Datasets:

---

DanielPFlorian

/

Transformers-Github-Issues

like 0

[]

[]

[ "Unfortunatelly I have the same error\r\n\r\nTraceback (most recent call last):\r\n File \"run_language_modeling.py\", line 29, in <module>\r\n from transformers import (\r\nImportError: cannot import name 'DataCollatorForLanguageModeling' from 'transformers' (E:\\PycharmProjects\\Ancient_BERT\\venv\\lib\\site-packages\\transformers\\__init__.py)", "> Unfortunatelly I have the same error\r\n> \r\n> Traceback (most recent call last):\r\n> File \"run_language_modeling.py\", line 29, in\r\n> from transformers import (\r\n> ImportError: cannot import name 'DataCollatorForLanguageModeling' from 'transformers' (E:\\PycharmProjects\\Ancient_BERT\\venv\\lib\\site-packages\\transformers__init__.py)\r\n\r\nI tried to do this (instead of !pip install transformers)\r\n\r\n!git clone https://github.com/huggingface/transformers\r\ncd transformers\r\npip install .\r\n\r\nAnd I get the following error:\r\n\r\n**Traceback (most recent call last):\r\nFile \"run_language_modeling.py\", line 280, in\r\nmain()\r\nFile \"run_language_modeling.py\", line 225, in main\r\nif training_args.do_train\r\nFile \"run_language_modeling.py\", line 122, in get_dataset\r\ntokenizer=tokenizer, file_path=file_path, block_size=args.block_size, local_rank=local_rank\r\nFile \"/usr/local/lib/python3.6/dist-packages/transformers/data/datasets/language_modeling.py\", line 84, in init\r\nassert os.path.isfile(file_path)\r\nAssertionError**\r\n\r\nI think it should be about **tokenizers-0.7.0.**\r\nBut I now still don't know how to fix it.", "Both PyCharm running and example in Colab script has the same problem:\r\nhttps://colab.research.google.com/github/huggingface/blog/blob/master/notebooks/01_how_to_train.ipynb\r\n\r\nTraceback (most recent call last):\r\n File \"run_language_modeling.py\", line 29, in <module>\r\n from transformers import (\r\nImportError: cannot import name 'DataCollatorForLanguageModeling'\r\nCPU times: user 36.2 ms, sys: 22.9 ms, total: 59.1 ms\r\nWall time: 15.2 s", "I change in Colab installation command to: \r\n!pip install git+https://github.com/huggingface/transformers\r\n\r\nand its running :), now only to solve problem with out of memory", "> !pip install git+https://github.com/huggingface/transformers\r\n\r\nYou have to change batch size.\r\n\r\ncmd =\t\"\"\"\r\n python run_language_modeling.py\r\n --train_data_file ./oscar.eo.txt\r\n --output_dir ./EsperBERTo-small-v1\r\n\t--model_type roberta\r\n\t--mlm\r\n\t--config_name ./EsperBERTo\r\n\t--tokenizer_name ./EsperBERTo\r\n\t--do_train\r\n\t--line_by_line\r\n\t--learning_rate 1e-4\r\n\t--num_train_epochs 1\r\n\t--save_total_limit 2\r\n\t--save_steps 2000\r\n\t**--per_gpu_train_batch_size 4**\r\n\t--seed 42\r\n\"\"\".replace(\"\\n\", \" \")\r\n\r\nAnd Thank you for your help.", "Maybe somone had error (error occurs after reaching save step value=2000):\r\n\r\nIteration: 16% 1997/12500 [10:38<49:42, 3.52it/s]\r\nIteration: 16% 1998/12500 [10:39<47:03, 3.72it/s]\r\n \r\n{\"learning_rate\": 8.4e-05, \"loss\": 7.684231301307678, \"step\": 2000}\r\nEpoch: 0% 0/1 [10:39<?, ?it/s]\r\nIteration: 16% 1999/12500 [10:39<49:56, 3.50it/s]04/23/2020 11:55:42 - INFO - transformers.trainer - Saving model checkpoint to ./EsperBERTo-small-v1/checkpoint-2000\r\n04/23/2020 11:55:42 - INFO - transformers.configuration_utils - Configuration saved in ./EsperBERTo-small-v1/checkpoint-2000/config.json\r\n04/23/2020 11:55:42 - INFO - transformers.modeling_utils - Model weights saved in ./EsperBERTo-small-v1/checkpoint-2000/pytorch_model.bin\r\nTraceback (most recent call last):\r\n File \"run_language_modeling.py\", line 280, in <module>\r\n main()\r\n File \"run_language_modeling.py\", line 254, in main\r\n trainer.train(model_path=model_path)\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/trainer.py\", line 363, in train\r\n self._rotate_checkpoints()\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/trainer.py\", line 458, in _rotate_checkpoints\r\n checkpoints_sorted = self._sorted_checkpoints(use_mtime=use_mtime)\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/trainer.py\", line 443, in _sorted_checkpoints\r\n regex_match = re.match(\".*{}-([0-9]+)\".format(checkpoint_prefix), path)\r\n File \"/usr/lib/python3.6/re.py\", line 172, in match\r\n return _compile(pattern, flags).match(string)\r\nTypeError: expected string or bytes-like object\r\n\r\nEpoch: 0% 0/1 [10:40<?, ?it/s]\r\nIteration: 16% 1999/12500 [10:40<56:04, 3.12it/s]\r\nCPU times: user 2.78 s, sys: 928 ms, total: 3.71 s\r\nWall time: 11min 33s\r\n\r\n\r\nMy configuration is:\r\nconfig = {\r\n\t\"architectures\": [\r\n\t\t\"RobertaForMaskedLM\"\r\n\t],\r\n\t\"attention_probs_dropout_prob\": 0.1,\r\n\t\"hidden_act\": \"gelu\",\r\n\t\"hidden_dropout_prob\": 0.1,\r\n\t\"hidden_size\": 768,\r\n\t\"initializer_range\": 0.02,\r\n\t\"intermediate_size\": 3072,\r\n\t\"layer_norm_eps\": 1e-05,\r\n\t\"max_position_embeddings\": 514,\r\n\t\"model_type\": \"roberta\",\r\n\t\"num_attention_heads\": 12,\r\n\t\"num_hidden_layers\": 6,\r\n\t\"type_vocab_size\": 1,\r\n\t\"vocab_size\": 52000\r\n}\r\ncmd = \"\"\"\r\npython run_language_modeling.py\r\n--train_data_file ./oscar.eo.txt\r\n--output_dir ./EsperBERTo-small-v1\r\n--model_type roberta\r\n--mlm\r\n--config_name ./EsperBERTo\r\n--tokenizer_name ./EsperBERTo\r\n--do_train\r\n--line_by_line\r\n--learning_rate 1e-4\r\n--num_train_epochs 1\r\n--save_total_limit 2\r\n--save_steps 2000\r\n--per_gpu_train_batch_size 4 \r\n--seed 42\r\n\"\"\".replace(\"\\n\", \" \")\r\n", "Closing in favor of #3920 ", "How to solve this? Why is this closed?" ]

[ { "id": 1990918270, "node_id": "MDU6TGFiZWwxOTkwOTE4Mjcw", "url": "https://api.github.com/repos/huggingface/transformers/labels/Good%20First%20Issue", "name": "Good First Issue", "color": "bbf794", "default": false, "description": "" } ]

[]

[]

[]

[ { "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 } ]

[ "These `next_token_logits[batch_idx, banned_tokens[batch_idx]]` are logits in the `range(-inf, inf)` not probabilities. You can try to set all other logits except your banned_tokens to `-float(\"inf\")`" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/3909?src=pr&el=h1) Report\n> Merging [#3909](https://codecov.io/gh/huggingface/transformers/pull/3909?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/cb3c2212c79d7ff0a4a4e84c3db48371ecc1c15d&el=desc) will **increase** coverage by `0.01%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/3909?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #3909 +/- ##\n==========================================\n+ Coverage 78.45% 78.47% +0.01% \n==========================================\n Files 111 111 \n Lines 18521 18521 \n==========================================\n+ Hits 14531 14534 +3 \n+ Misses 3990 3987 -3 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/3909?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/3909/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `91.06% <0.00%> (+0.12%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/3909/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `92.92% <0.00%> (+0.16%)` | :arrow_up: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/3909/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `68.86% <0.00%> (+0.36%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/3909?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/3909?src=pr&el=footer). Last update [cb3c221...3361612](https://codecov.io/gh/huggingface/transformers/pull/3909?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1845609017, "node_id": "MDU6TGFiZWwxODQ1NjA5MDE3", "url": "https://api.github.com/repos/huggingface/transformers/labels/seq2seq", "name": "seq2seq", "color": "fef2c0", "default": false, "description": "" }, { "id": 2009457320, "node_id": "MDU6TGFiZWwyMDA5NDU3MzIw", "url": "https://api.github.com/repos/huggingface/transformers/labels/translation", "name": "translation", "color": "b2d2f4", "default": false, "description": "machine translation utilities and models" } ]

[]

[ "I think that's awesome! Such little code changes to include that many models is amazing :-) \r\nDoes the model with the pre-trained weights produce 1-to-1 the same results as the official Marian models in C++?", "@patrickvonplaten the generations are very sensible, but not identical to marian, due to slightly different generation parameters: \r\n\r\nEquivalent (4/7):\r\n\r\nC+: Ich bin ein kleiner Frosch.\r\nHF: Ich bin ein kleiner Frosch.\r\n\r\nC+: Tom bat seinen Lehrer um Rat.\r\nHF: Tom bat seinen Lehrer um Rat.\r\n\r\nC+: Tom bewunderte Marias Mut wirklich.\r\nHF: Tom bewunderte Marias Mut wirklich.\r\n\r\nC+: Umdrehen und die Augen schließen.\r\nHF: Umdrehen und die Augen schließen.\r\n\r\nNot Equivalent (3/7):\r\nC+: Jetzt kann ich die 100 Wörter vergessen, die ich kenne.\r\nHF: Jetzt kann ich die 100 Wörter **des Deutschen** vergessen, die ich kenne.\r\n\r\nC+: **O** (Input=\"O\")\r\nHF: \r\n\r\nC+: So würde ich das **machen**.\r\nHF: So würde ich das **tun**.\r\n\r\nI'm investigating.", "> @patrickvonplaten the generations are very sensible, but not identical to marian, due to slightly different generation parameters:\r\n> \r\n> Equivalent (4/7):\r\n> \r\n> C+: Ich bin ein kleiner Frosch.\r\n> HF: Ich bin ein kleiner Frosch.\r\n> \r\n> C+: Tom bat seinen Lehrer um Rat.\r\n> HF: Tom bat seinen Lehrer um Rat.\r\n> \r\n> C+: Tom bewunderte Marias Mut wirklich.\r\n> HF: Tom bewunderte Marias Mut wirklich.\r\n> \r\n> C+: Umdrehen und die Augen schließen.\r\n> HF: Umdrehen und die Augen schließen.\r\n> \r\n> Not Equivalent (3/7):\r\n> C+: Jetzt kann ich die 100 Wörter vergessen, die ich kenne.\r\n> HF: Jetzt kann ich die 100 Wörter **des Deutschen** vergessen, die ich kenne.\r\n> \r\n> C+: **O** (Input=\"O\")\r\n> HF:\r\n> \r\n> C+: So würde ich das **machen**.\r\n> HF: So würde ich das **tun**.\r\n> \r\n> I'm investigating.\r\n\r\nSounds all very coherent :-) ", "Merging. Will add docs in next PR, when I add tons of models.", "@sshleifer, can you elaborate on \"investigate discrepancies in generation parameters: length_penalty, decoder_start_token_id, etc. These may explain small differences in generations (detailed in comment below)\"? I see https://github.com/huggingface/transformers/pull/3908#issuecomment-618473222 but was this resolved?", "_TLDR Was not resolved, the investigation continues._\r\nOur lib has many parameters to control generations, `length_penalty`, `repetition_penalty`, `bad_word_ids`, `decoder_start_token_id`.\r\nThey are documented [here](https://huggingface.co/transformers/model_doc/bart.html?#transformers.BartForConditionalGeneration.generate).\r\nThe marian lib has `normalize` and `word-penalty`, which are not identical, but similar. Part of the discrepancy may involve configuring our parameters to be closer to theirs.\r\n\r\n", "I see, thanks!", "Could anyone help with this issue: #5040 ?" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/3907?src=pr&el=h1) Report\n> Merging [#3907](https://codecov.io/gh/huggingface/transformers/pull/3907?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/cb3c2212c79d7ff0a4a4e84c3db48371ecc1c15d&el=desc) will **not change** coverage.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/3907?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #3907 +/- ##\n=======================================\n Coverage 78.45% 78.45% \n=======================================\n Files 111 111 \n Lines 18521 18521 \n=======================================\n Hits 14531 14531 \n Misses 3990 3990 \n```\n\n\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/3907?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/3907?src=pr&el=footer). Last update [cb3c221...f75b213](https://codecov.io/gh/huggingface/transformers/pull/3907?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[]

[]

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

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[ "It will load the weights, but will not load the multiple choice head as it's not in the checkpoint. That's why fine-tuning it is a good idea, so that this head's weights get trained!" ]

[]

[ { "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 @yuvalkirstain,\r\n\r\n\r\nThanks for your message! \r\n\r\nWhen you put in `input_ids` and `lm_labels` at the some time the following happens:\r\nthe `decoder_input_ids` are created automatically depending on the `lm_labels` (shift `lm_labels` to the right and pre-prend `<PAD>`) and look as follows: \r\n`<PAD> <extra_id_1> cute dog <extra_id_2> the <extra_id_3>`. \r\n\r\nsee https://github.com/huggingface/transformers/blob/4e817ff41885063e08bb3bcd63e5adfd835b9911/src/transformers/modeling_t5.py#L1060 \r\n\r\nNow in order for `<extra_id_1>` to show up, it should be guessed from `<PAD>`, but I'm not sure whether this makes much sense. After `<extra_id_1>` has been processed though from the input, it's much more likely that another `<extra_id_2>` is in the labels, so it does make more sense to me that `<extra_id_2>` is guessed.\r\n\r\nLet me know if it does not makes sense to you :-) \r\n", "ahh ok. Thank you so much and sorry that I bothered you with it. Really enjoying working with it so far :)" ]

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

[]

[ "How big are your batches? BERT can only accept tensors of maximum length of 512. Is it possible one of your inputs is longer than 512?", "I had the same issue while my `per_gpu_eval_batch_size=8` and `per_gpu_train_batch_size=8 `. ", "Yes, it's unrelated to batch size. Is your sequence length longer than 512?", "Also having this issue, getting a different error when I try to use GPU but I understand that the error is caused by the same thing and the CPU stack trace is just more informative. My sequence lengths are definitely <512", "> I had the same issue while my `per_gpu_eval_batch_size=8` and `per_gpu_train_batch_size=8 `.\r\n\r\nI got the same error when I use batch_size = 8. After I change it to 64, then no errors. ", "> Also having this issue, getting a different error when I try to use GPU but I understand that the error is caused by the same thing and the CPU stack trace is just more informative. My sequence lengths are definitely <512\r\n\r\ndid you find the solution?\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", " return torch.embedding(weight, input, padding_idx, scale_grad_by_freq, sparse)\r\n\r\nRuntimeError: index out of range: Tried to access index 0 out of table with 202 rows. at /pytorch/aten/src/TH/generic/THTensorEvenMoreMath.cpp:237\r\n\r\nis there anyone knows what's the problem ? , I checked the embedding matrix of shape (203, 16), it shows index 0 out of range, how it is possible???" ]

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

[]

[ "How did you obtain your pre-trained model? Does your pre-train model have a classification head? What is `model_class`? Is it `BertForSequenceClassification`?", "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 } ]

[]

[ "I assume you concatenate all paragraphs in a single text string, encode it with your question, and give the string as an input to your QA model (or pipeline). What you get back, is an answer span: indexes of the first and last character of the answer in the input text. To find a paragraph it belongs to, compute indexes of the paragraph spans in your concatenated text strings. ", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[]

[]

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

[ "DialoGPT is fine-tuned in the same way GPT2 in fine-tuned. Please take a look at the paper: \r\nhttps://arxiv.org/pdf/1911.00536.pdf . \r\n\r\nI will also add some training tips to the doc string." ]

[]

[]

[ "Duplicate of #3893 \r\n\r\nPlease install from source until we release a new version of the library on PyPI" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null }, { "id": 1990918270, "node_id": "MDU6TGFiZWwxOTkwOTE4Mjcw", "url": "https://api.github.com/repos/huggingface/transformers/labels/Good%20First%20Issue", "name": "Good First Issue", "color": "bbf794", "default": false, "description": "" } ]

[]

[ "Sorry, I think it is the same issue as #4103. ", "Please don't open duplicate issues. If you have more info about this issue, post it here. Also, please use [code blocks](https://help.github.com/en/github/writing-on-github/creating-and-highlighting-code-blocks).", "> Please don't open duplicate issues. If you have more info about this issue, post it here. Also, please use [code blocks](https://help.github.com/en/github/writing-on-github/creating-and-highlighting-code-blocks).\r\n\r\nSorry for this.", "Facing the same issue - \r\nEnvironment - \r\n\r\n- transformers version:2.2.2\r\n- Platform: \r\n- Python version:3.7.3\r\n- PyTorch version (GPU?): 1.3.1 (no GPU)\r\n- Tensorflow version (GPU?): 1.14.0 (no GPU)\r\n- Using GPU in script?:No\r\n- Using distributed or parallel set-up in script?: No\r\n\r\nWhile computing the initial importance score (When the head_mask is None) I am getting the following error. -\r\n \r\n File \"/Users/user/Desktop/org//prune_attention_heads.py\", line 226, in compute_heads_importance\r\n head_importance += head_mask.grad.abs().detach()\r\nAttributeError: 'NoneType' object has no attribute 'abs'\r\n\r\nOn putting the line above in a try and except block and printing the head_mask I get the following - \r\n\r\ntensor([[1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.],\r\n [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.],\r\n [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.],\r\n [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.],\r\n [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.],\r\n [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.],\r\n [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.],\r\n [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.],\r\n [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.],\r\n [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.],\r\n [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.],\r\n [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.]], requires_grad=True)\r\n\r\nHere, the head_mask has requires_grad=True and I am passing the head_mask into the model and calling loss.backward() as in the bertology script. \r\n", "I met the same problem, the head_mask has no grad at the second time.", "Seems that during the second pruning, the head_mask tensor becomes a non-leaf node, when the grad is None, I printed the `head_mask.is_leaf` attribute and get the warning(PyTorch 1.5.0) as below:\r\n> UserWarning: The .grad attribute of a Tensor that is not a leaf Tensor is being accessed. Its .grad attribute won't be populated during autograd.backward(). If you indeed want the gradient for a non-leaf Tensor, use .retain_grad() on the non-leaf Tensor. If you access the non-leaf Tensor by mistake, make sure you access the leaf Tensor instead. See github.com/pytorch/pytorch/pull/30531 for more informations.\r\n warnings.warn(\"The .grad attribute of a Tensor that is not a leaf Tensor is being accessed. Its .grad \"\r\nhead_mask is leaf: False", "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" ]

[]

[]

[ "Well, it looks like you don't have the \"mnli/dev_matched.tsv\" file downloaded? You can download the GLUE datasets using [utils/download_glue_data.py](https://github.com/huggingface/transformers/blob/master/utils/download_glue_data.py)" ]

[]

[]

[ "I was installing from pip and it isn't updated yet. Building from source solved this.", "I am trying to run this from source and still I am getting the same error! \r\nPlease have a look at this [here](https://github.com/huggingface/transformers/issues/3896#issue-604681806)", "It's because the pip package hasn't been updated. The script to train is changed fundamentally. so you can try building from scratch using \r\n`git clone https://github.com/huggingface/transformers\r\ncd transformers\r\npip install .` \r\nor \r\nYou can use old script of `run_language_modeling.py` from previous commits. " ]

[]

[ { "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 } ]

[ "No good reason.\r\nDo you feel comfortable sending a PR that shuffles for train loader only?\r\n" ]

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

[ { "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 @aced125, I agree that this functionality should be provided :-) \r\n\r\nI think in the PR, one has to include a `output_attention` argument to the `generate()` function and then make sure that the output idx are correct! \r\nBefore starting this PR, this PR should probably be solved before: https://github.com/huggingface/transformers/issues/3880", "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", "want to take a look at this 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", "#6735 is a first step to allow for this feature", "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/3890?src=pr&el=h1) Report\n> Merging [#3890](https://codecov.io/gh/huggingface/transformers/pull/3890?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eb5601b0a5a88824a2598956f96e06e7f2422bce&el=desc) will **decrease** coverage by `0.03%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/3890?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #3890 +/- ##\n==========================================\n- Coverage 78.57% 78.53% -0.04% \n==========================================\n Files 106 106 \n Lines 17962 17962 \n==========================================\n- Hits 14113 14106 -7 \n- Misses 3849 3856 +7 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/3890?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/3890/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `91.28% <0.00%> (-1.32%)` | :arrow_down: |\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/3890/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `91.06% <0.00%> (+0.12%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/3890?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/3890?src=pr&el=footer). Last update [eb5601b...fe606bb](https://codecov.io/gh/huggingface/transformers/pull/3890?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

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

[]

[]

[ { "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 } ]

[]

[ "I get the below error:\r\n\r\n```\r\nValidation sanity check: 0%| | 0/5 [00:00<?, ?it/s]Traceback (most recent call last):\r\n File \"run_pl_glue.py\", line 186, in <module>\r\n trainer = generic_train(model, args)\r\n File \"/home/jupyter/transformers/examples/transformer_base.py\", line 307, in generic_train\r\n trainer.fit(model)\r\n File \"/opt/conda/lib/python3.7/site-packages/pytorch_lightning/trainer/trainer.py\", line 701, in fit\r\n self.dp_train(model)\r\n File \"/opt/conda/lib/python3.7/site-packages/pytorch_lightning/trainer/distrib_parts.py\", line 540, in dp_train\r\n self.run_pretrain_routine(model)\r\n File \"/opt/conda/lib/python3.7/site-packages/pytorch_lightning/trainer/trainer.py\", line 843, in run_pretrain_routine\r\n False)\r\n File \"/opt/conda/lib/python3.7/site-packages/pytorch_lightning/trainer/evaluation_loop.py\", line 262, in _evaluate\r\n output = self.evaluation_forward(model, batch, batch_idx, dataloader_idx, test_mode)\r\n File \"/opt/conda/lib/python3.7/site-packages/pytorch_lightning/trainer/evaluation_loop.py\", line 430, in evaluation_forward\r\n output = model(*args)\r\n File \"/opt/conda/lib/python3.7/site-packages/torch/nn/modules/module.py\", line 532, in __call__\r\n result = self.forward(*input, **kwargs)\r\n File \"/opt/conda/lib/python3.7/site-packages/pytorch_lightning/overrides/data_parallel.py\", line 66, in forward\r\n return self.gather(outputs, self.output_device)\r\n File \"/opt/conda/lib/python3.7/site-packages/torch/nn/parallel/data_parallel.py\", line 165, in gather\r\n return gather(outputs, output_device, dim=self.dim)\r\n File \"/opt/conda/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py\", line 68, in gather\r\n res = gather_map(outputs)\r\n File \"/opt/conda/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py\", line 62, in gather_map\r\n for k in out))\r\n File \"/opt/conda/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py\", line 62, in <genexpr>\r\n for k in out))\r\n File \"/opt/conda/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py\", line 55, in gather_map\r\n return Gather.apply(target_device, dim, *outputs)\r\n File \"/opt/conda/lib/python3.7/site-packages/torch/nn/parallel/_functions.py\", line 54, in forward\r\n assert all(map(lambda i: i.is_cuda, inputs))\r\nAssertionError\r\n```\r\n@nateraw @williamFalcon", "update to the latest lightning version?\r\n0.7.4rc1", "@williamFalcon doesn't work with lightning version 0.7.4rc1, 0.7.4rc2 and even 0.7.3, 0.7.1\r\n", "ok, can you share a colab here? happy to take a look", "@williamFalcon Thanks. I'm running the code as per the given instructions in https://github.com/huggingface/transformers/tree/master/examples/glue\r\nI didn't make any changes, I just ran the same official example script in multi gpu's - https://github.com/huggingface/transformers/blob/master/examples/glue/run_pl.sh \r\nIt works in CPU and single GPU, but doesn't work in multi-gpu's ", "It is a bit unclear what is going on in there: the bash script installs lightning but the python code doesn't seem to use it?", "I am also facing the error but on a different custom learning model. My code is working properly on a single GPU, however, if I increase the number of GPUs to 2, it gives me the above error. I checked both PL 0.7.3 and 0.7.4rc3 \r\n\r\n**Update: Interestingly when I changed ``distributed_backend`` to ``ddp`` then it worked perfectly without any error** I think there is an issue with the **``dp``** distributed_backend", "\r\n\r\nrun_pl.sh runs fine.\r\n\r\nI ran without ANY changes to the file. Did you guys change anything in the file?", "@williamFalcon Didn't change anything, hope you ran it in multi-gpu's. The code seems to run fine in ddp, but not in dp, as mentioned by @mmiakashs . \r\n\r\nWhen I debugged, I found that when using dp (DataParallel) with 8 gpu's, it generates 8 different losses and since the training_step can't gather 8 losses, it showed error like this:\r\n``` TypeError: zip argument #1 must support iteration ```\r\n", "Ummm, yeah not sure... It looks ok to me.\r\n\r\n\r\n\r\n\r\n\r\n\r\nTry running dp on 2 GPUs? This test is on 2 GPUs", "It looks like hf sets ddp as the backend which is great because dp has a bunch of issues (this is a PyTorch problem, not lightning). Both PyTorch and lightning discourage dp use.\r\n\r\nJust ran this with the default ddp and it works well (although the run_pl.sh script has a bunch of usability issues, ie: i need the data in a different part of the cluster but that script doesn't do that, so I had to run from that directory in the cluster. Ideally --data_dir solves this issue but it doesn't).", "I can confirm that the issue occurs only when using multi-gpu's with dp as backend. Using ddp solves the issues.\r\n\r\nI found one more issue. If I use fast tokenizers with ddp as backend, I get the below error:\r\n\r\n```\r\nINFO:lightning:GPU available: True, used: True\r\nINFO:lightning:CUDA_VISIBLE_DEVICES: [0,1]\r\n/opt/conda/lib/python3.7/site-packages/pytorch_lightning/utilities/warnings.py:18: RuntimeWarning: You have defined a `val_dataloader()` and have defined a `validation_step()`, you may also want to define `validation_epoch_end()` for accumulating stats.\r\n warnings.warn(*args, **kwargs)\r\nTraceback (most recent call last):\r\n File \"run_pl_glue.py\", line 187, in <module>\r\n trainer = generic_train(model, args)\r\n File \"/home/jupyter/transformers/examples/transformer_base.py\", line 310, in generic_train\r\n trainer.fit(model)\r\n File \"/opt/conda/lib/python3.7/site-packages/pytorch_lightning/trainer/trainer.py\", line 734, in fit\r\n mp.spawn(self.ddp_train, nprocs=self.num_processes, args=(model,))\r\n File \"/opt/conda/lib/python3.7/site-packages/torch/multiprocessing/spawn.py\", line 162, in spawn\r\n process.start()\r\n File \"/opt/conda/lib/python3.7/multiprocessing/process.py\", line 112, in start\r\n self._popen = self._Popen(self)\r\n File \"/opt/conda/lib/python3.7/multiprocessing/context.py\", line 284, in _Popen\r\n return Popen(process_obj)\r\n File \"/opt/conda/lib/python3.7/multiprocessing/popen_spawn_posix.py\", line 32, in __init__\r\n super().__init__(process_obj)\r\n File \"/opt/conda/lib/python3.7/multiprocessing/popen_fork.py\", line 20, in __init__\r\n self._launch(process_obj)\r\n File \"/opt/conda/lib/python3.7/multiprocessing/popen_spawn_posix.py\", line 47, in _launch\r\n reduction.dump(process_obj, fp)\r\n File \"/opt/conda/lib/python3.7/multiprocessing/reduction.py\", line 60, in dump\r\n ForkingPickler(file, protocol).dump(obj)\r\nTypeError: can't pickle Tokenizer objects\r\n```", ">\r\n> I found one more issue. If I use fast tokenizers with ddp as backend, I get the below error:\r\n> \r\n@leslyarun I am also facing a similar issue with ddp backend (not exactly the same): [github issue](https://github.com/PyTorchLightning/pytorch-lightning/issues/1578)\r\nMy guess is that maybe there is an issue with the callback and the saving objects with pickle. At this moment I will try to manually save checkpoint without using the callbacks.\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", "@mmiakashs did that end up working?", "> @mmiakashs did that end up working?\r\n\r\ncurrently, I am using ddp_spwan mode and it is working fine. ", "@sshleifer can confirm A) the Lightning examples don't work at all with `dp` B) does run, but needs significant editing with `ddp`\r\n\r\nFor examples I've looked at it's not as simple as turning `ddp` on and all great. It seems whomever wrote the Lightning examples never tried multi-GPU. Happy to elaborate or share (though mine are not in great shape at the moment).\r\n\r\nAnd `ddp_spawn` definitely does not work for me. Gives several spawn-based errors -- says my model is not compliant. ", "A) don't know but that sounds very likely. @williamFalcon told me \"Dont use dp\".\r\n\r\nB) `examples/seq2seq/finetune.py` works in multigpu with two caveats:\r\n(a) versions need to be transformers=master, pl=0.8.1. \r\n(b) you cannot pass `--do_predict`. (`pl.Trainer.test` is broken for multi-gpu)\r\n\r\nFor the other two pl examples: ner, and glue, I haven't tested multi-gpu, but they should be at least close to working because they inherit from the same `BaseTransformer`. Which one of those were you trying to run/ are you interesting in running?\r\n\r\n", "Thanks @sshleifer. We're fine using `ddp` for everything -- only need one version to work, not multiple ways to do the same thing. Also according to the docs, `ddp` is the only one that works with FP16 anyway (have not tested yet, will do soon).\r\nhttps://pytorch-lightning.readthedocs.io/en/latest/multi_gpu.html\r\n\r\nI'm working off of `transformers` from GitHub... so should be a recent version. If that's not what you are saying couple you please be more specific?\r\n\r\nWe also don't necessarily \"need\" Lightning... but would be great if it worked (in single set of settings) for multi-GPU. As it is great having reasonable out of the box options for LR schedule, model synchronization, gradient accumulation, and all those other things I've grown tired of implementing for every project. ", "@moscow25 dp is NOT recommended by PyTorch\r\n\r\nhttps://pytorch.org/docs/master/generated/torch.nn.DataParallel.html\r\n\r\n\r\n\r\n2. The current base transformers has a few issues which I've submitted a PR for.\r\n3. Please let me know what example you are using / what code i can look at to reproduce the issues.", "> @sshleifer can confirm A) the Lightning examples don't work at all with `dp` B) does run, but needs significant editing with `ddp`\r\n> \r\n> For examples I've looked at it's not as simple as turning `ddp` on and all great. It seems whomever wrote the Lightning examples never tried multi-GPU. Happy to elaborate or share (though mine are not in great shape at the moment).\r\n> \r\n> And `ddp_spawn` definitely does not work for me. Gives several spawn-based errors -- says my model is not compliant.\r\n\r\nddp doesn't work for me and ddp_spawn gives a lot of errors. On using ddp, no error is shown but it doesn't start anything on the GPU - just the notebook cell being busy indefinitely. I am using the DistilBertTokenizer and DistilBertModel - has anyone been able to run pytorch lightning on multipe gpus with Distilbert?", "I suspect that your issue is ddp+jupyter rather than distillbert. Try running your command from the terminal.", "> I suspect that your issue is ddp+jupyter rather than distillbert. Try running your command from the terminal.\r\n\r\nWhy does running the code in Jupyter notebook create a problem? I was able to run the BertModels like SequenceClassification in the Jupyter notebook on multiple gpus without any problem - but running into this multiple gpu problem using pytorch lightning. It is nice to be able to use Pytorch lightning given all the built in options. It makes it easier to build the models interactively on the Jupyter notebook", "> > I suspect that your issue is ddp+jupyter rather than distillbert. Try running your command from the terminal.\r\n> \r\n> Why does running the code in Jupyter notebook create a problem? I was able to run the BertModels like SequenceClassification in the Jupyter notebook on multiple gpus without any problem - but running into this multiple gpu problem using pytorch lightning. It is nice to be able to use Pytorch lightning given all the built in options. It makes it easier to build the models interactively on the Jupyter notebook\r\n\r\nLooks like usage of ddp doesn't work in Jupyter notebook. and transformers don't work with dp parameter of pytorch lightning in Jupyter notebook. So looks like the only option to use pytorch lightning, multiple gpus and transformer is to run it as a python script.\r\nhttps://pytorch-lightning.readthedocs.io/en/latest/multi_gpu.html\r\nJupyter Notebooks\r\nUnfortunately any ddp_ is not supported in jupyter notebooks. Please use dp for multiple GPUs. This is a known Jupyter issue. If you feel like taking a stab at adding this support, feel free to submit a PR!", "i believe @nateraw is almost done updating the examples with the latest version of PL. \r\n\r\ncan you share the model that does work with multiple gpus in a jupyter notebook?", "I read somewhere on the pytorch lightning documents about being careful to checkpoint models when running on DDP mode - can't find that documentation now but is there something I need to be careful about checkpointing while running DDP on a single machine with 8 GPUs? It was something about the model getting split among multiple machines - not sure if that is valid if DDP used on a single machine. ", "nothing you have to worry about... we save the checkpoint correctly automatically ", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

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

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

[ "It won't be that easy since some words will be split into multiple tokens so you have to make two forward passes. \r\n\r\nIf you limit your `original_word` to just one token words (you can check that simply with `len(tokenizer.encode(original_word))==1`. Then your idea here should work. \r\n\r\nIf not it's gonna be trickier. Also this issue might be helpful: \r\nhttps://github.com/huggingface/transformers/issues/2311", "Thanks @patrickvonplaten for your response. \r\nYes, the code works for `len(tokenizer.encode(original_word))==1`, but not for those `original_word` , which consist of more than one tokens.\r\n\r\nI look at the shared issue, but I am confused, which selected word id, should I pass to the model again, as `next_word_logits.topk(5)` gives me 5 token ids?\r\n\r\nCan you please share any code snippet, which will work for the second part?", "Hi @patrickvonplaten,\r\n\r\ncan u plz let me know about any update?", "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" ]

[]

[]

[ "Seems fixed now--delete transformers installed via pip and install by cloning this repo." ]

[]

[]

[ "Never mind; this was an issue with the colab. It's sorted now." ]

[ { "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" } ]

[]

[ "I there any problem with this card?" ]

[]

[]

[ "(Hmm, GitHub was failing earlier today, and now it seems to have posted my comment multiple times. Sorry about that.)", "See cbbb3c43c55d2d93a156fc80bd12f31ecbac8520" ]

[ { "id": 1990918270, "node_id": "MDU6TGFiZWwxOTkwOTE4Mjcw", "url": "https://api.github.com/repos/huggingface/transformers/labels/Good%20First%20Issue", "name": "Good First Issue", "color": "bbf794", "default": false, "description": "" } ]

[]

[ "Hi, I would like to work on this issue.", "That's great :-) Do you want to open a PR and do the changes analogous to PR: #3734 ? ", "Is this still be worked on? If not, I'd be happy to make a first contribution here", "First PR first serve ;-) Still an open issue", "Any tips on how I should proceed? I was thinking of following the changes made for `config.output_past` (01c37dc), but for `config.output_attentions` instead.", "Oh sorry @drjosephliu didn't notice the comment as I was working on it earlier today, my apologies 😞 ", "Hey @patrickvonplaten, i noticed this issue has been closed. Any updates on what changes were made and any updates to the PR i still need to make?", "Hey @drjosephliu, I will take a closer look at your PR tomorrow :-) " ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null }, { "id": 1990918270, "node_id": "MDU6TGFiZWwxOTkwOTE4Mjcw", "url": "https://api.github.com/repos/huggingface/transformers/labels/Good%20First%20Issue", "name": "Good First Issue", "color": "bbf794", "default": false, "description": "" } ]

[]

[ "Hi, @patrickvonplaten I want to take up this issue. Can I move forward with it? ", "I think this could have side effects for libraries that use `config.output_hidden_states`, so I'm cc'ing @Timoeller and @brandenchan, because this parameter is used in [FARM](https://github.com/deepset-ai/FARM).", "> Hi, @patrickvonplaten I want to take up this issue. Can I move forward with it?\r\n\r\nThat would be great, feel free to open a PR and do a first model. The PR should be very similar to what was done in PR #3734", "Hi, @patrickvonplaten as I am new here I might take some time to get acquainted with the codebase and come up with a PR. Is it okay?", "Sorry @gaurav-singh1998, I saw this issue was still open so I made a PR for it.", "Okay, no issues @drjosephliu I'll find some other good first issues to solve. ", "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 } ]

[]

[ "Working on it as we speak :).\r\n\r\nI'd say it will be out in a few weeks at most.", "Is Albert pretraining from scratch is available? @LysandreJik \r\n", "@LysandreJik do you think the update will be available by the end of the month ? Maybe it has been postponed due to the recent addition of the Trainer and the refactor of the language_modeling script ?", "It has been postponed a bit due to the recent addition of the Trainer and the TPU work on it, but I definitely aim to have it out earlier than by the end of the month :)", "Was looking for Albert pre-training from scratch but I think there is support for Bert, Roberta and distillbert only as of now. \r\n@LysandreJik can you guide how can I do Albert pretraining from scratch?", "@LysandreJik Is the code for pretraining Electra from scratch available now?\r\n", "> @LysandreJik Is the code for pretraining Electra from scratch available now?\r\n\r\nNot yet. There's a PR about it;\r\nhttps://github.com/huggingface/transformers/pull/4656", "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", "Are there any updates to this, or plans to release the ELECTRA pre-training from scratch feature 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", "Hello. \r\nAre there any updates?", "Curious as well.", "The development of the ELECTRA pretraining from scratch is in a stale state with no plans to work on it further, see https://github.com/huggingface/transformers/pull/4656#issuecomment-711082850\r\n\r\nSee https://discuss.huggingface.co/t/electra-training-reimplementation-and-discussion/1004 by @richarddwang for a PyTorch implementation of the ELECTRA pretraining." ]

[]

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

[ "Should be fixed on master, please share if it does/your results" ]

[ { "id": 1260952223, "node_id": "MDU6TGFiZWwxMjYwOTUyMjIz", "url": "https://api.github.com/repos/huggingface/transformers/labels/Discussion", "name": "Discussion", "color": "22870e", "default": false, "description": "Discussion on a topic (keep it focused or open a new issue though)" }, { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "BART is used as a model for abstractive summarization so it can use different words than those used in the original text. But it should not go *off-topic*. You could use an extractive summarization model instead which does not generate new nouns. Also you might be interested in the methods from [*Controlling the Amount of Verbatim Copying in Abstractive Summarization*](https://arxiv.org/pdf/1911.10390.pdf) to control the degree of change in the summaries.", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

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

[ "Why do you use `decoder_start_token_id = tokenizer.eos_token_id` ? Is that stated in the examples somewhere? \r\n\r\nIf you do:\r\n```\r\nfrom transformers import T5Tokenizer, T5ForConditionalGeneration\r\ntokenizer = T5Tokenizer.from_pretrained('t5-base')\r\nmodel = T5ForConditionalGeneration.from_pretrained('t5-base')\r\ndata=\"translate English to German: Hello, my dog is cute\"\r\ninput_ids = tokenizer.encode(data, return_tensors=\"pt\") # Batch size 1\r\noutputs = model.generate(input_ids, decoder_start_token_id = tokenizer.eos_token_id)\r\nprint(tokenizer.decode(outputs[0]))\r\n```\r\n\r\nThe translation is correct. Also I recommend using the translation pipeline. This way T5 uses better generation paramaters.", "@patrickvonplaten Thanks a lot ! I used previous configuration file from https://s3.amazonaws.com/models.huggingface.co/bert/t5-base-config.json ; After using the new configuration file, translation error is gone ! \r\nMy code is : \r\n```from transformers import T5Tokenizer, T5ForConditionalGeneration\r\ntokenizer = T5Tokenizer.from_pretrained('t5-base')\r\nmodel = T5ForConditionalGeneration.from_pretrained('t5-base')\r\ndata=\"translate English to German: Hello, my dog is cute\"\r\ninput_ids = tokenizer.encode(data, return_tensors=\"pt\") # Batch size 1\r\noutputs = model.generate(input_ids)\r\nprint(tokenizer.decode(outputs[0]))" ]

[ { "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" } ]

[]

[ "It is! (though ideally you would add an example of use + details about training)\r\n\r\nWill merge this unless you add more in the next 24 hours.", "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/3874?src=pr&el=h1) Report\n> Merging [#3874](https://codecov.io/gh/huggingface/transformers/pull/3874?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/b1ff0b2ae7d368b7db3a8a8472a29cc195d278d8&el=desc) will **not change** coverage.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/3874?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #3874 +/- ##\n=======================================\n Coverage 78.57% 78.57% \n=======================================\n Files 106 106 \n Lines 17962 17962 \n=======================================\n Hits 14114 14114 \n Misses 3848 3848 \n```\n\n\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/3874?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/3874?src=pr&el=footer). Last update [b1ff0b2...8a40fb1](https://codecov.io/gh/huggingface/transformers/pull/3874?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "that makes a lot of sense. i'll update that. thank you", "[model page](https://huggingface.co/spentaur/yelp)" ]

[]

[]

[ "I think you are correct. Any way you'd open a PR? Otherwise we'll get on it in the next few days/weeks" ]

[]

[]

[]

[ { "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 } ]

[]

[ "HELLO eeegnu, I'm also facing the same issue.\r\n\r\n**Environment Info :**\r\nPlatform: Windows 10 64bit\r\nPython version: 3.6.10\r\nPyTorch version (GPU?): 1.4.0 \r\nTensorflow version (GPU?): not installed (NA)\r\n\r\n\r\n", "Hi!\r\nChange: `from .utils_summarization import (\r\n CNNDMDataset,\r\n build_mask,\r\n compute_token_type_ids,\r\n encode_for_summarization,\r\n truncate_or_pad,\r\n)` \r\nTo: `from utils_summarization import (\r\n CNNDMDataset,\r\n build_mask,\r\n compute_token_type_ids,\r\n encode_for_summarization,\r\n truncate_or_pad,\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" ]

[]

[]

[ "Hi, \r\n\r\nAs mentioned [here](https://github.com/huggingface/transformers/tree/master/examples#examples) and in the main README you need to install from source in order to run the latest versions of the examples" ]

[ { "id": 1834081910, "node_id": "MDU6TGFiZWwxODM0MDgxOTEw", "url": "https://api.github.com/repos/huggingface/transformers/labels/Usage", "name": "Usage", "color": "e28436", "default": false, "description": "General questions about the library" } ]

[]

[ "I'm assuming you read https://blog.einstein.ai/benchmarking-tensorrt-inference-server/ ?", "> \r\n> \r\n> I'm assuming you read https://blog.einstein.ai/benchmarking-tensorrt-inference-server/ ?\r\nok: torch.jit.trace(bert)\r\nFailed :torch.jit.script\r\nFailed:torch.jit.trace(bert+after)\r\nFailed: to onnx: \r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/modeling_bert.py\", line 686, in forward\r\n extended_attention_mask = extended_attention_mask.to(dtype=next(self.parameters()).dtype) # fp16 compatibility\r\nUnboundLocalError: local variable 'extended_attention_mask' referenced before assignment \r\n（pip transformers==2.2.0， fixed latested source）\r\n\r\nsolved by remove all none default option argments. minize uncertainty\r\n\r\n" ]

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

[]

[ ">Furthermore, I am also curious about what these 'Ġ' characters are in the RoBERTa encoding?\r\n\r\nIt's a feature of byte-level BPE (an encoded space character)\r\n[Ref-bart-fairseq](https://github.com/pytorch/fairseq/issues/1716), [Ref-openai-gpt](https://github.com/openai/gpt-2/issues/80)", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/3866?src=pr&el=h1) Report\n> Merging [#3866](https://codecov.io/gh/huggingface/transformers/pull/3866?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/a21d4fa410dc3b4c62f93aa0e6bbe4b75a101ee9&el=desc) will **increase** coverage by `0.00%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/3866?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #3866 +/- ##\n=======================================\n Coverage 78.61% 78.62% \n=======================================\n Files 106 106 \n Lines 17953 17953 \n=======================================\n+ Hits 14114 14115 +1 \n+ Misses 3839 3838 -1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/3866?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/3866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `91.06% <0.00%> (+0.12%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/3866?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/3866?src=pr&el=footer). Last update [a21d4fa...c84984c](https://codecov.io/gh/huggingface/transformers/pull/3866?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[ { "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 } ]

[ "Great, thanks @sshleifer ", "(Duplicate of https://github.com/huggingface/transformers/issues/3853)\r\n", "sorry, mentioned wrong issue" ]

[ { "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/3864?src=pr&el=h1) Report\n> Merging [#3864](https://codecov.io/gh/huggingface/transformers/pull/3864?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/a21d4fa410dc3b4c62f93aa0e6bbe4b75a101ee9&el=desc) will **increase** coverage by `0.00%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/3864?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #3864 +/- ##\n=======================================\n Coverage 78.61% 78.62% \n=======================================\n Files 106 106 \n Lines 17953 17953 \n=======================================\n+ Hits 14114 14115 +1 \n+ Misses 3839 3838 -1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/3864?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/3864/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `92.76% <0.00%> (+0.16%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/3864?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/3864?src=pr&el=footer). Last update [a21d4fa...7bbf47b](https://codecov.io/gh/huggingface/transformers/pull/3864?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Hi @julien-c !\r\n\r\nI hope you are well. My pull request is ready for review.\r\n\r\nI have tried my best to add license and language information to all model cards. I have added a few model cards as well.\r\n\r\nNote that my changes may have some downstream consequences:\r\n- the addition of a \"license\" key, the value being an atomic list)\r\n- the normalization of the \"languages\" key (I added the \"s\"), the value being a list of [ISO 639-1 codes](https://en.wikipedia.org/wiki/List_of_ISO_639-1_codes). For multilingual BERT, I had to simplify some rare languages by merging it with its ISO \"macro code\" (example: \"South Azerbaijani\" -> \"az\", \"Bavarian\" -> \"de\")\r\n \r\nOn the website, you may want to read those values and render them back in human-readable form.\r\n\r\nCheers,\r\n\r\nAlex\r\n", "Hi Alex,\r\n\r\nas mentioned in the previous issue, I'd rather use identifiers listed in https://help.github.com/en/github/creating-cloning-and-archiving-repositories/licensing-a-repository\r\n\r\nI can probably search and replace though.\r\n\r\nAlso, you pushed whitespace changes which make reviewing the actual changes slightly tedious.", "[EDIT] Sure, I have replaced licenses with these identifiers! \r\nFor whitespaces, I have autoformatting activated on sublime, that's why. Sorry for the inconvenience. ", "Good morning @julien-c,\r\n\r\nI hope all is well. What do you think of this PR?\r\n\r\nCheers,\r\n\r\nAlex", "I'm doing a partial merge (retaining your authorship information, @alexcombessie) of the licenses, as the languages will require some backend changes.\r\n\r\n(I'll do a search and replace at a later point)\r\n\r\nThank you for your contribution" ]

[]

[]

[ "I believe this is because `tf.Cumsum` is not a supported operation and not an issue relating to this repo. Here is a link to the tensorflow documentation on supported ops. [https://www.tensorflow.org/lite/guide/ops_compatibility]\r\nIn the past, I've been able to get around unsupported ops by reimplementing the operator with supported ops or replacing the unsupported portion with another op. ie. `relu` in place of `gelu`.", "Hey @will-rice, thank you for giving me an idea how to handle this issue. I managed to overcome this problem, by using custom _cumsum_ function implemented in pure python by @ibab in here https://github.com/tensorflow/tensorflow/issues/813.\r\nI just changed it to sum over rows not columns, as the way it is done in the Roberta model.\r\n\r\nHere is a cumsum function:\r\n```python\r\ndef cumsum(xs):\r\n values = tf.unstack(xs, axis=1)\r\n out = []\r\n prev = tf.zeros_like(values[0])\r\n for val in values:\r\n s = prev + val\r\n out.append(s)\r\n prev = s\r\n result = tf.stack(out, axis=1)\r\n return result\r\n```\r\nand it is used in the _modeling_tf_roberta.py_ file in line 69:\r\n```python\r\n # Original code / non tflite compatible way\r\n incremental_indicies = tf.math.cumsum(mask, axis=1) * mask)\r\n\r\n # My custom code / tflite compatible way\r\n incremental_indicies = cumsum(mask) * mask\r\n```\r\n\r\nHope it will help anyone aswell!", "Also cc'ing @Pierrci ", "@julien-c any updates on this feature? Was browsing through the later releases but could not find any reference.\r\n\r\nThanks!", "@dshahrokhian As mentioned by @will-rice, the issue is due to the lack of support for the `tf.Cumsum` operator by TFLite and thus not related to `transformers`. If you encounter the same problem you can implement the workaround posted by @kubux1 earlier, or implement a similar one if you're having this issue with a different operator.", "@Pierrci thanks! It also seems to be have been solved in the latest release of `tf-nightly`: https://github.com/tensorflow/tensorflow/issues/42382#issuecomment-675000451" ]

[ { "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" } ]

[]

[ "Thanks! [model page](https://huggingface.co/Hate-speech-CNERG/dehatebert-mono-english)", "A few things to add to the model card if you can (happy to help!)\r\n\r\n- which language(s) is it trained on?\r\n- How can one use it, i.e. is this a sequence classifier? " ]

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

[]

[ "You could do it very simply by passing the reference around:\r\n\r\n```py\r\nfrom transformers import BertModel\r\n\r\nmodel = BertModel.from_pretrained(\"bert-base-cased\")\r\nmodel2 = BertModel.from_pretrained(\"bert-base-cased\")\r\n\r\nmodel2.embeddings = model.embeddings\r\n\r\nprint(model2.embeddings.word_embeddings.weight)\r\n\r\nmodel.embeddings.word_embeddings.weight = torch.nn.Parameter(torch.zeros_like(model.embeddings.word_embeddings.weight))\r\n\r\nprint(model2.embeddings.word_embeddings.weight)\r\n```\r\n\r\nwhich outputs the result (note that I'm updating the `model.embeddings` and printing the `model2.embeddings`):\r\n\r\n```py\r\nParameter containing:\r\ntensor([[-0.0005, -0.0416, 0.0131, ..., -0.0039, -0.0335, 0.0150],\r\n [ 0.0169, -0.0311, 0.0042, ..., -0.0147, -0.0356, -0.0036],\r\n [-0.0006, -0.0267, 0.0080, ..., -0.0100, -0.0331, -0.0165],\r\n ...,\r\n [-0.0064, 0.0166, -0.0204, ..., -0.0418, -0.0492, 0.0042],\r\n [-0.0048, -0.0027, -0.0290, ..., -0.0512, 0.0045, -0.0118],\r\n [ 0.0313, -0.0297, -0.0230, ..., -0.0145, -0.0525, 0.0284]],\r\n requires_grad=True)\r\nParameter containing:\r\ntensor([[0., 0., 0., ..., 0., 0., 0.],\r\n [0., 0., 0., ..., 0., 0., 0.],\r\n [0., 0., 0., ..., 0., 0., 0.],\r\n ...,\r\n [0., 0., 0., ..., 0., 0., 0.],\r\n [0., 0., 0., ..., 0., 0., 0.],\r\n [0., 0., 0., ..., 0., 0., 0.]], requires_grad=True)\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" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/3860?src=pr&el=h1) Report\n> Merging [#3860](https://codecov.io/gh/huggingface/transformers/pull/3860?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/a21d4fa410dc3b4c62f93aa0e6bbe4b75a101ee9&el=desc) will **not change** coverage.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/3860?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #3860 +/- ##\n=======================================\n Coverage 78.61% 78.61% \n=======================================\n Files 106 106 \n Lines 17953 17953 \n=======================================\n Hits 14114 14114 \n Misses 3839 3839 \n```\n\n\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/3860?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/3860?src=pr&el=footer). Last update [a21d4fa...9e4fe33](https://codecov.io/gh/huggingface/transformers/pull/3860?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

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

[ "\r\n", "tokenizer.pad_token = 0", "You have to set the pad_token_id yourself as it's stated in the error message ;-). I would recommend using the `eos_token_id` as the `pad_token_id` for GPT2:\r\n```python \r\ntokenizer = GPT2Tokenizer.from_pretrained('gpt2')\r\ntokenizer.pad_token = tokenizer.eos_token\r\n```\r\nas it's written in the error message ;-)", "I hit the same issue after using [add_special_tokens](https://huggingface.co/transformers/main_classes/tokenizer.html#transformers.PreTrainedTokenizer.add_special_tokens) with `{\"pad_token\": \"PAD\"}` dictionary.\r\n\r\nI understand based on the error and the documentation it should not raise the error, right? @patrickvonplaten should the issue be reopened?", "For completeness: \r\n- patrickvonplaten tip `tokenizer.pad_token = tokenizer.eos_token` solved it\r\n- the whole error message for me was\r\n\r\n```\r\nValueError: Unable to set proper padding strategy as the tokenizer does not have a padding token. In this case please set the `pad_token` `(tokenizer.pad_token = tokenizer.eos_token e.g.)` \r\nor add a new pad token via the function add_special_tokens if you want to use a padding strategy\r\n```", "You didn't set pad_token. You can set it like this:\r\n```python\r\ntokenizer.pad_token = \"[PAD]\"\r\n```" ]

[]

[]

[ "Hi! We run on K80 and V100 GPUs. Running the 1.5B params model was very costly and hard to maintain, so we're not running it anymore, however. It should run fine on a single V100/Titan RTX however.\r\n\r\nDid you have issues with the T4 GPU because of memory?", "Thanks for you response. This is really beneficial. I think that it is due to memory. \r\nAre you serving your models with a flask server, TF-serving, or a different serving framework?\r\nWere you serving it using your PyTorch or Tensorflow implementation?\r\n\r\nThanks again", "We're serving our models in PyTorch, using a mix of gunicorn/falcon to handle requests. You can see the detailers [here](https://medium.com/huggingface/scaling-a-massive-state-of-the-art-deep-learning-model-in-production-8277c5652d5f)!", "Really clear blog post.\r\nThanks" ]

[]

[]

[ "Update: rm'ed inaccurate comment" ]

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

[]

[ "I have raised https://github.com/huggingface/transformers/pull/4940, waiting for approval.", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

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

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/3855?src=pr&el=h1) Report\n> Merging [#3855](https://codecov.io/gh/huggingface/transformers/pull/3855?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/a21d4fa410dc3b4c62f93aa0e6bbe4b75a101ee9&el=desc) will **increase** coverage by `0.00%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/3855?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #3855 +/- ##\n=======================================\n Coverage 78.61% 78.62% \n=======================================\n Files 106 106 \n Lines 17953 17953 \n=======================================\n+ Hits 14114 14115 +1 \n+ Misses 3839 3838 -1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/3855?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_bert.py](https://codecov.io/gh/huggingface/transformers/pull/3855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19iZXJ0LnB5) | `88.40% <100.00%> (ø)` | |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/3855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `92.76% <0.00%> (+0.16%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/3855?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/3855?src=pr&el=footer). Last update [a21d4fa...06fb495](https://codecov.io/gh/huggingface/transformers/pull/3855?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Good catch @Bharat123rox - thanks for the PR :-) " ]

[ { "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/3854?src=pr&el=h1) Report\n> Merging [#3854](https://codecov.io/gh/huggingface/transformers/pull/3854?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/a21d4fa410dc3b4c62f93aa0e6bbe4b75a101ee9&el=desc) will **increase** coverage by `0.00%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/3854?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #3854 +/- ##\n=======================================\n Coverage 78.61% 78.62% \n=======================================\n Files 106 106 \n Lines 17953 17953 \n=======================================\n+ Hits 14114 14115 +1 \n+ Misses 3839 3838 -1 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/3854?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/3854/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `92.76% <0.00%> (+0.16%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/3854?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/3854?src=pr&el=footer). Last update [a21d4fa...b5f2dc5](https://codecov.io/gh/huggingface/transformers/pull/3854?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Cherry-picked in 7f23af16840113fe137f42415a9daa7ce7f7f15f\r\n\r\nThank you, that looks great! cc @LysandreJik and @clarkkev " ]

[ { "id": 1260952223, "node_id": "MDU6TGFiZWwxMjYwOTUyMjIz", "url": "https://api.github.com/repos/huggingface/transformers/labels/Discussion", "name": "Discussion", "color": "22870e", "default": false, "description": "Discussion on a topic (keep it focused or open a new issue though)" }, { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "Facing a similar type of issue for T5. @sshleifer ", "The last ckpt file should be loaded into a `pl.LightningModule` if the --do_predict flag is specified.\r\n\r\nThere is a bug on master that messes up the loading, but it's fixed in #3866\r\n\r\nTo use that code immediately, you can run:\r\n```\r\ngit fetch\r\ngit checkout examples-summ-do-predict\r\n```\r\nthen your same `finetune.py` command\r\n with `--do_predict` (and not --do_train) and the proper `--output_dir`.\r\n\r\nWould love to know if that works!\r\n\r\ncc: @ethanjperez.", "Change is on master, let me know if this solves the problem!", "Config.json is still not generated while training.", "```python\r\n def log_hyperparams(model: pl.LightningModule):\r\n model.config.save_pretrained(model.hparams.output_dir)\r\n with open(os.path.join(model.hparams.output_dir, \"hparam.json\")) as f:\r\n json.dump(model.hparams, f)\r\n```\r\nYou can call this somewhere in your code, if that's helpful.", "@sshleifer, thank you - I can run ./run_train.sh with the --predict() option successfully. \r\n\r\nRegarding my original question, could you please specify how to load the checkpoint into the LighteningModule?\r\n\r\nAfter inspecting [transformer_base.py](https://github.com/huggingface/transformers/blob/master/examples/transformer_base.py), I think hparams is equivalent to the arguments provided in run_train.sh, so a separate hparams.json file does not need to be generated. Please correct me if I'm wrong.\r\n\r\nI am receiving the following error with my current code:\r\n\r\n`pytorch_lightning.utilities.exceptions.MisconfigurationException: Checkpoint contains hyperparameters but LightningModule's __init__ is missing the argument 'hparams'. Are you loading the correct checkpoint?`\r\n\r\nI've been using the following code, based on the discussion in https://github.com/PyTorchLightning/pytorch-lightning/issues/525 and https://pytorch-lightning.readthedocs.io/en/latest/weights_loading.html:\r\n```\r\n\r\n# load model\r\nimport pytorch_lightning as pl\r\n\r\nfrom argparse import Namespace\r\n\r\n# usually these come from command line args\r\nargs = Namespace(data_dir='CE_data/',\r\nmodel_type='bart',\r\nmodel_name_or_path='bart-large',\r\nlearning_rate='3e-5',\r\ntrain_batch_size=4,\r\neval_batch_size=4,\r\noutput_dir='transformers/examples/summarization/bart/bart_sum',\r\ndo_predict='do_predict')\r\n\r\npretrained_model = pl.LightningModule.load_from_checkpoint('bart_sum/checkpointepoch=2.ckpt', hparams=args)\r\npretrained_model.eval()\r\n\r\n# or for prediction\r\nout = model(inputs['input_ids'])\r\nprint(out)\r\n``'\r\n\r\nThank you for your time.", "Seems close to correct.\r\n\r\nhttps://github.com/huggingface/transformers/blob/7d40901ce3ad9e1c79fd9bb117f5b84bff42c33f/examples/summarization/bart/finetune.py#L164-L175\r\n\r\nis how we do it @riacheruvu", "@sshleifer \r\n1. Originally config.json is not created which is a requirement for prediction using fine-tuned model.\r\n*As shown in the screenshot, I add this code in transformer_base.py in end, config and hparam files are created.\r\n* Then try to predict with --do_predict, then it gives, \"\"We assumed '/content/t5' was a path, a model identifier, or url to a directory containing vocabulary files named ['spiece.model'] but couldn't find such vocabulary files at this path or url.\"\"\r\nWhat are the requirements to use fine-tuned model?\r\n<img width=\"696\" alt=\"Screenshot 2020-04-21 at 5 50 10 PM\" src=\"https://user-images.githubusercontent.com/30004110/79886728-c1d0bf80-83f9-11ea-90e5-400afc575da1.png\">\r\n\r\n----------------------------------------------------------------\r\n2. To predict for a single instance using the fine-tuned model, do I need to specify the test.target file also. I want to predict unknown instance without calculating the loss value.\r\n", "@sshleifer, thank you. I've got to the point where I can load the model and generate \"outputs\" using the forward() function, but I can't decode the outputs - using tokenizer.decoder() results in an error. Should I be using model.generate() instead of model.forward()? If so, it seems SummarizationTrainer does not support model.generate?\r\n\r\n\r\nRevised code:\r\n\r\n```\r\n tokenizer = BartTokenizer.from_pretrained('bart-large-cnn')\r\n ARTICLE_TO_SUMMARIZE = \"My friends are cool but they eat too many carbs.\"\r\n inputs = tokenizer.batch_encode_plus([ARTICLE_TO_SUMMARIZE], max_length=1024, return_tensors='pt')['input_ids']\r\n checkpoints = list(sorted(glob.glob(os.path.join(args.output_dir, \"checkpointepoch=*.ckpt\"), recursive=True)))\r\n model = model.load_from_checkpoint(checkpoints[-1])\r\n model.eval()\r\n model.freeze()\r\n outputs = model(inputs)\r\n print(outputs) #Successfully prints two 3D tensors in a tuple\r\n #print([tokenizer.decode(g, skip_special_tokens=True, clean_up_tokenization_spaces=False) for g in outputs]) #Results in ValueError: only one element tensors can be converted to Python scalars\r\n print([tokenizer.decode(g, skip_special_tokens=True, clean_up_tokenization_spaces=False) for g in outputs[0][0]])\r\n```\r\nThe error I'm encountering\r\n```\r\nTraceback (most recent call last):\r\n File \"finetune.py\", line 194, in <module>\r\n main(args)\r\n File \"finetune.py\", line 184, in main\r\n print([tokenizer.decode(g, skip_special_tokens=True, clean_up_tokenization_spaces=False) for g in outputs[1][0]])\r\n File \"finetune.py\", line 184, in <listcomp>\r\n print([tokenizer.decode(g, skip_special_tokens=True, clean_up_tokenization_spaces=False) for g in outputs[1][0]])\r\n File \"/home/ubuntu/anaconda3/envs/pytorch_p36/lib/python3.6/site-packages/transformers/tokenization_utils.py\", line 2141, in decode\r\n sub_texts.append(self.convert_tokens_to_string(current_sub_text))\r\n File \"/home/ubuntu/anaconda3/envs/pytorch_p36/lib/python3.6/site-packages/transformers/tokenization_gpt2.py\", line 235, in convert_tokens_to_string\r\n text = \"\".join(tokens)\r\nTypeError: sequence item 0: expected str instance, NoneType found\r\n```", "I found a solution. The model.generate() function is necessary to extract the predictions. I defined a separate function in the SummarizationTrainer() class to use self.model.generate(), and was able to use tokenizer.decoder() on the outputs.\r\n\r\nI was encountering issues when using self.tokenizer, so I assume using 'bart-large-cnn' tokenizer for similar custom summarization datasets is okay.\r\n\r\n@prabalbansal, I'm not sure if the same method will apply to T5, but it could work for predicting for a single instance, per one of your questions.\r\n\r\nMy code is below: \r\n\r\n```\r\n def text_predictions(self, input_ids):\r\n generated_ids = self.model.generate(\r\n input_ids=input_ids,\r\n num_beams=1,\r\n max_length=80,\r\n repetition_penalty=2.5,\r\n length_penalty=1.0,\r\n early_stopping=True,\r\n )\r\n preds = [\r\n self.tokenizer.decode(g, skip_special_tokens=True, clean_up_tokenization_spaces=True)\r\n for g in generated_ids\r\n ]\r\n return preds\r\n...\r\n # Optionally, predict on dev set and write to output_dir\r\n if args.do_predict:\r\n # See https://github.com/huggingface/transformers/issues/3159\r\n # pl use this format to create a checkpoint:\r\n # https://github.com/PyTorchLightning/pytorch-lightning/blob/master\\\r\n # /pytorch_lightning/callbacks/model_checkpoint.py#L169\r\n tokenizer = BartTokenizer.from_pretrained('bart-large-cnn')\r\n ARTICLE_TO_SUMMARIZE = \"My friends are cool but they eat too many carbs.\"\r\n inputs = tokenizer.batch_encode_plus([ARTICLE_TO_SUMMARIZE], max_length=1024, return_tensors='pt')['input_ids']\r\n checkpoints = list(sorted(glob.glob(os.path.join(args.output_dir, \"checkpointepoch=*.ckpt\"), recursive=True)))\r\n model = model.load_from_checkpoint(checkpoints[-1])\r\n model.eval()\r\n model.freeze()\r\n outputs = model.text_predictions(inputs)\r\n print(outputs)\r\n```\r\n\r\nThank you for the help, @sshleifer !", "@riacheruvu Thank You. It works for T5 also.", "I followed the steps given in this thread and am still facing an issue. I get an error saying the below when I try to use my fine-tuned model for prediction.\r\n\r\nOSError: Can't load '/home/bart/bart_1/checkpointepoch=3.ckpt'. Make sure that:\r\n\r\n- '/home/bart/bart_1/checkpointepoch=3.ckpt' is a correct model identifier listed on 'https://huggingface.co/models'\r\n\r\n- or '/home/bart/bart_1/checkpointepoch=3.ckpt' is the correct path to a directory containing a 'config.json' file\r\n", "@sangeethabal15, with my model, files were only generated up till the 2nd epoch. Just to confirm, do you have a checkpointepoch=3.ckpt file?\n\nAre you using the load_from_checkpoint() function?\n", "@riacheruvu yes I do have checkpoint=3.ckpt file. I gave my own number of epochs instead of the default 3. \r\n\r\nYes I am using the load_from_checkpoint() function", "Ok. Could you share your code here, @sangeethabal15? It might be easier to help debug. ", "@riacheruvu This is my modified code -\r\n\r\n\r\n\r\n # Optionally, predict on dev set and write to output_dir\r\n if args.do_predict:\r\n # See https://github.com/huggingface/transformers/issues/3159\r\n # pl use this format to create a checkpoint:\r\n # https://github.com/PyTorchLightning/pytorch-lightning/blob/master\\\r\n # /pytorch_lightning/callbacks/model_checkpoint.py#L169\r\n examples = [\" \" + x.rstrip() for x in open(\"/home/bart/input/test.source\").readlines()]\r\n fout = Path(\"output.txt\").open(\"w\")\r\n checkpoints = list(sorted(glob.glob(os.path.join(args.output_dir, \"checkpointepoch=*.ckpt\"), recursive=True)))\r\n model = model.load_from_checkpoint(checkpoints[-1])\r\n tokenizer = BartTokenizer.from_pretrained(\"bart-large\")\r\n\r\n max_length = 80\r\n min_length = 5\r\n\r\n for batch in tqdm(list(chunks(examples, 8))):\r\n dct = tokenizer.batch_encode_plus(batch, max_length=1024, return_tensors=\"pt\", pad_to_max_length=True)\r\n summaries = model.generate(\r\n input_ids=dct[\"input_ids\"].to(device),\r\n attention_mask=dct[\"attention_mask\"],\r\n num_beams=4,\r\n length_penalty=2.0,\r\n max_length=max_length + 2, # +2 from original because we start at step=1 and stop before max_length\r\n min_length=min_length + 1, # +1 from original because we start at step=1\r\n no_repeat_ngram_size=3,\r\n early_stopping=True,\r\n decoder_start_token_id=model.config.eos_token_id,\r\n )\r\n dec = [tokenizer.decode(g, skip_special_tokens=True, clean_up_tokenization_spaces=False) for g in summaries]\r\n for hypothesis in dec:\r\n fout.write(hypothesis + \"\\n\")\r\n fout.flush()", "Thank you, @sangeethabal15. From the error message you posted earlier, it seems load_from_checkpoint() is expecting a config.json file in the specified directory. \n\nI have a few more debug questions:\n\n- Do you have the latest version of the code?\n\n- Does load_from_checkpoint() work with the checkpoint file for the 2nd epoch? \n\n- If that fails, does your code run successfully if you use the default number of epochs?", "@riacheruvu \r\n\r\n- I do have the latest version of the code though I have not trained the model on the latest version of it.\r\n\r\n- load_from_checkpoint doesn't work with the 2nd either and expects a config.json file\r\n\r\n- and yes the code runs successfully on the default number of epochs as well.", " import json\r\n def log_hyperparams(model: pl.LightningModule):\r\n model.config.save_pretrained(model.hparams.output_dir)\r\n with open(os.path.join(model.hparams.output_dir, \"hparam.json\"),'w') as f:\r\n json.dump(model.hparams.__dict__, f)\r\n if args.do_train:\r\n trainer.fit(model)\r\n log_hyperparams(model)\r\n\r\n@sangeethabal15 Could you add this at the end of transformer_base.py. This works for me.\r\n", "@prabalbansal this is for when I am training my model. Since I have already fine-tuned my model, is there any workaround for test time when I am trying to predict my outputs?", "@riacheruvu I am currently working on a Text Summarization problem. I have collected a small dataset of my own. Implementing BART is very easy. I can generate a great summary. But I want to know how to how to use BART model for training my own custom dataset. Can you please kindly help me with this?\r\n\r\nI have browsed through internet. But I cannot any find any helpful resources as it is relatively new compared to other Transfer learning models.", "@murugeshmanthiramoorthi you can just use run_train.sh in the bart folder where you give in your parameters to run the fiinetune.py file", "@sangeethabal15 Thank you so much for your reply mam. I am completely new to transfer learning mam. I can't get what you are upto. Can you kindly explain more elaborately or share a resource so that I can follow up?\r\nThanks in advance mam. \r\n", "@sangeethabal15 I somehow managed to load the dataset. I run the run_train.sh file. But it is showing me error \"python3: can't open file 'finetune.py': [Errno 2] No such file or directory\". I even tried changing the data set from my custom dataset to default CNN/daily news dataset. Still, I am getting the same error. Can anyone help me out?", "@riacheruvu @prabalbansal did y'all finetune Bart on your own dataset?", "@sangeethabal15, I fine-tuned BART on my own custom dataset. It's strange that your code runs successfully on the default number of epochs, but load_from_checkpoint() does not work with the 2nd epoch .ckpt file with the original configuration. Where did you modify the default number of epochs?\r\n\r\n@murugeshmanthiramoorthi, \r\n\r\nPer the instructions given in https://github.com/huggingface/transformers/tree/master/examples/summarization/bart:\r\n\r\nThe steps I followed are cloning the transformers repo, navigating to the examples/summarization/bart directory, copying over a folder containing the data files (train.target, train.source, val.target, val.source, test.target, and test.source files), and then modifying run_train.sh to use this folder for the data_dir and filling in the other parameters.\r\n\r\nFor your .source and .target files, you need to structure them similar to the CNN/DM dataset: The .source files should have an article on each line, and the .target files should have a target summary on each line (corresponding to the article in the .source file).", "@riacheruvu I noticed that I get this warning for both training and while testing\r\n\r\nINFO:transformers.modeling_utils:Weights from pretrained model not used in BartForConditionalGeneration: ['encoder.version', 'decoder.version']\r\n\r\nSeems like my model hasn't been trained properly. Any idea how to go about this?\r\n\r\nAlso, I have the number of epochs in my run_train.sh. It is defined in the add_specific_args in the transformer_base.py", "that warning doesn't matter.", "@sangeethabal15, I agree that the warning does not matter as I saw that warning as well. It seems the issue might be when training the model with a different number of epochs compared to the default. @sshleifer, has the HuggingFace team tested the code with a different number of epochs before?", "@riacheruvu Thank you so much for your help. But when I proceeded with those steps, I get the error \r\n\r\nTraceback (most recent call last):\r\n File \"finetune.py\", line 10, in <module>\r\n from transformer_base import BaseTransformer, add_generic_args, generic_train, get_linear_schedule_with_warmup\r\nModuleNotFoundError: No module named 'transformer_base'\r\n\r\nDo you have any idea solving this issue." ]

[]

[ { "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 } ]

[ "From what I'm seeing, the `TFT5Model` **does** have [documentation](https://huggingface.co/transformers/model_doc/t5.html#transformers.TFT5Model.get_input_embeddings) for `get_input_embeddings` and `get_output_embeddings`.\r\n\r\nI believe the output embeddings and input embeddings should actually be the same. The embeddings are shared between input and output. Wdyt @patrickvonplaten? ", "Agree that the documentation is not the greatest, could definitely be improved :-). \r\n\r\nThe idea is that both `get_input_embeddings()` and `get_output_embeddings` return the **same** (this should be made clearer in the docs) embeddings matrix of dimension Vocab_size x Hidden_size. \r\n\r\nNow, to make the embeddings matrix work for both input and output, we need to be able to get a Vocab_size -> Hidden_size mapping (for the input embeddings) and Hidden_size -> Vocab_size mapping (for the output embeddings). In TF, we use a trick here, by wrapping the embedding in this layer: https://github.com/huggingface/transformers/blob/c53cc018de70436196858ca91c1a34f1b8947028/src/transformers/modeling_tf_utils.py#L1521\r\n\r\nAnd then by calling the embedding with different modes (\"linear\" and \"embedding\"), we get the correct mapping. See https://github.com/huggingface/transformers/blob/c53cc018de70436196858ca91c1a34f1b8947028/src/transformers/modeling_tf_t5.py#L1074 for example. \r\n\r\nSo IMO, the code is fine, I agree with you @parthe, that the documention should be cleaner and explain the logic I explained here a bit. \r\n\r\nIf you feel like it @parthe, it would be amazing if you could open a PR to straighten up the documentation here (the docstring). ", "Hi - I think I have related questions since I couldn't find answers in the documentation for get_input_embeddings().... \r\n\r\nI've been using the approach [here](https://mccormickml.com/2019/05/14/BERT-word-embeddings-tutorial/) to access the hidden states to obtain embeddings (I've also updated it to be based on `transformers` in my own notebook instead of `pytorch-pretrained-bert`) . I was wondering how the output of get_input_embeddings maps to the output of the hidden states there? I've not been able to figure that out. Also, what would be the advantage of using one over the other?\r\n\r\nThanks!", "So I would recommend that you take the `hidden_states` by setting `config.output_hidden_states=True` (Note: this API will be changed soon, see PR: #4538). \r\nThen you can map the `hidden_states` to `lm_logits` (non normalized scores for each word in the vocab) using:\r\n\r\n```python \r\nembed_tokens = self.get_output_embeddings()\r\nlm_logits = embed_tokens(<your_hidden_states>, mode=\"linear\")\r\n```\r\n\r\nLet me know if this isn't clear :-) ", "Hi @patrickvonplaten, referring to the quote below (from this [comment](https://github.com/huggingface/transformers/issues/3852#issuecomment-618852195)):\r\n> The idea is that both `get_input_embeddings()` and `get_output_embeddings` return the **same** (this should be made clearer in the docs) embeddings matrix of dimension Vocab_size x Hidden_size.\r\n> \r\n> Now, to make the embeddings matrix work for both input and output, we need to be able to get a Vocab_size -> Hidden_size mapping (for the input embeddings) and Hidden_size -> Vocab_size mapping (for the output embeddings). In TF, we use a trick here, by wrapping the embedding in this layer:\r\n> \r\n> https://github.com/huggingface/transformers/blob/c53cc018de70436196858ca91c1a34f1b8947028/src/transformers/modeling_tf_utils.py#L1521\r\n> \r\n> And then by calling the embedding with different modes (\"linear\" and \"embedding\"), we get the correct mapping. See\r\n> \r\n> https://github.com/huggingface/transformers/blob/c53cc018de70436196858ca91c1a34f1b8947028/src/transformers/modeling_tf_t5.py#L1074\r\n> \r\n> for example.\r\n\r\nDoes this only apply to TFT5Model or is the same across all models which has `get_input_embeddings` and `get_output_embeddings` method?", "It should be the same across models that share input and output embeddings :-) " ]

[]

[]

[ "This seems like more of a TF-related question rather than a Transformers-related question. The issue seems to stem from your code trying to get the value of a tensor which is not eager, using numpy. I believe the `tf.data.Dataset.map` method must trace inputs, resulting in the Tensors not being eager.\r\n\r\nCouldn't you build the `tf.data.Dataset` with already tokenized inputs instead?", "The ideal would be to follow the pipeline (read from the file >> generate batches >> tokenize >> train >> evaluate). It is the most efficient approach as pointed in the [TensorFlow tutorial](https://www.tensorflow.org/tutorials/customization/performance).\r\n\r\nTensorflow when dealing with texts generates string tensors that are stored as byte string:\r\n\r\n```python\r\n<tf.Tensor: shape=(2,), dtype=string, numpy=array(\r\n [b'Thê first utf-8 string of the batçh.',\r\n b'Thê secônd utf-8 string of the batçh.'], dtype=object)>\r\n```\r\n\r\nHowever, I didn't find an efficient way to decode this kind of tensor as a list of strings. It's even worse if the byte string containing a non-ascii character.\r\n\r\nWhat I really need is one of these two options:\r\n\r\n1. a tokenizer which is able to accept aforementioned byte string tensor as input to tokenize; or\r\n2. a vectorized approach to transforming a byte string tensor into a string list.\r\n\r\nThank you very much for all your help.", "@Ceceu I am running into this exact issue as well, and am wondering if you had found a good solution?", "@oja,\r\nThe best solution I could find was adapting an example from the Tensorflow tutorial: [Load Text](https://www.tensorflow.org/tutorials/load_data/text#encode_text_lines_as_numbers) which uses `tf.py_function`.\r\nLet me know if I can help more.", "@Ceceu got it, thank you!", "Tokenizers can now output `numpy` arrays with `return_tensors='np'` so I think this should work now.", "Thanks @thomwolf, I will check it out and if it works on TPU then it solves https://github.com/huggingface/transformers/issues/5066", "> Thanks @thomwolf, I will check it out and if it works on TPU then it solves #5066\r\n\r\nDid you check if it works on TPU?", "It does not work on TPU", "@oja, @Santosh-Gupta, @celsofranssa I too am facing this problem. Did you guys find any solution?", "cc @Rocketknight1 ", "Bump, I'm still having this issue (on a CPU)." ]

[]

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

[ "Hi @Akashdesarda, thanks for reporting this.\r\nYou should be able to pass `pad_to_max_length=True` when calling your pipeline: \r\n\r\n`pipe(train_data['comment_text'][:100].values.tolist(), pad_to_max_length=True)`\r\n\r\nCan you let us know if it works in your case ?\r\n", "Yes it worked, thanks for the solution." ]

[]

[]

[ "You need to install from source as specified in the README." ]

[]

[]

[ "You can basically copy+paste the code from BertForQuestionAnswering and just change it for ELECTRA. However, the original ELECTRA implementation to fine-tune on squad looks a bit different (it's more like in XLNet). \r\n\r\nIf you want to reproduce the official implementation it's probably best you take a look at the published code: https://github.com/google-research/electra/blob/master/finetune/qa/qa_tasks.py#L419", "Any updates about this? I managed the creation of ElectraForQuestionAnswering on my own and the code works. If the progress of this thread is in stand-by I can proceed submitting my pull request", "@volker42maru I implemented the same model as described in the official Electra Repository by google. I am still unable to reproduce the original paper results which is 75% EM on the squad v1 and 82% F1. The maximum I could get was 70% EM and 78% F1. @mmuffo94 Please let me know if you have successfully reproduced the results on the squad v1 dataset. It would be great help. I am using the Electra Small Model for now.", "@ankur19030 you are refering to ELECTRA-Small I assume? \r\n\r\nI actually finetuned and evaluated only on squad 2.0, but the score on squad 1.1 even with the Small model should be significantly higher. The different QA Head that is used for ELECTRA might not play such a big role in squad 1.1, because it's mostly used to get better predictions on answerability. You can try using a simple QA head first to check the performance on squad 1.1 with ELECTRA, e.g.:\r\n```\r\nclass ElectraForQuestionAnswering(ElectraPreTrainedModel):\r\n def __init__(self, config):\r\n super(ElectraForQuestionAnswering, self).__init__(config)\r\n self.num_labels = config.num_labels\r\n\r\n self.electra = ElectraModel(config)\r\n self.qa_outputs = nn.Linear(config.hidden_size, config.num_labels)\r\n\r\n self.init_weights()\r\n\r\n def forward(\r\n self,\r\n input_ids=None,\r\n attention_mask=None,\r\n token_type_ids=None,\r\n position_ids=None,\r\n head_mask=None,\r\n inputs_embeds=None,\r\n start_positions=None,\r\n end_positions=None,\r\n ):\r\n outputs = self.electra(\r\n input_ids, attention_mask, token_type_ids, position_ids, head_mask, inputs_embeds\r\n )\r\n\r\n sequence_output = outputs[0]\r\n\r\n logits = self.qa_outputs(sequence_output)\r\n start_logits, end_logits = logits.split(1, dim=-1)\r\n start_logits = start_logits.squeeze(-1)\r\n end_logits = end_logits.squeeze(-1)\r\n\r\n outputs = (start_logits, end_logits,) + outputs[2:]\r\n if start_positions is not None and end_positions is not None:\r\n # If we are on multi-GPU, split add a dimension\r\n if len(start_positions.size()) > 1:\r\n start_positions = start_positions.squeeze(-1)\r\n if len(end_positions.size()) > 1:\r\n end_positions = end_positions.squeeze(-1)\r\n # sometimes the start/end positions are outside our model inputs, we ignore these terms\r\n ignored_index = start_logits.size(1)\r\n start_positions.clamp_(0, ignored_index)\r\n end_positions.clamp_(0, ignored_index)\r\n\r\n loss_fct = torch.nn.CrossEntropyLoss(ignore_index=ignored_index)\r\n start_loss = loss_fct(start_logits, start_positions)\r\n end_loss = loss_fct(end_logits, end_positions)\r\n total_loss = (start_loss + end_loss) / 2\r\n outputs = (total_loss,) + outputs\r\n\r\n return outputs # (loss), start_logits, end_logits, (hidden_states), (attentions)\r\n```\r\n\r\nHowever, if you try to use ELECTRA-Base or Large you also want to use `layerwise_lr_decay`, as used in the official implementation (https://github.com/google-research/electra/blob/master/model/optimization.py#L48). For me that made quite a big difference in score.\r\n\r\nBTW, be sure to use `google/electra-small-discriminator` weights, not the generator.", "@volker42maru I initially tried with the simple QA head only, the same as you described but could not reproduce the results, not even near them even though I am using the same hyper-parameters as the ones used in the official implementation except layer wise LR decay. My EM is <70 on squad v1 which is around 5% less than the results from official repo. But I can say for sure that simply Layerwise LR decay in small model should not cause this much difference, as I have already tried removing Layerwise LR decay from official implementation and it still give the same results. So I have no idea where is the gap ?", "Are you sure that's for squad v1 and not v2?\r\n\r\nI just trained ELECTRA Small for 1 epoch using the simple QA head from above and I get the following score on squad v1 dev: `'exact': 73.46263008514664, 'f1': 82.46777637449017`\r\n\r\nI used mostly default parameters and no `layerwise_lr_decay`:\r\n```\r\n--num_train_epochs 1\r\n--per_gpu_train_batch_size 24\r\n--learning_rate 5e-5\r\n```", "@volker42maru I will cross check again then, Thanks ", "It's been added https://github.com/huggingface/transformers/pull/4913" ]

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

[]

[ "@xihui-wu To get the hyperparameters specific to the model (in this case `gpt2`), you can check the config file of `gpt2` with the code below:\r\n\r\n```\r\nfrom transformers import GPT2Config\r\nprint(GPT2Config())\r\n```\r\n\r\nSome higher level hyperparameters are still not included here (e.g. \"epochs\"). These can be set explicitly as arguments when running the CLI `run_language_modeling.py`; otherwise, the default values are used.\r\n\r\nYou can find the hyperparameters and their default values at the beginning of the `main` function in [run_language_modeling.py](https://github.com/huggingface/transformers/blob/master/examples/run_language_modeling.py). For example, the default epoch count (`num_train_epochs`) is 1.\r\n\r\nHope this helps!", "> @xihui-wu To get the hyperparameters specific to the model (in this case `gpt2`), you can check the config file of `gpt2` with the code below:\r\n> \r\n> ```\r\n> from transformers import GPT2Config\r\n> print(GPT2Config())\r\n> ```\r\n> \r\n> Some higher level hyperparameters are still not included here (e.g. \"epochs\"). These can be set explicitly as arguments when running the CLI `run_language_modeling.py`; otherwise, the default values are used.\r\n> \r\n> You can find the hyperparameters and their default values at the beginning of the `main` function in [run_language_modeling.py](https://github.com/huggingface/transformers/blob/master/examples/run_language_modeling.py). For example, the default epoch count (`num_train_epochs`) is 1.\r\n> \r\n> Hope this helps!\r\n\r\nThanks a lot @enzoampil! Do you know what hyper-parameters to get the result: \"This takes about half an hour to train on a single K80 GPU and about one minute for the evaluation to run. It reaches a score of ~20 perplexity once fine-tuned on the dataset.\" ?", "@xihui-wu This result comes from running the default training script with no explicitly specified hyperparameters; therefore, the **default** hyperparameters will apply. \r\n\r\n>You can find the hyperparameters and their default values at the beginning of the main function in [run_language_modeling.py](https://github.com/huggingface/transformers/blob/master/examples/run_language_modeling.py). For example, the default epoch count (num_train_epochs) is 1.\r\n\r\nFor reference, this is the code snippet that fine-tunes `gpt2` with the default hyperparameters.\r\n```\r\nexport TRAIN_FILE=/path/to/dataset/wiki.train.raw\r\nexport TEST_FILE=/path/to/dataset/wiki.test.raw\r\n\r\npython run_language_modeling.py \\\r\n --output_dir=output \\\r\n --model_type=gpt2 \\\r\n --model_name_or_path=gpt2 \\\r\n --do_train \\\r\n --train_data_file=$TRAIN_FILE \\\r\n --do_eval \\\r\n --eval_data_file=$TEST_FILE\r\n```", "> @xihui-wu This result comes from running the default training script with no explicitly specified hyperparameters; therefore, the **default** hyperparameters will apply.\r\n> \r\n> > You can find the hyperparameters and their default values at the beginning of the main function in [run_language_modeling.py](https://github.com/huggingface/transformers/blob/master/examples/run_language_modeling.py). For example, the default epoch count (num_train_epochs) is 1.\r\n> \r\n> For reference, this is the code snippet that fine-tunes `gpt2` with the default hyperparameters.\r\n> \r\n> ```\r\n> export TRAIN_FILE=/path/to/dataset/wiki.train.raw\r\n> export TEST_FILE=/path/to/dataset/wiki.test.raw\r\n> \r\n> python run_language_modeling.py \\\r\n> --output_dir=output \\\r\n> --model_type=gpt2 \\\r\n> --model_name_or_path=gpt2 \\\r\n> --do_train \\\r\n> --train_data_file=$TRAIN_FILE \\\r\n> --do_eval \\\r\n> --eval_data_file=$TEST_FILE\r\n> ```\r\n\r\nI got GPU memory error with k80 on this, what's the batch_size and how can I configure?", "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", "> > @xihui-wu This result comes from running the default training script with no explicitly specified hyperparameters; therefore, the **default** hyperparameters will apply.\r\n> > > You can find the hyperparameters and their default values at the beginning of the main function in [run_language_modeling.py](https://github.com/huggingface/transformers/blob/master/examples/run_language_modeling.py). For example, the default epoch count (num_train_epochs) is 1.\r\n> > \r\n> > \r\n> > For reference, this is the code snippet that fine-tunes `gpt2` with the default hyperparameters.\r\n> > ```\r\n> > export TRAIN_FILE=/path/to/dataset/wiki.train.raw\r\n> > export TEST_FILE=/path/to/dataset/wiki.test.raw\r\n> > \r\n> > python run_language_modeling.py \\\r\n> > --output_dir=output \\\r\n> > --model_type=gpt2 \\\r\n> > --model_name_or_path=gpt2 \\\r\n> > --do_train \\\r\n> > --train_data_file=$TRAIN_FILE \\\r\n> > --do_eval \\\r\n> > --eval_data_file=$TEST_FILE\r\n> > ```\r\n> \r\n> I got GPU memory error with k80 on this, what's the batch_size and how can I configure?\r\n\r\nYou can use a per_device_train_batch_size=1, worked for me on a K80" ]