. is a company based in New York City. Its headquarters are in DUMBO, therefore veryclose to the Manhattan Bridge.\r\n---Not grouped entities---\r\n{'word': 'Hu', 'score': 0.9995108246803284, 'entity': 'I-ORG', 'index': 1}\r\n{'word': '##gging', 'score': 0.989597499370575, 'entity': 'I-ORG', 'index': 2}\r\n{'word': 'Face', 'score': 0.9979704022407532, 'entity': 'I-ORG', 'index': 3}\r\n{'word': 'Inc', 'score': 0.9993758797645569, 'entity': 'I-ORG', 'index': 4}\r\n{'word': 'New', 'score': 0.9993405938148499, 'entity': 'I-LOC', 'index': 11}\r\n{'word': 'York', 'score': 0.9991927742958069, 'entity': 'I-LOC', 'index': 12}\r\n{'word': 'City', 'score': 0.9993411302566528, 'entity': 'I-LOC', 'index': 13}\r\n{'word': 'D', 'score': 0.986336350440979, 'entity': 'I-LOC', 'index': 19}\r\n{'word': '##UM', 'score': 0.9396238923072815, 'entity': 'I-LOC', 'index': 20}\r\n{'word': '##BO', 'score': 0.9121386408805847, 'entity': 'I-LOC', 'index': 21}\r\n{'word': 'Manhattan', 'score': 0.9839190244674683, 'entity': 'I-LOC', 'index': 29}\r\n{'word': 'Bridge', 'score': 0.9924242496490479, 'entity': 'I-LOC', 'index': 30}\r\n---Grouped entities---\r\n{'entity_group': 'I-ORG', 'score': 0.9966136515140533, 'word': 'Hugging Face Inc'}\r\n{'entity_group': 'I-LOC', 'score': 0.9992914994557699, 'word': 'New York City'}\r\n{'entity_group': 'I-LOC', 'score': 0.9460329612096151, 'word': 'DUMBO'}\r\n{'entity_group': 'I-LOC', 'score': 0.9881716370582581, 'word': 'Manhattan Bridge'}\r\n===================== sentence n°2\r\n---Sentence---\r\nIn addition , the Blabla Group has completed the acquisition of ISO / TS16949 certification .\r\n---Not grouped entities---\r\n{'word': 'B', 'score': 0.9997261762619019, 'entity': 'I-ORG', 'index': 5}\r\n{'word': '##la', 'score': 0.997683048248291, 'entity': 'I-ORG', 'index': 6}\r\n{'word': '##bla', 'score': 0.99888014793396, 'entity': 'I-ORG', 'index': 7}\r\n{'word': 'Group', 'score': 0.9992784261703491, 'entity': 'I-ORG', 'index': 8}\r\n{'word': 'ISO', 'score': 0.9711909890174866, 'entity': 'I-MISC', 'index': 14}\r\n{'word': 'T', 'score': 0.6591967344284058, 'entity': 'I-ORG', 'index': 16}\r\n{'word': '##S', 'score': 0.658642053604126, 'entity': 'I-MISC', 'index': 17}\r\n{'word': '##16', 'score': 0.5059574842453003, 'entity': 'I-MISC', 'index': 18}\r\n{'word': '##9', 'score': 0.5067382454872131, 'entity': 'I-MISC', 'index': 21}\r\n---Grouped entities---\r\n{'entity_group': 'I-ORG', 'score': 0.9988919496536255, 'word': 'Blabla Group'}\r\n{'entity_group': 'I-MISC', 'score': 0.9711909890174866, 'word': 'ISO'}\r\n{'entity_group': 'I-ORG', 'score': 0.6591967344284058, 'word': 'T'}\r\n{'entity_group': 'I-MISC', 'score': 0.5822997689247131, 'word': '##S16'}\r\n{'entity_group': 'I-MISC', 'score': 0.5067382454872131, 'word': '##9'}\r\n===================== sentence n°3\r\n---Sentence---\r\nProduct sales to the PSA Peugeot Citroën group totaled € 1 , 893 . 6 million in 2012 , down 8 . 1 % on a reported basis and 10 . 4 % on a like - for - like basis .\r\n---Not grouped entities---\r\n{'word': 'PS', 'score': 0.9970256686210632, 'entity': 'I-ORG', 'index': 5}\r\n{'word': '##A', 'score': 0.9927457571029663, 'entity': 'I-ORG', 'index': 6}\r\n{'word': 'P', 'score': 0.9980151653289795, 'entity': 'I-ORG', 'index': 7}\r\n{'word': '##eu', 'score': 0.9897757768630981, 'entity': 'I-ORG', 'index': 8}\r\n{'word': '##ge', 'score': 0.996147871017456, 'entity': 'I-ORG', 'index': 9}\r\n{'word': '##ot', 'score': 0.9928787350654602, 'entity': 'I-ORG', 'index': 10}\r\n{'word': '[UNK]', 'score': 0.5744695067405701, 'entity': 'I-ORG', 'index': 11}\r\n---Grouped entities---\r\n{'entity_group': 'I-ORG', 'score': 0.934436925819942, 'word': 'PSA Peugeot [UNK]'}\r\n===================== sentence n°4\r\n---Sentence---\r\nTo prepare as best as possible the decisions falling under its responsibilities , Faurecia ’ s Board of Directors has set up three committees : c Audit Committee ; c Strategy Committee ; c Appointments and Compensation Committee .\r\n---Not grouped entities---\r\n{'word': 'F', 'score': 0.9983997941017151, 'entity': 'I-ORG', 'index': 14}\r\n{'word': '##au', 'score': 0.9473735690116882, 'entity': 'I-ORG', 'index': 15}\r\n{'word': '##re', 'score': 0.9604568481445312, 'entity': 'I-ORG', 'index': 16}\r\n{'word': '##cia', 'score': 0.992807149887085, 'entity': 'I-ORG', 'index': 17}\r\n{'word': 'Board', 'score': 0.8452167510986328, 'entity': 'I-ORG', 'index': 20}\r\n{'word': 'of', 'score': 0.5921975374221802, 'entity': 'I-ORG', 'index': 21}\r\n{'word': 'Directors', 'score': 0.6778028607368469, 'entity': 'I-ORG', 'index': 22}\r\n{'word': 'Audi', 'score': 0.9764850735664368, 'entity': 'I-ORG', 'index': 30}\r\n{'word': '##t', 'score': 0.9692177772521973, 'entity': 'I-ORG', 'index': 31}\r\n{'word': 'Committee', 'score': 0.9959701299667358, 'entity': 'I-ORG', 'index': 32}\r\n{'word': 'Strategy', 'score': 0.9705951809883118, 'entity': 'I-ORG', 'index': 35}\r\n{'word': 'Committee', 'score': 0.994032621383667, 'entity': 'I-ORG', 'index': 36}\r\n{'word': 'A', 'score': 0.9764854907989502, 'entity': 'I-ORG', 'index': 39}\r\n{'word': '##oint', 'score': 0.7803319692611694, 'entity': 'I-ORG', 'index': 41}\r\n{'word': '##ments', 'score': 0.7828453779220581, 'entity': 'I-ORG', 'index': 42}\r\n{'word': 'and', 'score': 0.9625542163848877, 'entity': 'I-ORG', 'index': 43}\r\n{'word': 'Co', 'score': 0.9904180765151978, 'entity': 'I-ORG', 'index': 44}\r\n{'word': '##mp', 'score': 0.9140805602073669, 'entity': 'I-ORG', 'index': 45}\r\n{'word': '##ens', 'score': 0.8661588430404663, 'entity': 'I-ORG', 'index': 46}\r\n{'word': '##ation', 'score': 0.9150537252426147, 'entity': 'I-ORG', 'index': 47}\r\n{'word': 'Committee', 'score': 0.9888517260551453, 'entity': 'I-ORG', 'index': 48}\r\n---Grouped entities---\r\n{'entity_group': 'I-ORG', 'score': 0.9747593402862549, 'word': 'Faurecia'}\r\n{'entity_group': 'I-ORG', 'score': 0.7050723830858866, 'word': 'Board of Directors'}\r\n{'entity_group': 'I-ORG', 'score': 0.9805576602617899, 'word': 'Audit Committee'}\r\n{'entity_group': 'I-ORG', 'score': 0.9823139011859894, 'word': 'Strategy Committee'}\r\n{'entity_group': 'I-ORG', 'score': 0.9764854907989502, 'word': 'A'}\r\n{'entity_group': 'I-ORG', 'score': 0.9000368118286133, 'word': '##ointments and Compensation Committee'}\r\n```\r\n\r\nIt seems like the problem is still here for sentence n°4 : the last group should be \"Appointments and Compensation Committee\". For sentence n°2 it should be : \"TS16949\" as MISC or ORG at least it predicts the T in ORG and the other part in MISC. Even if both parts don't have the same entity tag, the ORG part should have been in one group \"S16949\" at least I think.\r\n\r\nAlso @dav009 \"trick\" to solve the [UNK] issue seems to not be working anymore : \r\n\r\n```Python\r\nmodel = AutoModelForTokenClassification.from_pretrained(\"dbmdz/bert-large-cased-finetuned-conll03-english\")\r\ntokenizer = AutoTokenizer.from_pretrained(\"bert-base-cased\", use_fast=True)\r\nnlp = TokenClassificationPipeline(model=model,\r\n tokenizer=tokenizer,\r\n grouped_entities=False)\r\n\r\nt=\"Product sales to the PSA Peugeot Citroën group totaled € 1 , 893 . 6 million in 2012 , down 8 . 1 % on a reported basis and 10 . 4 % on a like - for - like basis .\"\r\n\r\nnlp(t)\r\n```\r\n\r\nOutput : \r\n```\r\n\r\n[{'word': 'PS', 'score': 0.9970256686210632, 'entity': 'I-ORG', 'index': 5},\r\n {'word': '##A', 'score': 0.9927457571029663, 'entity': 'I-ORG', 'index': 6},\r\n {'word': 'P', 'score': 0.9980151653289795, 'entity': 'I-ORG', 'index': 7},\r\n {'word': '##eu', 'score': 0.9897757768630981, 'entity': 'I-ORG', 'index': 8},\r\n {'word': '##ge', 'score': 0.996147871017456, 'entity': 'I-ORG', 'index': 9},\r\n {'word': '##ot', 'score': 0.9928787350654602, 'entity': 'I-ORG', 'index': 10},\r\n {'word': '[UNK]',\r\n 'score': 0.5744695067405701,\r\n 'entity': 'I-ORG',\r\n 'index': 11}]\r\n```\r\n\r\nThe [UNK] token is back", "For sentence 4, this is because the ##pp in “Appointments”, is not being tagged as an entity. This will require a separate PR that assumes that all the word pieces attached to a tagged entity token, should also be tagged with the same entity, whether or not it was tagged.", "A similar situation is happening in sentence 2. The clue is in the value for “index”. You’ll notice that the tokens aren’t contiguous and so aren’t being grouped together. This implies that some middle word pieces aren’t being tagged as entities.", "For the [UNK] issue, this “might” be because that word piece token was out of vocabulary and so gets converted to [UNK] at the decoding step.\n\nSince this happens before entity grouping, I think safe to say this is unrelated to entity grouping and is related to how the raw NER forward pass is handled.\n\nPerhaps we can separate this from the above issue? Both will require separate PR’s to address.", "Actually you're right it seems that sentences n°2 and n°4 are showing a different issue : if the index is not contiguous (because a part is missing in the prediction : \"pp\" for n°4 and \"94\" for n°2) then the grouping fails. It's indeed a different issue.", "> For sentence 4, this is because the ##pp in “Appointments”, is not being tagged as an entity. This will require a separate PR that assumes that all the word pieces attached to a tagged entity token, should also be tagged with the same entity, whether or not it was tagged.\r\n\r\nAlthough I agree that it could be solved in a next PR, shouldn't this more 'holistic' view be preferable (and be the default). If one token in a word is 'missed' but the other four (e.g. PER-PER-O-PER-PER) are an entity the whole word is an entity (and not two separate entities). We 'know' what the word-level comprehends the model doesn't", "@HHoofs agree that this should be the default. If the \"word-level\" implementation is submitted as a PR, this should not be the default behaviour and should be explicitly set.", "I agree with that, what I meant however was the following case: `Italy`\nLet's say that this consists of three subtokens: `_It`, `a`, `ly`\nIf the first and last tokens are assigned as Country en the middle as None, it would now result in a splitted output (if I understand correctly).\nI would suggest that the outputs of all three subtokens are averaged and than the highest output class is selected.", "In pseudo-code, I would suggest the following (order):\r\n```\r\n...\r\n# first check if the user want to have grouped entities\r\nif self.grouped_entities:\r\n word_scores = []\r\n for token in tokens:\r\n # first input should always be a 'new word'\r\n if is_new_word(token):\r\n word_scores.append(score)\r\n score = np.zeros((0,?))\r\n score = np.sum(score, token['score'])\r\n # now you have a list of summed entity scores for each seperate word\r\n word_scores.argmax(axis=-1)\r\n ...\r\n \r\nelse:\r\n return ...\r\n", "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.\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "The same thing for me. Colab (using TPU) crashes when importing transformers package.\r\n\r\n\r\n\r\n", "Same issue.", "I'm trying to reproduce, but can't manage to make colab crash. @khalilRhouma, @amitness, did you have similar code to @antoniomastro1996? Would it be possible for you to show me the code you used?", "@LysandreJik yes, of course you can follow this:\r\nhttps://colab.research.google.com/drive/1jwXgtOXE8v8_qkiOCbjFQRFC5semK8T7?usp=sharing", "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.\n" ]

[]

[]

[ "Here is a related PyTorch issue: https://github.com/pytorch/pytorch/issues/32968.\r\n\r\nHere is a list of pretraineds model that can be exported to ONNX:\r\nbert-base-cased, distilbert-base-uncased, roberta-base, gpt2, distilgpt2, openai-gpt, albert-base-v2, xlnet-base-cased\r\n\r\nSo far, the following models have problem in exporting to ONNX:\r\nbart-large, transfo-xl-wt103, t5-base, xlm-mlm-en-2048\r\n\r\n\r\n\r\n", "Anyone had any luck resolving this issue? The Pytorch issue linked above has a couple of potential workarounds. I notice it affects Pegasus as well", "This pull request was closed by its author but doesn't raise the issue when I run convert on a Bart model. https://github.com/huggingface/transformers/pull/6334\r\n", "Can a custom distilbert model which is saved in local disk be converted to onnx?", "Can we convert Pix2Struct model into onnx i tried to do so manually it always show some error" ]

[]

[]

[ "They are the same as Roberta's \r\n\r\n```python\r\nvocab_url = \"https://s3.amazonaws.com/models.huggingface.co/bert/roberta-large-vocab.json\"\r\nmerges_url = \"https://s3.amazonaws.com/models.huggingface.co/bert/roberta-large-merges.txt\"\r\n```\r\n\r\nIs your question more general than just `bart-large`? (Feel free to close if not :) )", "Thanks and yes. I'm interested in a more general case to retrieve the URLs to weights for different models.\r\nCurrently, what I did actually is switching back to 2.10.0. And go to corresponding `modeling_xxx.py` and find the download link.", "You can find URLs to specific files on each model page (click on \"List all files\"): https://huggingface.co/distilbert-base-cased#list-files", "Hi @julien-c , as mentioned, I actually know this way, but \"list all files\" seems not giving me the `vocab.json`, `merge.txt`, OR are they not required? ", "Not all tokenizer types use those files. For instance Wordpiece (e.g. bert) is just one vocab.txt file.", "Yup. I understand that, just named an example. But just want to make sure that files in https://huggingface.co/distilbert-base-cased#list-files are sufficient for me to use. \r\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5073?src=pr&el=h1) Report\n> Merging [#5073](https://codecov.io/gh/huggingface/transformers/pull/5073?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/e4aaa4580515446cd5a2972ab42fec0b95819c84&el=desc) will **increase** coverage by `0.00%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5073?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5073 +/- ##\n=======================================\n Coverage 77.26% 77.26% \n=======================================\n Files 133 133 \n Lines 22146 22146 \n=======================================\n+ Hits 17110 17111 +1 \n+ Misses 5036 5035 -1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5073?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5073/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.42% <0.00%> (+0.15%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5073?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5073?src=pr&el=footer). Last update [e4aaa45...c2c5e07](https://codecov.io/gh/huggingface/transformers/pull/5073?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Thanks for catching it." ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "You are right, the number of optimization steps is already fixed in an another PR (#5051). And indeed the batch size for TPUs has to take into account the gradient accumulation, but I put this for later because we don't have yet a proper way to differenciate a run on TPU/CPU/GPU. Also the `unbatch` then `batch` is not needed. Same thing apply later when computing the loss in the training step.", "Thanks for your answer @jplu !\r\n\r\nCan I ask you clarification about this :\r\n\r\n>Also the unbatch then batch is not needed. Same thing apply later when computing the loss in the training step.\r\n\r\nWe still need the data to be batched based on the batch_size_per_device, not the total batch size, right ?\r\nIs the `floor` / `ceil` needed ?\r\n\r\nIs there any other changes I should apply locally to make it work ?", "I'm pretty sure that the TPUs as to be set for the full size of batches (including those with the accumulation).\r\n\r\n| Is the floor / ceil needed ?\r\n`floor` and `ceil` are also needed in order to be sure you get the proper approximation.\r\n\r\n| Is there any other changes I should apply locally to make it work ?\r\n\r\nThere are certainly other changes to make, but I still have to figure out which one :smile: Don't hesitate to participate on the PR I have opened ^^", "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.\n" ]

[]

[]

[ "Hi Ali!\r\nBased on your error info, it seems that a line in your `train.tsv` does not follow the MRPC format. Or more specifically, that line may contain an empty field. Could you add a try-except to catch and print that line? It'll look like:\r\n\r\n```python\r\ntry:\r\n text_a = line[3]\r\nexcept:\r\n print(line)\r\n```", "@JetRunner \r\nThank you very much for your suggestion. I found that there was a weird entry (that was not NaN nor empty) in some column of my data, and removing it with some modifications fixed the issue. \r\n\r\nThank you again for your clear explanation and suggestion. " ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "Hi, could you provide the versions of your software, like asked in the issue template?\r\n\r\nNamely tansformer version, python version, pytorch version, tensorflow version ...", "updated", "I still don't know what's your transformer version, which is arguably the most important version of the list :sweat_smile: \r\n\r\nDo you mind running `transformers-cli env` in your environment? It should output something along the lines of:\r\n\r\n```\r\n- `transformers` version: 2.11.0\r\n- Platform: Linux-5.6.15-1-MANJARO-x86_64-with-arch-Manjaro-Linux\r\n- Python version: 3.6.10\r\n- PyTorch version (GPU?): 1.5.0 (False)\r\n- Tensorflow version (GPU?): 2.2.0 (False)\r\n- Using GPU in script?: <fill in>\r\n- Using distributed or parallel set-up in script?: <fill in>\r\n```", "apologies :( , updated again.", "I met the same error with the same environment.\r\n", "@LousyLory Hi, have you solved this problem?\r\nI have try python3.6/3.7 + torch 1.5/1.4 in venv/conda.\r\nAnd I have checked my environment using `transformers-cli env`, which output same with @LousyLory .\r\nAll of them fail the test.\r\nFor conda + python3.6 + torch 1.4:\r\nthe failure case is: (The other cases are just like this)\r\n```\r\n======================================================================================================================== short test summary info =========================================================================================================================\r\nFAILED tests/test_modeling_electra.py::ElectraModelTest::test_multigpu_data_parallel_forward - RuntimeError: tensor.ndimension() == static_cast<int64_t>(expected_size.size()) INTERNAL ASSERT FAILED at /opt/conda/conda-bld/pytorch_1579022034529/work/torch/csrc/cud...\r\nFAILED tests/test_modeling_gpt2.py::GPT2ModelTest::test_multigpu_data_parallel_forward - RuntimeError: Gather got an input of invalid size: got [2, 2, 4, 7, 8], but expected [2, 4, 4, 7, 8] (gather at /opt/conda/conda-bld/pytorch_1579022034529/work/torch/csrc/cud...\r\nFAILED tests/test_modeling_xlnet.py::XLNetModelTest::test_multigpu_data_parallel_forward - RuntimeError: Gather got an input of invalid size: got [7, 2, 32], but expected [7, 4, 32] (gather at /opt/conda/conda-bld/pytorch_1579022034529/work/torch/csrc/cuda/comm.c...\r\nFAILED tests/test_modeling_ctrl.py::CTRLModelTest::test_multigpu_data_parallel_forward - RuntimeError: Gather got an input of invalid size: got [2, 2, 4, 7, 8], but expected [2, 4, 4, 7, 8] (gather at /opt/conda/conda-bld/pytorch_1579022034529/work/torch/csrc/cud...\r\nFAILED tests/test_modeling_mobilebert.py::MobileBertModelTest::test_multigpu_data_parallel_forward - RuntimeError: Caught RuntimeError in replica 0 on device 0.\r\n================================================================================================= 5 failed, 1054 passed, 560 skipped, 302 warnings in 253.59s (0:04:13) ==================================================================================================\r\n```\r\n", "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.\n" ]

[]

[]

[]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5068?src=pr&el=h1) Report\n> Merging [#5068](https://codecov.io/gh/huggingface/transformers/pull/5068?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/439aa1d6e9c953069f75fc23c737221d0df2c977&el=desc) will **increase** coverage by `0.98%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5068?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5068 +/- ##\n==========================================\n+ Coverage 76.45% 77.43% +0.98% \n==========================================\n Files 130 130 \n Lines 22024 22024 \n==========================================\n+ Hits 16839 17055 +216 \n+ Misses 5185 4969 -216 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5068?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/configuration\\_auto.py](https://codecov.io/gh/huggingface/transformers/pull/5068/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX2F1dG8ucHk=) | `92.85% <ø> (ø)` | |\n| [src/transformers/configuration\\_t5.py](https://codecov.io/gh/huggingface/transformers/pull/5068/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3Q1LnB5) | `96.42% <ø> (ø)` | |\n| [src/transformers/configuration\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5068/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3hsbmV0LnB5) | `94.00% <ø> (ø)` | |\n| [src/transformers/data/processors/utils.py](https://codecov.io/gh/huggingface/transformers/pull/5068/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9kYXRhL3Byb2Nlc3NvcnMvdXRpbHMucHk=) | `27.63% <ø> (ø)` | |\n| [src/transformers/modeling\\_auto.py](https://codecov.io/gh/huggingface/transformers/pull/5068/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19hdXRvLnB5) | `70.58% <ø> (ø)` | |\n| [src/transformers/modeling\\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5068/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19lbGVjdHJhLnB5) | `78.16% <ø> (ø)` | |\n| [src/transformers/modeling\\_longformer.py](https://codecov.io/gh/huggingface/transformers/pull/5068/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19sb25nZm9ybWVyLnB5) | `93.02% <ø> (ø)` | |\n| [src/transformers/modeling\\_reformer.py](https://codecov.io/gh/huggingface/transformers/pull/5068/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19yZWZvcm1lci5weQ==) | `88.19% <ø> (ø)` | |\n| [src/transformers/modeling\\_tf\\_albert.py](https://codecov.io/gh/huggingface/transformers/pull/5068/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9hbGJlcnQucHk=) | `75.33% <ø> (ø)` | |\n| [src/transformers/modeling\\_tf\\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5068/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `91.28% <ø> (ø)` | |\n| ... and [16 more](https://codecov.io/gh/huggingface/transformers/pull/5068/diff?src=pr&el=tree-more) | |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5068?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5068?src=pr&el=footer). Last update [439aa1d...6a52a4b](https://codecov.io/gh/huggingface/transformers/pull/5068?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "Hi! This is interesting. Could you resume what are the changes that would be needed in order to have our models scriptable?", "Sure, mostly my changes fall into these categories:\r\n\r\n### 1. Class members can only be basic types, None, nn.Modules, or list or tuple thereof\r\n- Solution: don't save whole config in the model, only individual entries you need, which are basic types\r\n- Solution for nn.functional: use the nn.Module equivalent of nn.functional\r\n- Solution for other functions: define and call the function globally, not as a class member\r\n\r\n### 2. Inputs are assumed to be Tensors\r\n- Solution: use typing to tell TorchScript the types (note - requires typing to be supported. I checked in python 3.7, but not 3.5 or 3.6)\r\n\r\n### 3. TorchScript can't figure out that an Optional is not None\r\n- Solution: add assertions to help TorchScript\r\n\r\n### 4. Variable types are not allowed to change depending on conditionals\r\n- Solution: use consistent types (with Optional to tell TorchScript that a variable/argument can be None) - this is where I had to change the interface, since current BERT models can optionally return attention probabilities. Had to change so that they always return the same sized output tuple, with None values, instead).\r\n\r\n### 5. TorchScript can't handle the expand (*) operator on lists\r\n- Solution: explicitly enumerate the arguments\r\n\r\n### 6. You can't use nn.Modules as local variables (take variable number of args)\r\n- Solution: use the nn.functional equivalents of the modules.\r\n\r\n### 7. TorchScript doesn't know about nn.ModuleList's enumerate \r\n- Solution: use a regular for loop\r\n\r\nMost of these are pretty small changes and do not affect the logic. #4 and #1c can be tricky, and #5 might be an issue with recent changes made here: https://github.com/huggingface/transformers/pull/4874", "Hi @sbrody18, \r\n\r\nThanks for opening this issue and taking the time to dive into our TorchScript support.\r\n\r\nRegarding **_A scriptable model would allow for variable-length input, offering big speedup gains and simplification_**:\r\n\r\nDo you have some numbers to compare against the current transformers library? We ran some TorchScript tests and the differences where not that huge at that time, may be this has changed since? I (and probably others) would be very interested in knowing more on this aspect.\r\n\r\nRegarding the list of changes you suggested: \r\n\r\nI'm currently not really in favour of such changes as they are almost changing all the way the library is designed and would have an impact on all the models. Some of them might be further discussed if there are real performance benefits.", "Hi @mfuntowicz,\r\nMy co-workers and I have run the experiments that show that inference time scales more-or-less linearly with the input size (also supported in the linked article below).\r\n\r\nAssuming you are trying to run in C++ (which is the reason to use TorchScript), the current solution, using `trace()` means that you can only use fixed length input - you have to set a large value for max_length to support your longest expected input, and zero-pad all input to the max-length.\r\nThat means if your max_length is 1000 tokens and your average length is 20 tokens, your inference is taking 50x longer than it should.\r\nYou can see an example of how big a difference this makes, [here](https://medium.com/roblox-tech-blog/how-we-scaled-bert-to-serve-1-billion-daily-requests-on-cpus-d99be090db26), under 'Scenario #3: Smaller Inputs (Dynamic Shapes)'.\r\n\r\nI'm guessing the tests you ran were focused specifically on the technical behavior of the models on a fixed input set and didn't take into account the max-length issue. Also, this is only an issue if you need to use TorchScript in order to run in C++.\r\n\r\nRe. the change to design, my intention is to keep the model changes to a minimum (e.g., adding type hints and asserts does not change the design at all) and make sure they are fully backwards compatible. There would still be some changes required, but I don't think they are drastic.\r\n\r\nAs I said in the original post, I have a PR where I did a lot of the work, and I'd be happy to work with someone to figure out how to get it to a state where it can be merged.", "@sbrody18 do you mind sharing your fork ? ", "Yes, I can do so, but it may have to wait a week or two - things are busy at the moment.", "I am very interested in this work as well. Our team would like to be able to use TorchScript so we can train without depending on Python. If there's any way I can be of help, I would gladly offer some time here!", "Sorry for the delay. I hope to have a reasonable PR later this week.", "My change is available at https://github.com/sbrody18/transformers/tree/scripting\r\n\r\nNote that it is based off of a commit from earlier this month:\r\nhttps://github.com/huggingface/transformers/compare/ef0e9d806c51059b07b98cb0279a20d3ba3cbc1d...sbrody18:scripting\r\nSince then there have been changes made to the BertModel interface adding a return_tuple argument and changing the return type of the forward method, and this would require more effort to resolve.\r\n\r\nI listed the principles I used in https://github.com/huggingface/transformers/issues/5067#issuecomment-644989375. The original components tended to return different sized tuples, depending on arguments, which is problematic for TorchScript. When a component BertX required an interface change to be scriptable, I made a BertScriptableX version with the modifications, and had the BertX component inherit from it and just modify the output so it is compatible with the original API.\r\n\r\nI made scriptable versions of BertModel and all the BertFor\\<Task\\> classes, except BertForMaskedLM (some complexities there were too much work for a proof of concept).\r\nI added a [test](https://github.com/sbrody18/transformers/blob/scripting/tests/test_modeling_bert.py#L529) to demonstrate the scripting capability.\r\n\r\nNote that my change disables the [gradient_checkpoint path](https://github.com/sbrody18/transformers/blob/scripting/src/transformers/modeling_bert.py#L474-492) in the encoder. I think this can be resolved, but I didn't have the time to work on it.", "@sgugger @joeddav: see comment above for preliminary PR. \r\nProbably too big and complicated to try to merge as is, but would be happy to work with someone to break things down into reasonable chunks.", "Thanks for all the work. Looking at this and our recent changes in the model API (in particular the return_dict argument) I think we probably won't be able to have the models be fully compatible with TorchScript. What is possible however would be to have a second version of the models that don't have the option of return_dict (we can also remove output_hiddens/output_attentions if it makes life easier) and would be fully scriptable.\r\n\r\nSince you already started with some components in a different class, I think we should have two models (let's say `BertModel` and `ScriptableBertModel`) with the same named parameters so you can seemlessly save/load from one to the other (a workflow would then be to experiment with `BertModel`, save the fine-tuned model and then go to `ScriptableBertModel` for inference for instance).\r\n\r\nThen I'm not sure what's easiest:\r\n- have the two inherit from some base class and have a minimal of methods that need to be different (probably just the forward?)\r\n- or have the second class be a complete rewrite.\r\n\r\nI think we should focus on having a proof of concept on one model before moving forward with others. ", "That makes sense to me. It will probably result in some amount of code duplication, and we'd need to make sure we keep the named parameters in sync, but probably easier to maintain.\r\nSo would you suggest the ScriptableBertModel is a separate file?", "Not necessarily a separate file, I guess it depends on the amount of code to rewrite. I think we can worry about this in a second stage, once we have a good poc.", "@sgugger Please see POC implementation in PR above.", "@sbrody18 in the original PR https://github.com/huggingface/transformers/pull/6846 you created for this issue, you mentioned you saw a large perf increase with dynamic sequences. What did you use as a test to make that determination?", "@kevinstephano - see discussion and conclussions [here](https://github.com/huggingface/transformers/pull/6907#issuecomment-687343119)\r\nWe saw a large perfomance increase with an older version of PyTorch, where traced models required the input to be the same length as the one used for tracing, making it necessary to pad short sequences at inference, and adding a lot of unnecessary computation overhead.\r\nWith recent versions of PyTorch (>=1.3, I think), this is no longer the case.", "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.\n", "Hi,\r\n\r\nI just tried `jit.script()`ing using Bert from the PR (just copied modeling_bert, modeling_utils and replaced relative imports of other dependencies with imports from transformers master branch)\r\nI see there are try blocks left in the code, which cause `jit.script` to fail:\r\n```\r\nUnsupportedNodeError: try blocks aren't supported:\r\n File \"/zhome/1d/8/153438/experiments/master-thesis/export_model/modeling_utils_script_proof.py\", line 131\r\n Get torch.device from module, assuming that the whole module has one device.\r\n \"\"\"\r\n try:\r\n ~~~ <--- HERE\r\n return next(self.parameters()).device\r\n except StopIteration:\r\n```\r\n\r\n@sbrody18 how did you export the model? I guess the workaround would be to remove try blocks, but apparently it did work for you as it is.\r\n", "@fteufel you can see #6846 for a stand-alone implementation that worked **at a previous version of the transformers library**. Maybe that's good enough for your purposes?\r\nThe transformers library has changed significantly since these PRs and I'm not sure if that try was added. If you are using code from the transformers master branch in the model itself, it's likely you will encounter several unscriptable bits.\r\nSpecifically for the next function, you can either:\r\na. remove the try block, since there should always be at least one parameter on the model\r\nb. use the next with default:\r\n first_param = next(self.parameters(), None)\r\n if not first_param: <handle it>\r\n return first_param.device\r\nc. figure out a better way to decide the model device :)", "@sbrody18 It seems have not been merged to official transformers ? My transformers Version: 4.21.3, and it can not use `jit.script` to convert BERT model to TorchScript.\r\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null }, { "id": 1834056635, "node_id": "MDU6TGFiZWwxODM0MDU2NjM1", "url": "https://api.github.com/repos/huggingface/transformers/labels/Core:%20Tokenization", "name": "Core: Tokenization", "color": "FF4446", "default": false, "description": "Internals of the library; Tokenization." } ]

[ { "login": "LysandreJik", "id": 30755778, "node_id": "MDQ6VXNlcjMwNzU1Nzc4", "avatar_url": "https://avatars.githubusercontent.com/u/30755778?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LysandreJik", "html_url": "https://github.com/LysandreJik", "followers_url": "https://api.github.com/users/LysandreJik/followers", "following_url": "https://api.github.com/users/LysandreJik/following{/other_user}", "gists_url": "https://api.github.com/users/LysandreJik/gists{/gist_id}", "starred_url": "https://api.github.com/users/LysandreJik/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/LysandreJik/subscriptions", "organizations_url": "https://api.github.com/users/LysandreJik/orgs", "repos_url": "https://api.github.com/users/LysandreJik/repos", "events_url": "https://api.github.com/users/LysandreJik/events{/privacy}", "received_events_url": "https://api.github.com/users/LysandreJik/received_events", "type": "User", "site_admin": false } ]

[ "This makes sense! Do you have some examples of some tokenization that can happen on TPUs with TensorFlow, so that we may consider what we have to do to enable this?", "I am not totally sure, but one thing to look at would probably be TF's native `tf.keras.preprocessing.text.Tokenizer`, which (I think) works when used within tf.data.Dataset maps on a TPU. \r\n\r\nhttps://www.tensorflow.org/api_docs/python/tf/keras/preprocessing/text/Tokenizer", "Assigning myself to keep this in mind", "\r\n> I am not totally sure, but one thing to look at would probably be TF's native `tf.keras.preprocessing.text.Tokenizer`, which (I think) works when used within tf.data.Dataset maps on a TPU.\r\n> \r\n> https://www.tensorflow.org/api_docs/python/tf/keras/preprocessing/text/Tokenizer\r\n\r\nInteresting. Is there a way to translate HF's tokenziers to this?\r\n\r\nI am wondering it's close to being possible that we don't need to convert text to Ids on the HF's data processors. ", "TensorFlow 2.3.x adds the Sentencepiece tokenizer to `tensorflow_text`. You can use this short script to turn a HuggingFace tokenizer into a `tensorflow_text.SentencepieceTokenizer`: https://gist.github.com/noahtren/6f9f6ecf2f81d0975c4f54afaeb95318\r\n\r\nI tested it on TPU and it's been working for me. My experience is that HuggingFace tokenizers are wrappers for https://github.com/google/sentencepiece so it's really simple to make it compatible with TensorFlow graph mode. Not sure yet if this works for all huggingface pretrained tokenizers.", "Did you try on the hf rust tokenizers as well? ", "@Santosh-Gupta No, I haven't tried that", "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.\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5065?src=pr&el=h1) Report\n> Merging [#5065](https://codecov.io/gh/huggingface/transformers/pull/5065?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/fc24a93e6493c2689e5585d12b7c43730ad9b3ea&el=desc) will **decrease** coverage by `0.05%`.\n> The diff coverage is `12.28%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5065?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5065 +/- ##\n==========================================\n- Coverage 79.02% 78.96% -0.06% \n==========================================\n Files 138 138 \n Lines 24064 24089 +25 \n==========================================\n+ Hits 19017 19023 +6 \n- Misses 5047 5066 +19 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5065?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/trainer\\_tf.py](https://codecov.io/gh/huggingface/transformers/pull/5065/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyX3RmLnB5) | `17.92% <8.00%> (-0.77%)` | :arrow_down: |\n| [src/transformers/training\\_args\\_tf.py](https://codecov.io/gh/huggingface/transformers/pull/5065/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmluZ19hcmdzX3RmLnB5) | `53.19% <42.85%> (+2.02%)` | :arrow_up: |\n| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/5065/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `39.38% <0.00%> (-0.24%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5065/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.92% <0.00%> (+0.29%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5065?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5065?src=pr&el=footer). Last update [fc24a93...9326d27](https://codecov.io/gh/huggingface/transformers/pull/5065?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "@Colanim Can you summarize your findings here? It would be better suited than going across multiple issues.", "@jplu Sure !\r\n\r\n* [x] Optimizer was passed as arguments to `_step` function, but we can pass only Tensors (issue #4994). This is fixed by ef02be8.\r\n\r\n* [x] Optimization steps were not computed correctly : it was ignoring gradient accumulation (issue #5072). This is fixed by a897e09.\r\n\r\n* [x] Computation of number of steps should take into account `dataloader_drop_last` (mentioned in #5072), and use `math.floor` in that case (instead of `math.ceil`). Fixed in #5051\r\n\r\n* [ ] I'm still having error as described in #4996, but I didn't figure the reason yet. Maybe it's my model that has a problem, not `TFTrainer`.", "Thanks for having sum up the issues here.\r\n\r\n> * [ ] Computation of number of steps should take into account `dataloader_drop_last` (mentioned in #5072), and use `math.floor` in that case (instead of `math.ceil`). Not fixed yet.\r\n\r\nThis is fixed in PR #5051.\r\n\r\n> * [ ] I'm still having error as described in #4996, but I didn't figure the reason yet. Maybe it's my model that has a problem, not `TFTrainer`.\r\n\r\nThis is what I'm currently looking for, but still not figuring out why :/\r\n\r\nWhich TPUs are you using, over ctpu, Cloud AI or Colab?", "> Which TPUs are you using, over ctpu, Cloud AI or Colab?\r\n\r\nI'm currently using ctpu, but I could see similar issue when using Colab.", "For me, this is the problem : https://github.com/huggingface/transformers/blob/5f721ad6e48c9d846de25c3fefa0e50a306cbf10/src/transformers/trainer_tf.py#L388-L389\r\n\r\nSomehow the Exception is not catched on TPU, which crash the training. Using `max_steps` argument instead of `num_train_epochs` fix the problem because we repeat the dataset, and therefore we never have an out of range error.\r\n\r\nIn the eval code, since we iterate the dataset without repeat, it's causing the same problem.", "Can I see all your logs output during the training?", "What is the command line you are using to create your TPU and run the process? In order to be aligned with the same errors :)", "@jplu finally I was wrong : even when using `max_steps` I'm having the problem.\r\n\r\nIt always happen at the beginning of the end of first epoch (for both validation and training). I think it's due to how the dataset is iterated. \r\n\r\nAccording to [this link](https://www.kaggle.com/mgornergoogle/custom-training-loop-with-100-flowers-on-tpu), we should have a single iterator for the whole training procedure (iterator should not be reset after each epoch), and always `repeat` the dataset indefinitely. Even for validation dataset.\r\n\r\nUnfortunately I'm having trouble with GCP right now, I can't try things on my end..", "Ok, can you share your command lines here please. Because even the size computation cannot be run on TPU with TF 2.2.", "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.\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5064?src=pr&el=h1) Report\n> Merging [#5064](https://codecov.io/gh/huggingface/transformers/pull/5064?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/d5477baf7d87b9bdad386f2f317732b85277b06b&el=desc) will **decrease** coverage by `0.05%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5064?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5064 +/- ##\n==========================================\n- Coverage 77.41% 77.36% -0.06% \n==========================================\n Files 130 130 \n Lines 22023 22023 \n==========================================\n- Hits 17050 17037 -13 \n- Misses 4973 4986 +13 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5064?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5064/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ194bG5ldC5weQ==) | `76.35% <0.00%> (-2.30%)` | :arrow_down: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5064/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `75.00% <0.00%> (-0.41%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5064/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.57% <0.00%> (+0.31%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5064?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5064?src=pr&el=footer). Last update [d5477ba...d7a3d5d](https://codecov.io/gh/huggingface/transformers/pull/5064?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "Hi! [This StackOverflow issue](https://datascience.stackexchange.com/questions/14812/making-keras-tensorflow-code-execution-deterministic-on-a-gpu) might be of interest to you.\r\n\r\nNamely:\r\n\r\n> In fact, the randomness(non-determinstic) is a behavior of GPU.\r\n> \r\n> The reason behind is that cuDNN(and othere CUDA stuffs) uses a non-deterministic algorithm to compute gradients, thus we can't determine anything.", "Note that this issue is being addressed as an [issue in the tensorflow-determinism repo](https://github.com/NVIDIA/tensorflow-determinism/issues/19). I have also added a reference to that repo in the above-mentioned Stack Exchange / Data Science 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.\n" ]

[]

[ { "login": "patrickvonplaten", "id": 23423619, "node_id": "MDQ6VXNlcjIzNDIzNjE5", "avatar_url": "https://avatars.githubusercontent.com/u/23423619?v=4", "gravatar_id": "", "url": "https://api.github.com/users/patrickvonplaten", "html_url": "https://github.com/patrickvonplaten", "followers_url": "https://api.github.com/users/patrickvonplaten/followers", "following_url": "https://api.github.com/users/patrickvonplaten/following{/other_user}", "gists_url": "https://api.github.com/users/patrickvonplaten/gists{/gist_id}", "starred_url": "https://api.github.com/users/patrickvonplaten/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/patrickvonplaten/subscriptions", "organizations_url": "https://api.github.com/users/patrickvonplaten/orgs", "repos_url": "https://api.github.com/users/patrickvonplaten/repos", "events_url": "https://api.github.com/users/patrickvonplaten/events{/privacy}", "received_events_url": "https://api.github.com/users/patrickvonplaten/received_events", "type": "User", "site_admin": false } ]

[ "I added a description of `d_kv`. `d_model` is as you say the size of the embedding. `d_kv` is the size of the key, query, value projections. \r\n\r\nIf you look at this blog post: http://jalammar.github.io/illustrated-transformer/\r\n\r\n`d_model` corresponds to the size of a vector `x_1` and `d_kv` of a vector `q_1, k_1, v_1`" ]

[ { "id": 2155169140, "node_id": "MDU6TGFiZWwyMTU1MTY5MTQw", "url": "https://api.github.com/repos/huggingface/transformers/labels/trainer", "name": "trainer", "color": "2ef289", "default": false, "description": "" } ]

[]

[ "Yeah I agree with the motivations here. From my experience I like to initialize wandb as soon as the main script starts which has benefits like \r\n- Capturing all the console logs, so if something failed before the execution reached the trainer class one could debug the issue through Wandb. This is specially useful for doing training in a kubernetes environment where the logs are not very easily available after a crash.\r\n- Instantiating at the beginning also allows us to capture all the CLI arguments properly before they might have been modified. This is a problem right now because Wandb only captures the training_args leaving out model_args or data args. This is important for replicating a training run in the future.", "Doesn't it work already when initializing wandb outside?\r\n\r\nI believe that `wandb.init` does not create a new run if one is already running so all the functions should be available anywhere in the script.", "When looking over the code, I was pretty sure it will initialize a new run. However, I checked and everything (mentioned in the issue) works smoothly. ", "Ok, thanks for checking @Guitaricet " ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5060?src=pr&el=h1) Report\n> Merging [#5060](https://codecov.io/gh/huggingface/transformers/pull/5060?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/d5477baf7d87b9bdad386f2f317732b85277b06b&el=desc) will **increase** coverage by `0.01%`.\n> The diff coverage is `84.21%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5060?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5060 +/- ##\n==========================================\n+ Coverage 77.41% 77.43% +0.01% \n==========================================\n Files 130 130 \n Lines 22023 22029 +6 \n==========================================\n+ Hits 17050 17059 +9 \n+ Misses 4973 4970 -3 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5060?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/5060/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `39.62% <50.00%> (+0.09%)` | :arrow_up: |\n| [src/transformers/data/data\\_collator.py](https://codecov.io/gh/huggingface/transformers/pull/5060/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9kYXRhL2RhdGFfY29sbGF0b3IucHk=) | `98.33% <93.33%> (+8.67%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5060?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5060?src=pr&el=footer). Last update [d5477ba...857975c](https://codecov.io/gh/huggingface/transformers/pull/5060?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Refactored on @julien-c suggestion and removed the little hack to handle tensors and lists of ints (which was cauding a 4x slowdown on my tests). I just use the first features to test if we have a tensor (then stack) or not (then use torch.tensor).", "LGTM!" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5059?src=pr&el=h1) Report\n> Merging [#5059](https://codecov.io/gh/huggingface/transformers/pull/5059?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/d5477baf7d87b9bdad386f2f317732b85277b06b&el=desc) will **not change** coverage.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5059?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5059 +/- ##\n=======================================\n Coverage 77.41% 77.41% \n=======================================\n Files 130 130 \n Lines 22023 22023 \n=======================================\n Hits 17050 17050 \n Misses 4973 4973 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5059?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5059/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `75.00% <0.00%> (-0.41%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5059/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `87.42% <0.00%> (+0.15%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5059?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5059?src=pr&el=footer). Last update [d5477ba...6bef869](https://codecov.io/gh/huggingface/transformers/pull/5059?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Merging so that external contrib can pick up the next steps without conflicts. Feel free to add comments here or in #5057 " ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "Hi! Could you show what command you used to launch the script?", "I was runing the token classification example : https://github.com/huggingface/transformers/blob/master/examples/token-classification/run.sh with flaubert-large-cased as model name. \r\n\r\nI tried to use the downloaded model from https://huggingface.co/flaubert models, but after 100 epochs the results are very bad, the model did not learn any thing. \r\nI Don't understand what is the problem with flaubert-large-cased model. \r\nNote that flaubert-base-cased model give good results on NER taks.", "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.\n" ]

[ { "id": 1108649053, "node_id": "MDU6TGFiZWwxMTA4NjQ5MDUz", "url": "https://api.github.com/repos/huggingface/transformers/labels/Help%20wanted", "name": "Help wanted", "color": "008672", "default": false, "description": "Extra attention is needed, help appreciated" }, { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null }, { "id": 1936351150, "node_id": "MDU6TGFiZWwxOTM2MzUxMTUw", "url": "https://api.github.com/repos/huggingface/transformers/labels/Examples", "name": "Examples", "color": "d4c5f9", "default": false, "description": "Which is related to examples in general" }, { "id": 1990918270, "node_id": "MDU6TGFiZWwxOTkwOTE4Mjcw", "url": "https://api.github.com/repos/huggingface/transformers/labels/Good%20First%20Issue", "name": "Good First Issue", "color": "bbf794", "default": false, "description": "" }, { "id": 2139563322, "node_id": "MDU6TGFiZWwyMTM5NTYzMzIy", "url": "https://api.github.com/repos/huggingface/transformers/labels/cleanup", "name": "cleanup", "color": "e7fc49", "default": false, "description": "" } ]

[ { "login": "sshleifer", "id": 6045025, "node_id": "MDQ6VXNlcjYwNDUwMjU=", "avatar_url": "https://avatars.githubusercontent.com/u/6045025?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sshleifer", "html_url": "https://github.com/sshleifer", "followers_url": "https://api.github.com/users/sshleifer/followers", "following_url": "https://api.github.com/users/sshleifer/following{/other_user}", "gists_url": "https://api.github.com/users/sshleifer/gists{/gist_id}", "starred_url": "https://api.github.com/users/sshleifer/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/sshleifer/subscriptions", "organizations_url": "https://api.github.com/users/sshleifer/orgs", "repos_url": "https://api.github.com/users/sshleifer/repos", "events_url": "https://api.github.com/users/sshleifer/events{/privacy}", "received_events_url": "https://api.github.com/users/sshleifer/received_events", "type": "User", "site_admin": false } ]

[ "Hi, @sshleifer I would like to work on this issue. Shall I take this up.\r\n\r\n", "Yes, I would pick one item from the list to start with.\r\nMake sure you pull first, I just merged some improvements.", "@sshleifer I will work on the first one.\r\n\r\nJust to be clear I will note down what I have understood and what I have in mind to do.\r\n\r\n1. The issue as per my understanding: The tests in the example folder are not up to the mark and we have to add certain parts to fix this. For this, as the first point suggests when running tests in the examples folder the tests is not checking if Cuda or fp16 is available.\r\n\r\n2. There are 4 tests in the `test_examples.py`\r\n- text-classification(run_glue)\r\n- language-modeling(run_language_modeling)\r\n- question-answering(run_squad)\r\n- text-generation(run_generation)\r\nso each should run with cuda or fp16 if available.\r\n\r\ncorrect if I am wrong.", "Good idea.\r\n\r\n1) Yes. I think the desired behavior is\r\nif `torch.cuda.is_available()`:\r\n- assume fp16 is available\r\n- run the code with fp16 and cude.\r\n\r\nTry to do that for all tests. Some will likely break. You can add a TODO to those and keep them running on CPU for now.\r\n\r\n1b) You probably need a GPU to do this PR.\r\n\r\n2) There are more tests than that:\r\n\r\n```bash\r\n$ ls examples/**/test*.py\r\n\r\nexamples/adversarial/test_hans.py\r\nexamples/summarization/bertabs/test_utils_summarization.py\r\nexamples/summarization/test_summarization_examples.py\r\nexamples/test_examples.py\r\nexamples/token-classification/test_ner_examples.py\r\nexamples/translation/t5/test_t5_examples.py\r\n```\r\n", "You don't need to cover all those tests. Feel free to break the work into very small PRs and tag me on them.", "Thanks, @sshleifer for the clarification \r\n\r\nI will start working on this.", "> 2. The `@slow` decorator used in the main tests is not importable, so there are no @slow tests.\r\n\r\nThis is no longer the case. \r\n\r\n```from transformers.testing_utils import slow```\r\n\r\nThis item can be removed.", "> 3. `test_run_glue` uses distilbert-case-cased. It should use a smaller model, one of the `tiny` family [here](https://huggingface.co/models?search=sshleifer/tiny) or a new tiny model.\r\n\r\nI tried a few and either they have a wrong head dimension as in `sshleifer/tiny-distilbert-base-cased` (9x2), but tests are (2x2), so it won't load as is (`size mismatch for classifier.weight:` and `size mismatch for classifier.bias`), or they perform terribly with the current test settings.\r\n\r\nI also did an experiment for the same for the suggested inside the existing test:\r\n```\r\n def test_run_language_modeling(self):\r\n stream_handler = logging.StreamHandler(sys.stdout)\r\n logger.addHandler(stream_handler)\r\n # TODO: switch to smaller model like sshleifer/tiny-distilroberta-base\r\n```\r\nwith terrible results (perplexity > 5,000, whereas the current one < 35).\r\n\r\nSo when these tiny models are suggested as a replacement to speed things up, what things are to be sacrificed?\r\n", "Happy to do big models and mark slow. I just don't want to do big models when we are only testing output shape.", "> Happy to do big models and mark slow. I just don't want to do big models when we are only testing output shape.\r\n\r\nSo then we could write a test that uses a tiny model that does just that? i.e. no outcome quality checks. Leaving big models for quality checks with @slow.", "Yes!", "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.\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5056?src=pr&el=h1) Report\n> Merging [#5056](https://codecov.io/gh/huggingface/transformers/pull/5056?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/b28b53713161a6299c757c32f7179a2cb2d8cbd7&el=desc) will **increase** coverage by `0.05%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5056?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5056 +/- ##\n==========================================\n+ Coverage 77.96% 78.02% +0.05% \n==========================================\n Files 138 138 \n Lines 23838 23847 +9 \n==========================================\n+ Hits 18585 18606 +21 \n+ Misses 5253 5241 -12 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5056?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/tokenization\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5056/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHMucHk=) | `92.47% <100.00%> (+0.96%)` | :arrow_up: |\n| [src/transformers/tokenization\\_utils\\_base.py](https://codecov.io/gh/huggingface/transformers/pull/5056/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHNfYmFzZS5weQ==) | `92.82% <100.00%> (+1.95%)` | :arrow_up: |\n| [src/transformers/tokenization\\_utils\\_fast.py](https://codecov.io/gh/huggingface/transformers/pull/5056/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHNfZmFzdC5weQ==) | `94.28% <100.00%> (+2.31%)` | :arrow_up: |\n| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/5056/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `38.38% <0.00%> (-1.19%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5056/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `86.00% <0.00%> (ø)` | |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5056?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5056?src=pr&el=footer). Last update [b28b537...8f87f25](https://codecov.io/gh/huggingface/transformers/pull/5056?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[]

[ "Duplicate of #4144 ", "Thanks!" ]