Datasets:

---

DanielPFlorian

/

Transformers-Github-Issues

like 0

[ { "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" ]

[]

[]

[ "^^ I borked your PR unfortunately (💩). Sorry about that. Care to open a new one?" ]

[]

[]

[ "T5 uses CrossEntropy loss so I think you can use `torch.exp(loss)` to calculate perplexity.", "Thanks!", "@patil-suraj Sorry to bother you again, I thought my problem has been solved, but I don't know how to deal with the loss. It's a (1,1,vocab_size) tensor. I don't even have an intuition about what that means. Can you help me?", "It seems that you are not passing `lm_labels`. When you pass `lm_labels` to T5/BART's forward methods then the first element of the output tuple is the loss, otherwise its `prediction_scores` tensor which you are receiving.", "hi @patil-suraj bart doesnt have lm_labels has argument", "`lm_labels` is renamed to `labels`", "Thanks for the prompt response, @patil-suraj , previously before trying Bart, I trained it with t5,the valid loss gets to almost zero, but still found it difficult generating correct translation. I would love to know reasons for that. Thanks", "I'm not sure which problem you are talking about, hard to answer without knowing the context.", "Hi patil,\n\nI am working on a translation task/competition, so as a text2text\nframework, I used t5 even though it is not pretrained on my source language\njust for the knowledge sake,\nThe notebook link here\n\nhttps://colab.research.google.com/drive/1w0brh9KULV4zDmgMme4csXeCpfAR960N?usp=sharing\n\nthe test and valid loss was declining to zero even though the later\ngenerated predictions wasn't good so I changed to Bart even though no\nimprovement, I have no idea why? I would love to know the reason.\n\nI found MarianMT just yesterday which was pretrained for my task yo-en, I\nhave started working with it. I think it would also be the case for MarianMt\nThank you.\n\nOn Fri, Jan 1, 2021, 8:38 AM Suraj Patil <[email protected]> wrote:\n\n> I'm not sure which problem you are talking about, hard to answer without\n> knowing the context.\n>\n> —\n> You are receiving this because you commented.\n> Reply to this email directly, view it on GitHub\n> <https://github.com/huggingface/transformers/issues/4210#issuecomment-753280527>,\n> or unsubscribe\n> <https://github.com/notifications/unsubscribe-auth/AOP3ZJFLJLLFGUSFCU3WKB3SXV3X7ANCNFSM4M3PUPXA>\n> .\n>\n" ]

[]

[]

[]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/4208?src=pr&el=h1) Report\n> Merging [#4208](https://codecov.io/gh/huggingface/transformers/pull/4208?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/cafa6a9e29f3e99c67a1028f8ca779d439bc0689&el=desc) will **decrease** coverage by `0.01%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/4208?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #4208 +/- ##\n==========================================\n- Coverage 78.32% 78.31% -0.02% \n==========================================\n Files 120 120 \n Lines 19854 19854 \n==========================================\n- Hits 15551 15548 -3 \n- Misses 4303 4306 +3 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/4208?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/4208/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `39.88% <ø> (-0.30%)` | :arrow_down: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/4208/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `73.96% <0.00%> (-0.42%)` | :arrow_down: |\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/4208/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `90.96% <0.00%> (-0.13%)` | :arrow_down: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/4208?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/4208?src=pr&el=footer). Last update [cafa6a9...bfbed71](https://codecov.io/gh/huggingface/transformers/pull/4208?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Superseded and fixed by #4213 " ]

[ { "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" ]

[ { "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" ]

[ { "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" ]

[]

[]

[ "How many labels/entities are there in your model ? Can you check if the key 422 exists in the models config's id2label by printing it or by opening the config file. You can print config using `model.config` ", "Hello\r\nIt seems I had simply been misusing my model : I was instantiating an `AutoModel`, when I should have been using the `BertForTokenClassification` class. \r\nThanks anyway ! :) " ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/4203?src=pr&el=h1) Report\n> Merging [#4203](https://codecov.io/gh/huggingface/transformers/pull/4203?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/026097b9ee7862905ec3f3b0e729c5dbc95a0fd9&el=desc) will **increase** coverage by `0.01%`.\n> The diff coverage is `0.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/4203?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #4203 +/- ##\n==========================================\n+ Coverage 78.43% 78.44% +0.01% \n==========================================\n Files 120 120 \n Lines 19800 19797 -3 \n==========================================\n Hits 15530 15530 \n+ Misses 4270 4267 -3 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/4203?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/hf\\_argparser.py](https://codecov.io/gh/huggingface/transformers/pull/4203/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9oZl9hcmdwYXJzZXIucHk=) | `61.97% <0.00%> (+1.16%)` | :arrow_up: |\n| [src/transformers/trainer.py](https://codecov.io/gh/huggingface/transformers/pull/4203/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmVyLnB5) | `41.83% <0.00%> (-0.35%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/4203/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `92.77% <0.00%> (+0.16%)` | :arrow_up: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/4203/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `74.38% <0.00%> (+0.41%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/4203?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/4203?src=pr&el=footer). Last update [026097b...a73d487](https://codecov.io/gh/huggingface/transformers/pull/4203?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Thanks!" ]

[ { "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" } ]

[]

[]

[]

[]

[]

[ { "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" ]

[]

[]

[ "Thanks for your contribution @girishponkiya , merging!" ]

[]

[]

[ "This google colab shows nicely how the new benchmarking utils can be used: https://colab.research.google.com/drive/1GM3WPXmZ5wZHwhdPyf7RykJS3Y12iceJ?usp=sharing", "A plotter file is included in `examples/benchmarking` which can give the following results:\r\n\r\n### Memory\r\n\r\n\r\n\r\n### Time\r\n\r\n\r\n", "Can `csv_memory_filename_inference` be called `save_csv_path`?", "> Can `csv_memory_filename_inference` be called `save_csv_path`?\r\n\r\nYeah about the exact naming I think we can decide a bit later. Wanna see first if you guys are ok with the general design. But we have up to four csv files that are saved - so `save_csv_path` is too general.", "> This is really cool, I'm looking forward to using it.\r\n> \r\n> It seems you've implemented TPU support as well, did you get a chance to try it?\r\n\r\nI didn't play around with TPU very much yet. I'd like to merge this PR which works for GPU / CPU for PyTorch now and then maybe clean things for TPU in a second PR? ", "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/4198?src=pr&el=h1) Report\n> Merging [#4198](https://codecov.io/gh/huggingface/transformers/pull/4198?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/4bd7cfc30952922b6329ec2d7637f13729317014&el=desc) will **not change** coverage.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/4198?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #4198 +/- ##\n=======================================\n Coverage 77.39% 77.39% \n=======================================\n Files 128 128 \n Lines 20939 20939 \n=======================================\n Hits 16205 16205 \n Misses 4734 4734 \n```\n\n\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/4198?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/4198?src=pr&el=footer). Last update [4bd7cfc...4bd7cfc](https://codecov.io/gh/huggingface/transformers/pull/4198?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 } ]

[]

[ "It looks like when you load a tokenizer from a dir it's also looking for files to load it's related model config via `AutoConfig.from_pretrained`. It does this because it's using the information from the config to to determine which model class the tokenizer belongs to (BERT, XLNet, etc ...) since there is no way of knowing that with the saved tokenizer files themselves. This is deceptive, as intuitively it would seem you should be able to load a tokenizer without needing it's model config. Plus in the documentation of `AutoTokenizer` under examples it states:\r\n\r\n```\r\n# If vocabulary files are in a directory (e.g. tokenizer was saved using `save_pretrained('./test/saved_model/')`)\r\ntokenizer = AutoTokenizer.from_pretrained('./test/bert_saved_model/')\r\n```\r\n\r\nEither:\r\n1. The documentation needs an update to clarify that `config` must be a kwarg of `from_pretrained` or the dir must also have the files for `AutoConfig` (if loading from a dir)\r\n2. `AutoTokenizer` should be changed to be independent from `AutoConfig` (maybe the model base class name should be stored along with \"tokenizer_config.json\")", "I had the same issue loading a saved tokenizer. Separately loading and saving the config file to the tokenizer directory worked.\r\n```\r\nconfig_file = AutoConfig.from_pretrained('distilgpt2')\r\nconfig_file.save_pretrained(saved_tokenizer_directory)\r\n```", "+1 to @jaymody. It seems as though that `AutoTokenizer` is coupled to the model that uses it, and I don't see why it should be impossible to load a tokenizer without an explicit model configuration.\r\n\r\nFor those that happen to be loading a tokenizer and model using the `Auto` classes, I was able to use the following workaround where my tokenizer relies on the configuration in the model directory:\r\n\r\n```\r\ntokenizer = AutoTokenizer.from_pretrained(<tokenizer-dir>, config=AutoConfig.from_pretrained(<model-dir>))\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": 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" } ]

[]

[]

[ { "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", "Closing since already implemented" ]

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

[]

[ "GPT2 Attention has `merge_heads` method which takes an input as `(batch, heads, seq, repr)` and outputs the flattened _(merged)_ representation of it `(batch, seq, heads * repr)`. \r\n\r\n```python\r\n def merge_heads(self, x):\r\n x = x.transpose(2, 1).contiguous()\r\n new_x_shape = x.size()[:-2] + (x.size(-2) * x.size(-1),)\r\n return x.view(*new_x_shape) # in Tensorflow implem: fct merge_states\r\n```\r\n\r\nThis can be simplified with `flatten()` to avoid constructing a tuple and accessing shape information.\r\n\r\n```python\r\n def merge_heads(self, x):\r\n x = x.transpose(2, 1).contiguous()\r\n return x.flatten(-2) # in Tensorflow implem: fct merge_states\r\n```", "GPT2 Attention has a `Conv1D` projection which is nothing else than a linear projection of the input. \r\n\r\nThe current implementation use BLAS acceleration function `addmm` (Matrix Multiplication + Add) to handle the affine projection `w * x.T + b`. \r\n\r\nThis is currently implemented with the non batched BLAS call and requires flattening the sequence over the batch axis and unflattening after the call: \r\n\r\n```python\r\ndef forward(self, x):\r\n size_out = x.size()[:-1] + (self.nf,)\r\n torch.addmm(self.bias, x, self.weight)\r\n return x.view(*size_out)\r\n```\r\n\r\nWe can optimize this call to remove all the reshape operations by leveraging the batched BLAS call (`baddbmm` : Batched Matrix Multiply + Batched Add) which takes 3D inputs as (`b, m, n)` and thus avoid the reshapings\r\n\r\n```python\r\ndef forward(self, x):\r\n return torch.baddbmm(self.bias[None, None], x, self.weight[None])\r\n```\r\n", "Use tuples instead of list for returning objects from layers (refer to [this SO post](https://stackoverflow.com/a/22140115))", "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 } ]

[]

[ "This is currently not supported. There used to be some community maintained examples IIRC. You can have a look here: https://github.com/huggingface/transformers/tree/1.1.0/examples/lm_finetuning\r\n\r\n" ]

[]

[]

[]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/4191?src=pr&el=h1) Report\n> Merging [#4191](https://codecov.io/gh/huggingface/transformers/pull/4191?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/96c78396ce1baf5e19c0618689005f93c7f42d79&el=desc) will **decrease** coverage by `0.00%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/4191?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #4191 +/- ##\n==========================================\n- Coverage 78.41% 78.41% -0.01% \n==========================================\n Files 120 120 \n Lines 19785 19785 \n==========================================\n- Hits 15515 15514 -1 \n- Misses 4270 4271 +1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/4191?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_reformer.py](https://codecov.io/gh/huggingface/transformers/pull/4191/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19yZWZvcm1lci5weQ==) | `88.17% <ø> (ø)` | |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/4191/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `92.77% <0.00%> (-0.17%)` | :arrow_down: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/4191?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/4191?src=pr&el=footer). Last update [96c7839...488d787](https://codecov.io/gh/huggingface/transformers/pull/4191?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[]

[]

[]

[]

[ "By the way, when I downgrade Pytorch 1.5.0 to 1.4.0, the error disappears. ", "The same issue: #3936", "Closing in favor of #3936" ]

[ { "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/4188?src=pr&el=h1) Report\n> Merging [#4188](https://codecov.io/gh/huggingface/transformers/pull/4188?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/877fc56410e3a0495f62e07e66a73e6b3b9629bc&el=desc) will **increase** coverage by `0.00%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/4188?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #4188 +/- ##\n=======================================\n Coverage 78.10% 78.11% \n=======================================\n Files 117 117 \n Lines 18962 18963 +1 \n=======================================\n+ Hits 14811 14813 +2 \n+ Misses 4151 4150 -1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/4188?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_albert.py](https://codecov.io/gh/huggingface/transformers/pull/4188/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19hbGJlcnQucHk=) | `75.38% <100.00%> (+0.06%)` | :arrow_up: |\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/4188/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `90.92% <0.00%> (-0.13%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/4188/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `92.93% <0.00%> (+0.32%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/4188?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/4188?src=pr&el=footer). Last update [877fc56...7c64f23](https://codecov.io/gh/huggingface/transformers/pull/4188?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" ]

[ { "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" ]

[]

[]

[ "This supercedes https://github.com/huggingface/transformers/pull/2840, where I added patience to the outdated `run_language_modeling.py` script.", "Looking good! Can you add a reference to your original post that this closes https://github.com/huggingface/transformers/issues/4894? Thanks", "Hello, when this feature will be merged? I would like to use it. Thank you.", "> Hello, when this feature will be merged? I would like to use it. Thank you.\r\n\r\nThere are some changes requested that @thesamuel should fix before this can be merged.", "Bump. Early stopping is critical for an automated Trainer that reliably gives us the best model. Current way of figuring out the training stopping point seems to be specifying a static train_epochs but the training duration a model can take depends on way too many factors like learning rate, data complexity, model, model size, optimizer and so on that it is unreasonable to ask the user to specify the epochs in advance.\r\nI believe the current assumption is that people train with very small learning rates so that the loss always seems to keep decreasing very slowly but according to my experience (and on my data) it is a sub-optimal schedule which takes too much time. I see that training with higher learning rates with larger batch sizes and stopping at the early stopping point results in an equally good if not better models. Although this requires use of early stopping.", "I would like to use this early stopping on downstream training.\r\nThe current implementation only stops training by monitoring loss. IMO it should also be possible to monitor other metrics like F1 and ROC-AUC. \r\n\r\nI also would like to add a feature that stores the model each time when the monitored metric improves and then optionaly loads the model after training. Then later evaluation can be done on this \"best\" model.\r\n\r\n@thesamuel @julien-c @kevin-yauris what do you think?", "I plan to work on this once I'm finished with the Funnel Transformer model @PhilipMay (so end of this week, beginning of the next).", "> \r\n> \r\n> I plan to work on this once I'm finished with the Funnel Transformer model @PhilipMay (so end of this week, beginning of the next).\r\n\r\n@sgugger That would be awsome. Maybe you want to get some inspiration from the FARM training loop which is pretty nice IMO:\r\n\r\nhttps://github.com/deepset-ai/FARM/blob/master/farm/train.py#L262-L370", "I just found this PR that was already merged: #7431\r\nI think it solved this...", "Not quite, but it makes implementing it easier.", "> Not quite, but it makes implementing it easier.\r\n\r\nYes - you are right. The patience part is still missing. ", "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", "@sgugger Should we keep this open? You wrote in this thread you will work on this if you find the time, but I am not sure if you plan to use another PR for that.", "There has been a PR merged adding the `EarlyStoppingCallback` (#8581) so I think this can be closed now.", "Thanks @cbrochtrup @sgugger! Sorry I didn't get around to this...", "You're welcome, happy to help!" ]

[ { "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": "" } ]

[]

[ "https://github.com/lonePatient/MobileBert_PyTorch", "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 was completed in #4901" ]

[ { "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" } ]

[]

[]

[ { "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" } ]

[]

[ "Great work and great model cards, congrats on beating the state-of-the-art of Polish NLU!", "Thanks!" ]

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

[]

[ "Not from Huggingface, but for this, I would use something like SentenceBERT. As far as I know.", "@moinnadeem Thank you for pointing me in that direction! But would sentencebert really work for documents?", "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": 2039044877, "node_id": "MDU6TGFiZWwyMDM5MDQ0ODc3", "url": "https://api.github.com/repos/huggingface/transformers/labels/marian", "name": "marian", "color": "30cc95", "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 } ]

[ "marian C++ code uses `<unk>` in this case" ]

[ { "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" ]

[ { "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/4179?src=pr&el=h1) Report\n> Merging [#4179](https://codecov.io/gh/huggingface/transformers/pull/4179?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/ff8ed52dd8c6268f2535c0721cde9e360fbb0ce0&el=desc) will **not change** coverage.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/4179?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #4179 +/- ##\n=======================================\n Coverage 78.79% 78.79% \n=======================================\n Files 114 114 \n Lines 18712 18712 \n=======================================\n Hits 14745 14745 \n Misses 3967 3967 \n```\n\n\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/4179?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/4179?src=pr&el=footer). Last update [ff8ed52...c11fb93](https://codecov.io/gh/huggingface/transformers/pull/4179?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Really cool. Do they have a downstream task that we can fine-tune on, and get some eval results?", "I will check it again and will let you know.\r\nI am also working on a Molecule Drug Transformer for drug target interaction system. It is more powerful and they provide several downstream tasks and evaluation datasets. Something like this: https://arxiv.org/abs/2004.11424" ]

[ { "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/4178?src=pr&el=h1) Report\n> Merging [#4178](https://codecov.io/gh/huggingface/transformers/pull/4178?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/ff8ed52dd8c6268f2535c0721cde9e360fbb0ce0&el=desc) will **decrease** coverage by `0.00%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/4178?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #4178 +/- ##\n==========================================\n- Coverage 78.79% 78.79% -0.01% \n==========================================\n Files 114 114 \n Lines 18712 18712 \n==========================================\n- Hits 14745 14744 -1 \n- Misses 3967 3968 +1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/4178?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/4178/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `92.61% <0.00%> (-0.17%)` | :arrow_down: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/4178?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/4178?src=pr&el=footer). Last update [ff8ed52...dc9a6bc](https://codecov.io/gh/huggingface/transformers/pull/4178?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "I am working on a NMT en to es following this format https://github.com/facebookresearch/XLM/#iii-applications-supervised--unsupervised-mt\r\n\r\nIs there any way to integrate my final model on the HUB? (script for conversion or smth like that)", "This creates a lot of files but I think it's fine. Thoughts?", "I went through all the 995 files and they look good", "> I went through all the 995 files and they look good\r\n\r\nhaha", "> This creates a lot of files but I think it's fine. Thoughts?\r\n\r\nI think it's fine as well!", "fwiw, it's 5MB of data. Previously, model_cards was 1MB of data. I think it's fine to merge @julien-c ", "Gunna merge this 7pm EST barring objections @julien-c ", "Can you add them to the S3 bucket instead of the repo? I'll add code tomorrow to display them on the website from the repo (not currently the case)", "Sure, will add once the group naming question is resolved." ]

[]

[]

[ "You need to make sure you install from source as documented in the README" ]

[]

[]

[]

[ { "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" ]

[]

[]

[ "Thanks for contributing #4173 :)." ]

[]

[]

[]

[]

[]

[ "Got the same result with 2.9.0, i.e. with `# pip install tensorflow==2.0 transformers==2.9.0`", "`AutoModel` is the equivalent of `TFAutoModel` but for PyTorch model classes. If you don't have pytorch installed this is expected.\r\n\r\nUse `TFAutoModel` instead =)\r\n\r\n", "Thank you!", "@julien-c \r\nBut there still is a problem with `transformers-cli`.\r\nNow, people who want to use `transformers-cli download` are required to install PyTorch even when they use Tensorflow only.\r\n\r\nedit:\r\nIt could be the problem specific to `t5-large`. I'll try it later.", "I have pytorch installed but still show the error:\r\n`ImportError: cannot import name 'AutoModelForSequenceClassification'`\r\n\r\nEnvironment info:\r\n* Ubuntu 16.04.6\r\n* Python 3.6.8\r\n* Torch 1.2.0\r\n* CUDA 10.0\r\n* transformers 4.8.2", "@JTWang2000, even without PyTorch you should be able to import `AutoModelForSequenceClassification`:\r\n\r\n```py\r\nPython 3.8.6 (default, Dec 4 2020, 09:21:54) \r\n[GCC 10.2.0] on linux\r\nType \"help\", \"copyright\", \"credits\" or \"license\" for more information.\r\n>>> import torch\r\nTraceback (most recent call last):\r\n File \"<stdin>\", line 1, in <module>\r\nModuleNotFoundError: No module named 'torch'\r\n>>> from transformers import AutoModelForSequenceClassification\r\n>>>\r\n```\r\n\r\nCan you offer a reproducible example? Thanks!", "@LysandreJik Thank you for your reply\r\n### Steps to reproduce the behavior:\r\n\r\n* Python: 3.6.8\r\n* Ubuntu 16.04.6\r\n* CUDA 10.0\r\n` pip install torch==1.2.0 torchvision==0.4.0`\r\n` pip install transformers`\r\n```\r\nPython 3.6.8 (default, Jul 6 2021, 15:59:52)\r\n[GCC 5.4.0 20160609] on linux\r\nType \"help\", \"copyright\", \"credits\" or \"license\" for more information.\r\n>>> from transformers import AutoModel\r\nTraceback (most recent call last):\r\n File \"<stdin>\", line 1, in <module>\r\nImportError: cannot import name 'AutoModel'\r\n>>> from transformers import AutoModelForSequenceClassification\r\nTraceback (most recent call last):\r\n File \"<stdin>\", line 1, in <module>\r\nImportError: cannot import name 'AutoModelForSequenceClassification'\r\n```", "The may be som error as I am getting:\r\n\r\n```\r\nTraceback (most recent call last):\r\n File \"/home/usename/dev/speech_recognition.py\", line 1, in <module>\r\n from transformers import Wav2Vec2Tokenizer, Wav2Vec2ForCTC\r\nImportError: cannot import name 'Wav2Vec2ForCTC' from 'transformers' (unknown location)\r\n```\r\n\r\nThe line of code: \r\n\r\n```\r\nfrom transformers import Wav2Vec2Tokenizer, Wav2Vec2ForCTC\r\n```\r\n\r\nTransformers: 4.3.0\r\nPython 3.9.6\r\nOS: Fedora 34", "@JTWang2000 Recent versions of `transformers` are compatible with more recent PyTorch versions, see the [README](https://github.com/huggingface/transformers#with-pip).\r\n\r\n@kennylajara could you try upgrading `transformers` to a more recent version to see if it fixes your issue? (Wav2Vec2 has been improved since v4.3.0) - if not, do you mind linking me to a reproducible code example (with pip installations)/colab notebook or something similar that reproduces your issue?", "**EDIT:** Well... **This** issue has been solved by updating dependencies but now it produces another issue. I will post the link to the reproducible issue in a minute.", "@LysandreJik There you go: [https://github.com/kennylajara/speech-recognition](https://github.com/kennylajara/speech-recognition)", "@LysandreJik Thank you for the reply.\r\nI reinstall PyTorch with version 1.7.0 and now it works. Thanks! ", "@kennylajara answered in a separate issue (please keep issues on different subjects separate, thank you): https://github.com/kennylajara/speech-recognition/issues/1", "In my case I both installed pytorch and tensorflow (noob). Creating new environment just for pytorch help me solve it." ]

[]

[ { "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 } ]

[ "Are you using this exact line\r\n```\r\nmodel = EncoderDecoderModel.from_encoder_decoder_pretrained('bert-base-uncased', 'bert-base-uncased') # initialize Bert2Bert\r\n```\r\nIf yes, then please use paths for your saved model. Few other things to try: verify data pipeline,\r\ntry using beam search or sampling in generate", "Thanks! I am using that exact line. I saved my trained model using save_pretrained() and it saved everything as one file. How would I separate this, or should I just retrain and re-save the encoder and decoder separately? Also, does the untrained model not work due to the untrained cross attention layer? ", "If you saved your model using `.save_pretrained` then you can load it using just `.from_pretrained` as you load any other HF model. Just pass the path of your saved model. You won't need to use `.from_encoder_decoder_pretrained`", "Hi @anishthite,\r\n\r\nHow did you train your Bert2Bert model? Can you post the code you used to train your model here? Dontt worry if it's a very long code snippet :-)", "Hello! I managed to figure out the issue. I retrained and saved the encoder and decoder in their own folders. I then was able to load it in as @patil-suraj suggested. I guess earlier it was loading in the untrained model. Would it be helpful to redefine save_pretrained() for EncoderDecoder models to automatically split it into an encoder and decoder folder I can submit a PR if you want. \r\n```\r\n dataset = QADataset(dataset=args.traindataset, block_size=args.maxseqlen)\r\n qa_loader = DataLoader(dataset, batch_size=args.batch, shuffle=True)\r\n model.train()\r\n optimizer = AdamW(model.parameters(), lr=LEARNING_RATE)\r\n t_total = len(qa_loader) // args.gradient_acums * args.epochs\r\n scheduler = get_linear_schedule_with_warmup(optimizer, num_warmup_steps=WARMUP_STEPS, num_training_steps = t_total)\r\n proc_seq_count = 0\r\n sum_loss = 0.0\r\n batch_count = 0\r\n models_folder = \"combinerslargeencoder\"\r\n models_folder2 = \"combinerslargedecoder\"\r\n if not os.path.exists(models_folder):\r\n os.mkdir(models_folder)\r\n if not os.path.exists(models_folder2):\r\n os.mkdir(models_folder2)\r\n for epoch in range(args.epochs):\r\n \r\n print(f\"EPOCH {epoch} started\" + '=' * 30)\r\n \r\n for idx,qa in enumerate(qa_loader):\r\n print(str(idx) + ' ' + str(len(qa_loader)))\r\n inputs, labels = (qa[0], qa[1])\r\n inputs = inputs.to(device)\r\n labels = labels.to(device)\r\n outputs = model(input_ids=inputs, decoder_input_ids=labels, lm_labels=labels)\r\n loss, logits = outputs[:2]\r\n loss = loss / args.gradient_acums\r\n loss.backward()\r\n sum_loss = sum_loss + loss.detach().data\r\n \r\n #proc_seq_count = proc_seq_count + 1\r\n #if proc_seq_count == args.gradient_acums:\r\n # proc_seq_count = 0 \r\n batch_count += 1\r\n if (idx + 1) % args.gradient_acums == 0:\r\n optimizer.step()\r\n scheduler.step() \r\n optimizer.zero_grad()\r\n model.zero_grad()\r\n\r\n if batch_count == 100:\r\n print(f\"sum loss {sum_loss}\")\r\n batch_count = 0\r\n sum_loss = 0.0\r\n \r\n # Store the model after each epoch to compare the performance of them\r\n torch.save(model.state_dict(), os.path.join(models_folder, f\"combined_mymodel_{args.maxseqlen}{epoch}{args.gradient_acums}.pt\"))\r\n model.save_pretrained(models_folder)\r\n model.encoder.save_pretrained(models_folder)\r\n model.decoder.save_pretrained(models_folder2)\r\n evaluate(args, model, tokenizer)\r\n```", "Why do you save the encoder and decoder model seperately?:\r\n```\r\n model.encoder.save_pretrained(models_folder)\r\n model.decoder.save_pretrained(models_folder2)\r\n```\r\nThis line:\r\n```\r\n model.save_pretrained(models_folder)\r\n```\r\nshould be enough. \r\n\r\nWe moved away from saving the model to two separate folders, see: https://github.com/huggingface/transformers/pull/3383. Also the docs: https://huggingface.co/transformers/model_doc/encoderdecoder.html might be useful." ]

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

[]

[ "You don't need to provide position IDs to the model, the model will create them on its own. Similarly to the `attention_mask` and `token_type_ids`, you only need to provide them if you want them to be different to the default.\r\n\r\nIn the case of position IDs, provide them if you want to use a special scheme/different to what was used when pre-training the 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" ]

[ { "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", "The issue was resolved." ]

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

[]

[ "Hi, you can import it as such:\r\n```py\r\nfrom transformers.modeling_albert import AlbertMLMHead\r\n```\r\n\r\nIt's not in the `__init.py__` as it's not a model but part of one." ]

[]

[]

[]

[ { "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": "" } ]

[]

[ "hey, @thomwolf this would be a great addition! have a look into it.", "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 think @NielsRogge is working on this over here: https://github.com/NielsRogge/transformers/tree/modeling_tapas" ]

[]

[]

[ "(note that this will need to be refined to work with TF / beam search if we decide on the direction)", "Closing in favor of #5416 " ]

[ { "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 } ]

[ "Very interesting idea! I think we eventually have to make the generation function more general anyways. \r\n\r\nMaybe it's time to move this whole code:\r\n\r\n```python \r\n # repetition penalty from CTRL paper (https://arxiv.org/abs/1909.05858)\r\n if repetition_penalty != 1.0:\r\n self.enforce_repetition_penalty_(next_token_logits, batch_size, 1, input_ids, repetition_penalty)\r\n if no_repeat_ngram_size > 0:\r\n # calculate a list of banned tokens to prevent repetitively generating the same ngrams\r\n # from fairseq: https://github.com/pytorch/fairseq/blob/a07cb6f40480928c9e0548b737aadd36ee66ac76/fairseq/sequence_generator.py#L345\r\n banned_tokens = calc_banned_ngram_tokens(input_ids, batch_size, no_repeat_ngram_size, cur_len)\r\n for batch_idx in range(batch_size):\r\n next_token_logits[batch_idx, banned_tokens[batch_idx]] = -float(\"inf\")\r\n if bad_words_ids is not None:\r\n # calculate a list of banned tokens according to bad words\r\n banned_tokens = calc_banned_bad_words_ids(input_ids, bad_words_ids)\r\n for batch_idx in range(batch_size):\r\n next_token_logits[batch_idx, banned_tokens[batch_idx]] = -float(\"inf\")\r\n # set eos token prob to zero if min_length is not reached\r\n if eos_token_id is not None and cur_len < min_length:\r\n next_token_logits[:, eos_token_id] = -float(\"inf\")\r\n if do_sample:\r\n # Temperature (higher temperature => more likely to sample low probability tokens)\r\n if temperature != 1.0:\r\n next_token_logits = next_token_logits / temperature\r\n # Top-p/top-k filtering\r\n next_token_logits = top_k_top_p_filtering(next_token_logits, top_k=top_k, top_p=top_p)\r\n # Sample\r\n probs = F.softmax(next_token_logits, dim=-1)\r\n next_token = torch.multinomial(probs, num_samples=1).squeeze(1)\r\n\r\n if sampling_transform:\r\n next_token = sampling_transform(input_ids, probs, next_token)\r\n else:\r\n # Greedy decoding\r\n next_token = torch.argmax(next_token_logits, dim=-1)\r\n```\r\n\r\neven to a `Sampler` class with a `sampler.sample(input_ids, next_token_logits)` which can also include a generic function as proposed. \r\n\r\nWhat are your thoughts on this @yjernite @thomwolf @sshleifer @LysandreJik ?", "Having a formal class sounds good to me; personally what I had in mind was a torchvision transforms type interface - so something like\r\n\r\n```\r\nSampler.Compose([\r\n temperature(2) ,\r\n top_p(75),\r\n early_eos(),\r\n my_crazy_custom_transform(),\r\n])\r\n```\r\n\r\nThe current operations are order-dependent which isn't necessarily apparent to the user in params. \r\n", "@turtlesoupy that's definitely a feature we want! Would it be possible to apply the transform at the logit level rather than have it sample too? It seems like it fits the proposed use case and it would make beam search significantly easier.", "@yjernite good call; the interface would have to be modified slightly to `(input_ids, next_probs, next_token) -> (next_probs)`. You might be able to fold the sampling procedure into the transforms. I know in my case I want to operate after a sample has been chosen\r\n\r\n```\r\nSampler.Compose([\r\n one_hot_draw(),\r\n my_crazy_custom_transform(),\r\n])\r\n```\r\n", "@patrickvonplaten @yjernite finally got around to doing this -- can you take a look at #5420 and let me know if it can be merged? ", "Thanks a lot for the PR :-) I will take a look this week!", "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": "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 } ]

[]

[ { "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" } ]

[]

[]

[]

[ { "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 think in your case it might make more sense to use the `TextGenerationPipeline`. It's a two liner that does the same as the script. You can wrap a `open(inpt_text_file)` and `file.write()` around these two lines in your use case. \r\n\r\nSee documentation for the generation pipeline here: https://huggingface.co/transformers/usage.html#text-generation", "Closing for now" ]

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

[]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/4159?src=pr&el=h1) Report\n> Merging [#4159](https://codecov.io/gh/huggingface/transformers/pull/4159?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/7822cd38a0e18004ab1a55bfe85e8b3bc0d8857a&el=desc) will **increase** coverage by `0.00%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/4159?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #4159 +/- ##\n=======================================\n Coverage 78.14% 78.15% \n=======================================\n Files 120 120 \n Lines 20053 20053 \n=======================================\n+ Hits 15670 15672 +2 \n+ Misses 4383 4381 -2 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/4159?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/tokenization\\_marian.py](https://codecov.io/gh/huggingface/transformers/pull/4159/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fbWFyaWFuLnB5) | `87.32% <ø> (-0.35%)` | :arrow_down: |\n| [src/transformers/tokenization\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/4159/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHMucHk=) | `89.51% <100.00%> (+0.02%)` | :arrow_up: |\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/4159/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `90.60% <0.00%> (+0.12%)` | :arrow_up: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/4159/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `73.85% <0.00%> (+0.41%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/4159?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/4159?src=pr&el=footer). Last update [7822cd3...cc21682](https://codecov.io/gh/huggingface/transformers/pull/4159?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "I'm in favor of this since we also have a `batch_encode_plus` method. ", "One small thing: the `encode_plus` method uses the `batch_encode_plus` whereas here it's the other way around. But this does not bother me too much.", "I don't have a strong knowledge of the API design for tokenizers so will let others chime in instead.\r\n" ]

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

[ "This is only an issue for BPE models, which we are not supporting." ]

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

[ "Not true, fixed on master." ]

[]

[]

[ "Could you please tell what does this mean ?\r\n```\r\n#!/bin/bash -eo pipefail\r\nblack --check --line-length 119 --target-version py35 examples templates tests src utils\r\n\r\nwould reformat /home/circleci/transformers/examples/pplm/run_pplm.py\r\nOh no! 💥 💔 💥\r\n1 file would be reformatted, 266 files would be left unchanged.\r\n\r\nExited with code exit status 1\r\n```\r", "You need to run `make style` and `make quality` locally and push the resulting changes", "Also this file was moved around so you'll need to re-apply your changes to new location", "I ran both make commands as you said.\r\n```\r\n(base) u37216@s001-n007:~/transformers$ make style\r\nblack --line-length 119 --target-version py35 examples templates tests src utils\r\nreformatted /home/u37216/transformers/src/transformers/__init__.py\r\nreformatted /home/u37216/transformers/templates/adding_a_new_example_script/run_xxx.py\r\nreformatted /home/u37216/transformers/templates/adding_a_new_example_script/utils_xxx.py\r\nAll done! ✨ 🍰 ✨\r\n3 files reformatted, 272 files left unchanged.\r\nisort --recursive examples templates tests src utils\r\nFixing /home/u37216/transformers/templates/adding_a_new_example_script/run_xxx.py\r\nFixing /home/u37216/transformers/templates/adding_a_new_example_script/utils_xxx.py\r\nFixing /home/u37216/transformers/src/transformers/__init__.py\r\n(base) u37216@s001-n007:~/transformers$ make quality\r\nblack --check --line-length 119 --target-version py35 examples templates tests src utils\r\nwould reformat /home/u37216/transformers/src/transformers/__init__.py\r\nwould reformat /home/u37216/transformers/templates/adding_a_new_example_script/run_xxx.py\r\nwould reformat /home/u37216/transformers/templates/adding_a_new_example_script/utils_xxx.py\r\nOh no! 💥 💔 💥\r\n3 files would be reformatted, 272 files would be left unchanged.\r\nMakefile:6: recipe for target 'quality' failed\r\nmake: *** [quality] Error 1\r\n```\r\nIs this related to my code ? I didn't even touch other parts of the file. I think the formatting issue came from master itself (I checked it). ", "@julien-c Can you please see this ?" ]

[]

[]

[ "Hi, did you find out the reason for this error? I get the same error even on Wikitext-2 corpus ", "Yeah, it is an issue with how your data is read. Try the --line_by_line argument. " ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/4154?src=pr&el=h1) Report\n> Merging [#4154](https://codecov.io/gh/huggingface/transformers/pull/4154?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/818463ee8eaf3a1cd5ddc2623789cbd7bb517d02&el=desc) will **increase** coverage by `0.03%`.\n> The diff coverage is `96.42%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/4154?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #4154 +/- ##\n==========================================\n+ Coverage 78.79% 78.83% +0.03% \n==========================================\n Files 114 114 \n Lines 18711 18726 +15 \n==========================================\n+ Hits 14743 14762 +19 \n+ Misses 3968 3964 -4 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/4154?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/pipelines.py](https://codecov.io/gh/huggingface/transformers/pull/4154/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9waXBlbGluZXMucHk=) | `76.27% <96.42%> (+1.32%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/4154/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `92.77% <0.00%> (+0.16%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/4154?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/4154?src=pr&el=footer). Last update [818463e...95a257c](https://codecov.io/gh/huggingface/transformers/pull/4154?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Ok so this should be ready to merge, right @mfuntowicz?", "+1 Would love a merge here so I can more easily improve `TranslationPipeline`" ]

[]

[]

[ "I am running into the same error on my own script. Interestingly it only appears on CPU... Did you find a solution?", "No, I want to get a French Q&A pipeline, surprinsingly, with the hugging face pipeline everything works great, I can plug the code in a local server and make requests on it. \r\n\r\nBut when I try to use the same code in a docker envrionement to ship it, it fail with this error (only in french with camembert, classic bert works fine)\r\n\r\nI get the error locally as well if I try not to use the hugging face pipeline but write my own inference (as described above)", "I can confirm it's working on GPU local (and even in a docker) but still stuck on CPU", "I actually figured out my error. I was adding special tokens to the tokenizer (like begin-of-sequence) but did not resize the models token embeddings via: \r\n`model.resize_token_embeddings(len(self.tokenizer))`\r\nJust in case someone else is not reading the documentation carefully enough :see_no_evil: \r\nConsidering that, the error message did actually make sense. ", "Hi @Ierezell, there is indeed an issue which I'm patching in #4289. Please be aware that you're using `CamembertModel` which cannot be used for question answering. Please use `CamembertForQuestionAnswering` instead.", "It's patched now, please install from source and there should be no error anymore!", "Hi @LysandreJik, I'm concious that I used it with a non QA model but it was to try the base model supported by hugging face. \r\n\r\nI tried as well with `illuin/camembert-base-fquad` (large as well) and with `fmikaelian/camembert-base-fquad` \r\n\r\nI will install the latest version and try it.\r\n\r\nThnaks a lot for the fast support !", "I tried your fix but it lead to key errors : \r\n\r\n```\r\n File \"/home/pedro/.local/lib/python3.8/site-packages/transformers/pipelines.py\", line 1156, in __call__\r\n answers += [\r\n File \"/home/pedro/.local/lib/python3.8/site-packages/transformers/pipelines.py\", line 1159, in <listcomp>\r\n \"start\": np.where(char_to_word == feature.token_to_orig_map[s])[0][0].item(),\r\nKeyError: 0\r\n```", "Could you provide a reproducible script? I can't reproduce.", "My problem here was surely linked with #4674 everything seems to work now, thanks a lot" ]

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

[]

[ "### Screenshots of Documentation\r\n\r\n\r\n\r\n\r\n[ x] Shows up in TOC:\r\n\r\n", "The documentation looks great!", "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/4152?src=pr&el=h1) Report\n> Merging [#4152](https://codecov.io/gh/huggingface/transformers/pull/4152?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/a01a3fecb4fe203c086d3d7450b76bbcfa035725&el=desc) will **not change** coverage.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/4152?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #4152 +/- ##\n=======================================\n Coverage 77.52% 77.52% \n=======================================\n Files 120 120 \n Lines 19932 19932 \n=======================================\n Hits 15453 15453 \n Misses 4479 4479 \n```\n\n\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/4152?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/4152?src=pr&el=footer). Last update [a01a3fe...a01a3fe](https://codecov.io/gh/huggingface/transformers/pull/4152?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 }, { "id": 1834052847, "node_id": "MDU6TGFiZWwxODM0MDUyODQ3", "url": "https://api.github.com/repos/huggingface/transformers/labels/Ex:%20LM%20(Finetuning)", "name": "Ex: LM (Finetuning)", "color": "26FFF8", "default": false, "description": "Related to language modeling fine-tuning" }, { "id": 1834053007, "node_id": "MDU6TGFiZWwxODM0MDUzMDA3", "url": "https://api.github.com/repos/huggingface/transformers/labels/Ex:%20LM%20(Pretraining)", "name": "Ex: LM (Pretraining)", "color": "76FFAF", "default": false, "description": "Related to language modeling pre-training" } ]

[ { "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 } ]

[ "We still need to provide a good docstring/notebook for this. It's on our ToDo-List. :-) \r\n\r\nOr @sshleifer - is there already something for Bart? ", "Nothing yet, would be good to add!", "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 have seen the same [issue](https://github.com/pytorch/fairseq/issues/1899) in fairseq BART!.", "Hi, any news about bart pre-training? ", " who can tell me how to pre-train the bart on my own dataset? I am so confused ....\r\nthank you so much", "Maybe this comment can help: https://github.com/huggingface/transformers/issues/5096#issuecomment-645860271", "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", "Any news on this please?\r\n", "not so far, would be great to have it. Thanks.", "I and my co-worker wrote a demo according to roberta pretraining demo.\r\n```\r\n#encoding=utf-8\r\n\r\nfrom transformers import (\r\n BartForConditionalGeneration, BartTokenizer, BartForCausalLM,\r\n Seq2SeqTrainingArguments, Seq2SeqTrainer\r\n )\r\n\r\nimport torch\r\nfrom torch.utils.data import random_split\r\n\r\n\r\n# ## Initiating model and trainer for training\r\nfrom transformers import BartModel, BartConfig\r\nfrom transformers import BartTokenizerFast\r\n\r\n\r\nconfiguration = BartConfig(\r\n vocab_size=52000,\r\n max_position_embeddings=258,\r\n d_model=256,\r\n encoder_layers=3,\r\n decoder_layers=3,\r\n encoder_attention_heads=4,\r\n decoder_attention_heads=4,\r\n decoder_ffn_dim=1024,\r\n encoder_ffn_dim=1024,\r\n)\r\nmodel = BartForCausalLM(configuration)\r\ntokenizer = BartTokenizerFast.from_pretrained(\"./dic\", max_len=256, additional_special_tokens=['[CH]', '[OTHER]', '[VAR]', '[NUM]'])\r\n\r\n\r\n# ### HTTP Request DataPreparing & Modeling\r\ndata = []\r\nwith open(\"../data/sample.txt\") as f1:\r\n for src in f1:\r\n data.append(\r\n {\r\n \"seq2seq\": {\r\n \"input\": src.strip()\r\n }\r\n }\r\n )\r\nprint(f'total size of data is {len(data)}')\r\n\r\n\r\n# splitting dataset into train, validation\r\nsplit = 0.2\r\ntrain_dataset, eval_dataset = random_split(data, lengths=[int((1-split)*len(data))+1, int(split*len(data))])\r\n\r\n\r\n# defining collator functioon for preparing batches on the fly ..\r\ndef data_collator(features:list):\r\n inputs = [f[\"seq2seq\"][\"input\"] for f in features]\r\n batch = tokenizer.prepare_seq2seq_batch(src_texts=inputs, max_length=256, padding='max_length')\r\n batch[\"labels\"] = batch[\"input_ids\"].copy()\r\n for k in batch:\r\n batch[k] = torch.tensor(batch[k])\r\n return batch\r\n\r\n\r\nbatch_out = data_collator(eval_dataset)\r\nprint(batch_out)\r\nprint(batch_out['input_ids'].shape,batch_out['labels'].shape,batch_out['attention_mask'].shape)\r\n\r\n\r\n# defining training related arguments\r\nargs = Seq2SeqTrainingArguments(output_dir=\"clm-checkpoints\",\r\n do_train=True,\r\n do_eval=True,\r\n evaluation_strategy=\"epoch\",\r\n per_device_train_batch_size=8,\r\n per_device_eval_batch_size=8,\r\n learning_rate=5e-5,\r\n num_train_epochs=1,\r\n logging_dir=\"./logs\")\r\n\r\n\r\n# defining trainer using 🤗\r\ntrainer = Seq2SeqTrainer(model=model, \r\n args=args, \r\n data_collator=data_collator, \r\n train_dataset=train_dataset, \r\n eval_dataset=eval_dataset)\r\n\r\n\r\n# ## Training time\r\ntrainer.train()\r\n# It will take hours to train this model on this dataset\r\n\r\n\r\n# lets save model\r\ntrainer.evaluate(eval_dataset=eval_dataset)\r\ntrainer.save_model(\"clm-checkpoints\")\r\n\r\n```", "> I and my co-worker wrote a demo according to roberta pretraining demo.\r\n> \r\n> ```\r\n> #encoding=utf-8\r\n> \r\n> from transformers import (\r\n> BartForConditionalGeneration, BartTokenizer, BartForCausalLM,\r\n> Seq2SeqTrainingArguments, Seq2SeqTrainer\r\n> )\r\n> \r\n> import torch\r\n> from torch.utils.data import random_split\r\n> \r\n> \r\n> # ## Initiating model and trainer for training\r\n> from transformers import BartModel, BartConfig\r\n> from transformers import BartTokenizerFast\r\n> \r\n> \r\n> configuration = BartConfig(\r\n> vocab_size=52000,\r\n> max_position_embeddings=258,\r\n> d_model=256,\r\n> encoder_layers=3,\r\n> decoder_layers=3,\r\n> encoder_attention_heads=4,\r\n> decoder_attention_heads=4,\r\n> decoder_ffn_dim=1024,\r\n> encoder_ffn_dim=1024,\r\n> )\r\n> model = BartForCausalLM(configuration)\r\n> tokenizer = BartTokenizerFast.from_pretrained(\"./dic\", max_len=256, additional_special_tokens=['[CH]', '[OTHER]', '[VAR]', '[NUM]'])\r\n> \r\n> \r\n> # ### HTTP Request DataPreparing & Modeling\r\n> data = []\r\n> with open(\"../data/sample.txt\") as f1:\r\n> for src in f1:\r\n> data.append(\r\n> {\r\n> \"seq2seq\": {\r\n> \"input\": src.strip()\r\n> }\r\n> }\r\n> )\r\n> print(f'total size of data is {len(data)}')\r\n> \r\n> \r\n> # splitting dataset into train, validation\r\n> split = 0.2\r\n> train_dataset, eval_dataset = random_split(data, lengths=[int((1-split)*len(data))+1, int(split*len(data))])\r\n> \r\n> \r\n> # defining collator functioon for preparing batches on the fly ..\r\n> def data_collator(features:list):\r\n> inputs = [f[\"seq2seq\"][\"input\"] for f in features]\r\n> batch = tokenizer.prepare_seq2seq_batch(src_texts=inputs, max_length=256, padding='max_length')\r\n> batch[\"labels\"] = batch[\"input_ids\"].copy()\r\n> for k in batch:\r\n> batch[k] = torch.tensor(batch[k])\r\n> return batch\r\n> \r\n> \r\n> batch_out = data_collator(eval_dataset)\r\n> print(batch_out)\r\n> print(batch_out['input_ids'].shape,batch_out['labels'].shape,batch_out['attention_mask'].shape)\r\n> \r\n> \r\n> # defining training related arguments\r\n> args = Seq2SeqTrainingArguments(output_dir=\"clm-checkpoints\",\r\n> do_train=True,\r\n> do_eval=True,\r\n> evaluation_strategy=\"epoch\",\r\n> per_device_train_batch_size=8,\r\n> per_device_eval_batch_size=8,\r\n> learning_rate=5e-5,\r\n> num_train_epochs=1,\r\n> logging_dir=\"./logs\")\r\n> \r\n> \r\n> # defining trainer using 🤗\r\n> trainer = Seq2SeqTrainer(model=model, \r\n> args=args, \r\n> data_collator=data_collator, \r\n> train_dataset=train_dataset, \r\n> eval_dataset=eval_dataset)\r\n> \r\n> \r\n> # ## Training time\r\n> trainer.train()\r\n> # It will take hours to train this model on this dataset\r\n> \r\n> \r\n> # lets save model\r\n> trainer.evaluate(eval_dataset=eval_dataset)\r\n> trainer.save_model(\"clm-checkpoints\")\r\n> ```\r\n\r\nThanks for the code example, I am also planning on implementing pretrained from scratch, and I've got several questions for the code\r\n- I noticed that you use pretrained bart tokenizer, how can I pretrain it for different language?\r\n- How much compute did you use for your implementation?", "> > I and my co-worker wrote a demo according to roberta pretraining demo.\r\n> > ```\r\n> > #encoding=utf-8\r\n> > \r\n> > from transformers import (\r\n> > BartForConditionalGeneration, BartTokenizer, BartForCausalLM,\r\n> > Seq2SeqTrainingArguments, Seq2SeqTrainer\r\n> > )\r\n> > \r\n> > import torch\r\n> > from torch.utils.data import random_split\r\n> > \r\n> > \r\n> > # ## Initiating model and trainer for training\r\n> > from transformers import BartModel, BartConfig\r\n> > from transformers import BartTokenizerFast\r\n> > \r\n> > \r\n> > configuration = BartConfig(\r\n> > vocab_size=52000,\r\n> > max_position_embeddings=258,\r\n> > d_model=256,\r\n> > encoder_layers=3,\r\n> > decoder_layers=3,\r\n> > encoder_attention_heads=4,\r\n> > decoder_attention_heads=4,\r\n> > decoder_ffn_dim=1024,\r\n> > encoder_ffn_dim=1024,\r\n> > )\r\n> > model = BartForCausalLM(configuration)\r\n> > tokenizer = BartTokenizerFast.from_pretrained(\"./dic\", max_len=256, additional_special_tokens=['[CH]', '[OTHER]', '[VAR]', '[NUM]'])\r\n> > \r\n> > \r\n> > # ### HTTP Request DataPreparing & Modeling\r\n> > data = []\r\n> > with open(\"../data/sample.txt\") as f1:\r\n> > for src in f1:\r\n> > data.append(\r\n> > {\r\n> > \"seq2seq\": {\r\n> > \"input\": src.strip()\r\n> > }\r\n> > }\r\n> > )\r\n> > print(f'total size of data is {len(data)}')\r\n> > \r\n> > \r\n> > # splitting dataset into train, validation\r\n> > split = 0.2\r\n> > train_dataset, eval_dataset = random_split(data, lengths=[int((1-split)*len(data))+1, int(split*len(data))])\r\n> > \r\n> > \r\n> > # defining collator functioon for preparing batches on the fly ..\r\n> > def data_collator(features:list):\r\n> > inputs = [f[\"seq2seq\"][\"input\"] for f in features]\r\n> > batch = tokenizer.prepare_seq2seq_batch(src_texts=inputs, max_length=256, padding='max_length')\r\n> > batch[\"labels\"] = batch[\"input_ids\"].copy()\r\n> > for k in batch:\r\n> > batch[k] = torch.tensor(batch[k])\r\n> > return batch\r\n> > \r\n> > \r\n> > batch_out = data_collator(eval_dataset)\r\n> > print(batch_out)\r\n> > print(batch_out['input_ids'].shape,batch_out['labels'].shape,batch_out['attention_mask'].shape)\r\n> > \r\n> > \r\n> > # defining training related arguments\r\n> > args = Seq2SeqTrainingArguments(output_dir=\"clm-checkpoints\",\r\n> > do_train=True,\r\n> > do_eval=True,\r\n> > evaluation_strategy=\"epoch\",\r\n> > per_device_train_batch_size=8,\r\n> > per_device_eval_batch_size=8,\r\n> > learning_rate=5e-5,\r\n> > num_train_epochs=1,\r\n> > logging_dir=\"./logs\")\r\n> > \r\n> > \r\n> > # defining trainer using 🤗\r\n> > trainer = Seq2SeqTrainer(model=model, \r\n> > args=args, \r\n> > data_collator=data_collator, \r\n> > train_dataset=train_dataset, \r\n> > eval_dataset=eval_dataset)\r\n> > \r\n> > \r\n> > # ## Training time\r\n> > trainer.train()\r\n> > # It will take hours to train this model on this dataset\r\n> > \r\n> > \r\n> > # lets save model\r\n> > trainer.evaluate(eval_dataset=eval_dataset)\r\n> > trainer.save_model(\"clm-checkpoints\")\r\n> > ```\r\n> \r\n> Thanks for the code example, I am also planning on implementing pretrained from scratch, and I've got several questions for the code\r\n> \r\n> * I noticed that you use pretrained bart tokenizer, how can I pretrain it for different language?\r\n> * How much compute did you use for your implementation?\r\n\r\n\r\n\r\nFor the first question, just like this:\r\n\r\n```\r\nfrom tokenizers import (ByteLevelBPETokenizer,SentencePieceBPETokenizer,BertWordPieceTokenizer)\r\n\r\ntokenizer = ByteLevelBPETokenizer()\r\npaths = ['./data/corpus.txt']\r\ntokenizer.train(files=paths, vocab_size = 15000, min_frequency=6, special_tokens = [\"[UNK]\", \"[CLS]\", \"[SEP]\", \"[PAD]\", \"[MASK]\"])\r\n\r\ntokenizer.save_model(\"./data/dic/\")\r\n```\r\n\r\nFor the other question, i trained it with 12G gpu memory，but it may be completed with samller gpu memory. And also you could adjust you parameters to your server environment.", "@myechona Thanks for your code. I have a question about it. There are some tasks like text-filling and sentence-permutation during pretrain stage, i want to know whether the \"input_ids\" is for masked sentence and the \"labels\" is for origin sentence?", "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!", "> I and my co-worker wrote a demo according to roberta pretraining demo.\r\n> \r\n> ```\r\n> #encoding=utf-8\r\n> \r\n> from transformers import (\r\n> BartForConditionalGeneration, BartTokenizer, BartForCausalLM,\r\n> Seq2SeqTrainingArguments, Seq2SeqTrainer\r\n> )\r\n> \r\n> import torch\r\n> from torch.utils.data import random_split\r\n> \r\n> \r\n> # ## Initiating model and trainer for training\r\n> from transformers import BartModel, BartConfig\r\n> from transformers import BartTokenizerFast\r\n> \r\n> \r\n> configuration = BartConfig(\r\n> vocab_size=52000,\r\n> max_position_embeddings=258,\r\n> d_model=256,\r\n> encoder_layers=3,\r\n> decoder_layers=3,\r\n> encoder_attention_heads=4,\r\n> decoder_attention_heads=4,\r\n> decoder_ffn_dim=1024,\r\n> encoder_ffn_dim=1024,\r\n> )\r\n> model = BartForCausalLM(configuration)\r\n> tokenizer = BartTokenizerFast.from_pretrained(\"./dic\", max_len=256, additional_special_tokens=['[CH]', '[OTHER]', '[VAR]', '[NUM]'])\r\n> \r\n> \r\n> # ### HTTP Request DataPreparing & Modeling\r\n> data = []\r\n> with open(\"../data/sample.txt\") as f1:\r\n> for src in f1:\r\n> data.append(\r\n> {\r\n> \"seq2seq\": {\r\n> \"input\": src.strip()\r\n> }\r\n> }\r\n> )\r\n> print(f'total size of data is {len(data)}')\r\n> \r\n> \r\n> # splitting dataset into train, validation\r\n> split = 0.2\r\n> train_dataset, eval_dataset = random_split(data, lengths=[int((1-split)*len(data))+1, int(split*len(data))])\r\n> \r\n> \r\n> # defining collator functioon for preparing batches on the fly ..\r\n> def data_collator(features:list):\r\n> inputs = [f[\"seq2seq\"][\"input\"] for f in features]\r\n> batch = tokenizer.prepare_seq2seq_batch(src_texts=inputs, max_length=256, padding='max_length')\r\n> batch[\"labels\"] = batch[\"input_ids\"].copy()\r\n> for k in batch:\r\n> batch[k] = torch.tensor(batch[k])\r\n> return batch\r\n> \r\n> \r\n> batch_out = data_collator(eval_dataset)\r\n> print(batch_out)\r\n> print(batch_out['input_ids'].shape,batch_out['labels'].shape,batch_out['attention_mask'].shape)\r\n> \r\n> \r\n> # defining training related arguments\r\n> args = Seq2SeqTrainingArguments(output_dir=\"clm-checkpoints\",\r\n> do_train=True,\r\n> do_eval=True,\r\n> evaluation_strategy=\"epoch\",\r\n> per_device_train_batch_size=8,\r\n> per_device_eval_batch_size=8,\r\n> learning_rate=5e-5,\r\n> num_train_epochs=1,\r\n> logging_dir=\"./logs\")\r\n> \r\n> \r\n> # defining trainer using 🤗\r\n> trainer = Seq2SeqTrainer(model=model, \r\n> args=args, \r\n> data_collator=data_collator, \r\n> train_dataset=train_dataset, \r\n> eval_dataset=eval_dataset)\r\n> \r\n> \r\n> # ## Training time\r\n> trainer.train()\r\n> # It will take hours to train this model on this dataset\r\n> \r\n> \r\n> # lets save model\r\n> trainer.evaluate(eval_dataset=eval_dataset)\r\n> trainer.save_model(\"clm-checkpoints\")\r\n> ```\r\n\r\nThanks for your code, it really helps.", "I'm most interested in sentence infilling, which this script doesn't really seem to address (though my understanding was that BART training generally involves masking and permutation). Is there an additional step I need to add for the infilling functionality?", "\r\n> We still need to provide a good docstring/notebook for this. It's on our ToDo-List. :-)\r\n> \r\n> Or @sshleifer - is there already something for Bart?\r\n\r\nHi, any update on this? @vanpelt \r\n", "I actually decided to jump over to T5 and use the `run_t5_mlm_flax.py` script. Seems to be working so far, though it's very new, so missing some conveniences.... it sounds like that stuff is underway!", "> I actually decided to jump over to T5 and use the `run_t5_mlm_flax.py` script. Seems to be working so far, though it's very new, so missing some conveniences.... it sounds like that stuff is underway!\r\n\r\n\r\nGreat, I was initially looking at those scripts to get some ideas about the pre-training script, but since then thought the Huggingface guys might have come up with a resource to do this. Apparently, it's still underway! :)\r\n\r\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." ]

[]

[ { "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 } ]

[ "You get the default config file as written on AWS with \r\n```python \r\nconfig = T5Config.from_pretrained(\"t5-small\")\r\n```\r\n\r\nIf you want to change anything on the config after you can update it with a dict of your specifications:\r\n```python \r\nconfig.update(your_dict)\r\n```", "Hello @patrickvonplaten !\r\nI have tried to update the config file using new dict \r\n```\r\nnew_dict={ \"d_ff\": 2048,\r\n \"d_kv\": 64,\r\n \"d_model\": 512,\r\n \"dropout_rate\": 0.1,\r\n \"eos_token_id\": 1,\r\n \"feed_forward_proj\": \"relu\",\r\n \"initializer_factor\": 1.0,\r\n \"is_encoder_decoder\": True,\r\n \"layer_norm_epsilon\": 1e-06,\r\n \"model_type\": \"t5\",\r\n \"num_decoder_layers\": 6,\r\n \"num_heads\": 8,\r\n \"num_layers\": 6,\r\n \"pad_token_id\": 0,\r\n \"relative_attention_num_buckets\": 32,\r\n \"transformers_version\": \"4.5.1\",\r\n \"use_cache\": False,\r\n \"vocab_size\": 32128,\r\n \"num_input_sent\": 4, # TODO: ADDED\r\n \"max_source_len \": 25, # TODO: ADDED\r\n \"max_target_len\": 25, # TODO: ADDED\r\n \r\n }\r\n```\r\nprint(onfig.d_ff) gives 2048\r\n\r\nbut: \r\nprint(max_source_len) gives me an error \"AttributeError: 'T5Config' object has no attribute 'max_source_len'\"\r\n\r\nhow to use the new added configurations after they are added?", "Hey @Arij-Aladel, \r\n\r\nWhat would be the purpose of adding parameters to the config, such as `max_target_len` that are not used in the corresponding model `T5Model`?", "@patrickvonplaten I am trying to build my model depending on t5 hugging face API and I need to build my own config. Whatever the purpose is , the question still what is th benefit of updating the parameters if I can not reuse them?", "Ok thanks I have solved it it was my bad, so sorry! every thing works well" ]

[ { "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 } ]

[ "Hi @tqdo, \r\n\r\nGood question. We will add better explanation for TF T5 soon :-) ", "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", "@tqdo. I open sourced fine-tuning T5 by customizing training loop in Tensorflow2.0+: https://github.com/wangcongcong123/ttt\r\nHope this helps.", "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" ]

[]

[]

[ "Solved : Had to update transformers & torch\r\nthis code is working fine :\r\n```python\r\nfrom transformers import DistilBertTokenizer, DistilBertForQuestionAnswering\r\nimport torch\r\n\r\ntokenizer = DistilBertTokenizer.from_pretrained(\r\n 'distilbert-base-uncased-distilled-squad')\r\nmodel = DistilBertForQuestionAnswering.from_pretrained(\r\n 'distilbert-base-uncased-distilled-squad')\r\ninput_ids = torch.tensor(tokenizer.encode(\r\n question, corpus, max_length=256, add_special_tokens=True)).unsqueeze(0) # Batch size 1\r\nstart_positions = torch.tensor([1])\r\nend_positions = torch.tensor([3])\r\n\r\noutputs = model(input_ids, start_positions=start_positions,\r\n end_positions=end_positions)\r\nloss, start_scores, end_scores = outputs[:3]\r\n\r\nall_tokens = tokenizer.convert_ids_to_tokens(input_ids[0])\r\nanswer = tokenizer.convert_tokens_to_string(\r\n all_tokens[torch.argmax(start_scores): torch.argmax(end_scores) + 1])\r\nprint('*************')\r\nprint(answer)\r\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 } ]

[ "The longest input length a pretrained GPT2 model can treat depends on its `n_position` value. You can look it up here e.g. https://s3.amazonaws.com/models.huggingface.co/bert/gpt2-config.json . \r\nIf you use a pretrained-model you sadly can only treat sequences <= 1024.\r\n\r\nFor you own model you can increase `n_position` and retrain the longer position encoding matrix this way. \r\n\r\nIf you are just interested in the perplexity you could also simply cut the input_ids into smaller input_ids and average the loss over them. It will not exactly be the same, but a good approximation.", "Is it being calculated in the same way for the evaluation of training on validation set?\r\n\r\nSecondly, if we calculate perplexity of all the individual sentences from corpus \"xyz\" and take average perplexity of these sentences? will it be the same by calculating the perplexity of the whole corpus by using parameter \"eval_data_file\" in language model script?", "1) Yes \r\n\r\n2) No -> since you don't take into account the probability `p(first_token_sentence_2 | last_token_sentence_1)`, but it will be a very good approximation.\r\n\r\nHope this answers your question" ]