Datasets:

---

DanielPFlorian

/

Transformers-Github-Issues

like 0

[ { "id": 1897896961, "node_id": "MDU6TGFiZWwxODk3ODk2OTYx", "url": "https://api.github.com/repos/huggingface/transformers/labels/Migration", "name": "Migration", "color": "e99695", "default": false, "description": "" } ]

[]

[ "Hey @chenlin038,\r\n\r\nCan you copy paste the configs of `bert-base-1536` and `long-bert-1536` below? This way I can see exactly which configs you are using. Otherwise, I also can only refer to the answers in https://github.com/allenai/longformer/issues/106 .", "> Hey @chenlin038,\r\n> \r\n> Can you copy paste the configs of `bert-base-1536` and `long-bert-1536` below? This way I can see exactly which configs you are using. Otherwise, I also can only refer to the answers in [allenai/longformer#106](https://github.com/allenai/longformer/issues/106) .\r\n\r\nSorry for the late reply！This is the config of ### bert-base-1536:\r\n{\r\n \"architectures\": [\r\n \"BertForMaskedLM\"\r\n ],\r\n \"attention_probs_dropout_prob\": 0.1,\r\n \"directionality\": \"bidi\",\r\n \"gradient_checkpointing\": false,\r\n \"hidden_act\": \"gelu\",\r\n \"hidden_dropout_prob\": 0.1,\r\n \"hidden_size\": 768,\r\n \"initializer_range\": 0.02,\r\n \"intermediate_size\": 3072,\r\n \"layer_norm_eps\": 1e-12,\r\n \"max_position_embeddings\": 1536,\r\n \"model_type\": \"bert\",\r\n \"num_attention_heads\": 12,\r\n \"num_hidden_layers\": 12,\r\n \"pad_token_id\": 0,\r\n \"pooler_fc_size\": 768,\r\n \"pooler_num_attention_heads\": 12,\r\n \"pooler_num_fc_layers\": 3,\r\n \"pooler_size_per_head\": 128,\r\n \"pooler_type\": \"first_token_transform\",\r\n \"type_vocab_size\": 2,\r\n \"vocab_size\": 21128\r\n}\r\n\r\nThe following is the config of ### long-bert-1536:\r\n{\r\n \"architectures\": [\r\n \"BertLongForSequenceClassification\"\r\n ],\r\n \"attention_probs_dropout_prob\": 0.1,\r\n \"attention_window\": [\r\n 512,\r\n 512,\r\n 512,\r\n 512,\r\n 512,\r\n 512,\r\n 512,\r\n 512,\r\n 512,\r\n 512,\r\n 512,\r\n 512\r\n ],\r\n \"directionality\": \"bidi\",\r\n \"gradient_checkpointing\": false,\r\n \"hidden_act\": \"gelu\",\r\n \"hidden_dropout_prob\": 0.1,\r\n \"hidden_size\": 768,\r\n \"initializer_range\": 0.02,\r\n \"intermediate_size\": 3072,\r\n \"layer_norm_eps\": 1e-12,\r\n \"max_position_embeddings\": 1536,\r\n \"model_type\": \"bert\",\r\n \"num_attention_heads\": 12,\r\n \"num_hidden_layers\": 12,\r\n \"pad_token_id\": 0,\r\n \"pooler_fc_size\": 768,\r\n \"pooler_num_attention_heads\": 12,\r\n \"pooler_num_fc_layers\": 3,\r\n \"pooler_size_per_head\": 128,\r\n \"pooler_type\": \"first_token_transform\",\r\n \"type_vocab_size\": 2,\r\n \"vocab_size\": 21128\r\n}\r\n", "Ok, yeah I'm not very surprised that `bert-base-1536` is faster in your case. Longformer should mainly be used to prevent out-of-memory problems for long sequences. To do so it uses a complex attention mechanism than BERT which makes it a bit slower (especially for shorter sequences). So, in your case I would expect the Longformer model to use less memory than the BERT model, but not necessarily to be faster. If `bert-base-1536` fits in memory, then I think it's a good idea to use the model", "> Ok, yeah I'm not very surprised that `bert-base-1536` is faster in your case. Longformer should mainly be used to prevent out-of-memory problems for long sequences. To do so it uses a complex attention mechanism than BERT which makes it a bit slower (especially for shorter sequences). So, in your case I would expect the Longformer model to use less memory than the BERT model, but not necessarily to be faster. If `bert-base-1536` fits in memory, then I think it's a good idea to use the model\r\n\r\nDoes the longformer use less memory only have an effect on the GPU, or even specific GPU types, such as Nvidia Ampere? If I use CPU as a inference device, can I save more memory? Is there currently any optimization for the storage or calculation of sparse matrices by the CPU?", "You should definitely see an improvement in CPU memory usage when using longformer! ", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored.", "Hi, It seems I have a similar issue. However, the time difference between Longformer and Roberta is nearly a factor 10. Does it seem normal? After checking it is in the longformer.encoder step that this is quite slow. Here is the config I use (I have similar factor whatever the config):\r\nLongFormer_TOY_MODEL_HPARAMS = {\r\n \"vocab_size\": len(LongFormer_VOCAB),\r\n \"hidden_size\": 64,\r\n \"num_hidden_layers\": 3,\r\n \"num_attention_heads\": 8,\r\n \"intermediate_size\": 32,\r\n \"hidden_act\": \"gelu\",\r\n \"hidden_dropout_prob\": 0.1,\r\n \"attention_probs_dropout_prob\": 0.1,\r\n \"max_position_embeddings\": 512\r\n + 2, # tokenizer's model_max_length + 2 (<s> / </s> tokens of sequence)\r\n \"initializer_range\": 0.02,\r\n \"layer_norm_eps\": 1e-12,\r\n \"attention_window\": 512\r\n}\r\n\r\n\r\n\r\nThanks!" ]

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

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

[]

[]

[ "Hello! Could you provide the commands you used to launch the script, and where you obtained the file from? Thanks.", "> Hello! Could you provide the commands you used to launch the script, and where you obtained the file from? Thanks.\r\n\r\nI found that I made a mistake when saving the pretrained model, after fix this bug, the script converts the pytorch model correctly. \r\nThanks for your time!", "@QixinLi , how did you converti this one? I am having similar type of problem while converting xlmroberta to tf. \r\n\r\nmy code: https://colab.research.google.com/drive/17mOz39gXNHjeGN9tT4oJMBLvSWkgKlDN?usp=sharing\r\n\r\n```\r\nAttributeError Traceback (most recent call last)\r\n<ipython-input-6-37732b1a66b9> in <module>()\r\n 2 '/content/drive/MyDrive/Colab_Notebooks/models/pytorch_xlmr/pytorch_model.bin',\r\n 3 '/content/drive/MyDrive/Colab_Notebooks/models/pytorch_xlmr/config.json',\r\n----> 4 '/content/drive/MyDrive/Colab_Notebooks/models/pytorch_xlmr')\r\n\r\n2 frames\r\n/usr/local/lib/python3.7/dist-packages/transformers/modeling_tf_pytorch_utils.py in load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs, allow_missing_keys)\r\n 179 continue\r\n 180 \r\n--> 181 raise AttributeError(f\"{name} not found in PyTorch model\")\r\n 182 \r\n 183 array = pt_state_dict[name].numpy()\r\n\r\nAttributeError: lm_head.bias not found in PyTorch model\r\n```\r\n" ]

[]

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

[ "Hey @ZhaoQianfeng I think I know what you mean! So in short you are saying that `beam_hyp.add(...)` should be behaving differently depending on whether it finished with or without EOS token right? \r\n\r\nStill not sure whether this is a real problem though... -> Could you maybe open a PR that quickly shows the changes you would want beam search to have and we can take a look at some code? This would be super helpful! Thanks a lot for diving into the code :-) ", "Hello @patrickvonplaten. Glad you know what i mean. Those hypotheses which finished without EOS are calculated with longer length than they should be, so they have higher scores than they should be, this may cause them to compete unfairly with those with EOS and make us incorrectly throw away those \"with EOS\" hypotheses . \r\nMaybe it is not a severe problem, above situation maybe not common. Anyway, I modify the code here #8890 but haven't test it.I will be happy if it helps.", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

[]

[]

[ "The comment you are mentioning was about the old `run_language_modeling` script, and probably with some more options for a K80 that what you are running the script with (we should probably remove it or update with a proper command that gives those results). This doesn't look like a memory leak problem, you just don't have enough GPU memory to run the this large model with its full sequence length (of 1,024). You could try:\r\n- a smaller batch size with `--per_device_batch_size 4` or even 2 (or use gradient accumulation)\r\n- a smaller sequence length with `--block_size 512` or even 256\r\n- a smaller model with `--model_name_or_path gpt2-medium` or even distilgpt2.\r\n", "The smaller `--per_device_train_batch_size 2` batch size seems to be working for me. Just started the training process. Thank you very much for the extremely quick response, and for being an OSS maintainer @sgugger! \r\n\r\nI'll likely drop one more update in this thread to confirm that it worked all the way through.", "Can confirm - your advice works for me. \r\n\r\nIn fact, I managed to retrain even the XL on T100 GPUs on the new p4d.24xl instances. Definitely high mem requirements, but doable with `--model_name_or_path gpt2-xl --per_device_train_batch_size 1 --block_size 512`\r\n\r\nThanks, team! Y'all have a https://buymeacoffee.com account I can send some brews to? I appreciate your work.", "Hi, \r\n\r\n@sgugger I'm getting the same out of memory error on G Colab, as @erik-dunteman mentioned and I am using the smallest model of distilgpt2, I followed the advice here and added the additional argument to my command:\r\n\r\n```python\r\n!python /content/transformers/examples/language-modeling/run_clm.py \\\r\n --model_name_or_path distilgpt2 \\\r\n --train_file /content/train.txt \\\r\n --per_device_batch_size 2 \\\r\n --do_train \\\r\n --output_dir model_output\r\n```\r\n\r\nbut am now getting the error:\r\n\r\n`ValueError: Some specified arguments are not used by the HfArgumentParser: ['--per_device_batch_size', '2']`\r\n\r\nAlso, I could not find many parameters that were previously supported by ``run_language_modeling.py`` such as ``--line_by_line``. Were these removed in ``run_clm``? Is there a place where all possible arguments are listed?\r\n\r\nThanks", "The correct argument name is `--per_device_train_batch_size` or `--per_device_eval_batch_size`.\r\n\r\nThee is no `--line_by_line` argument to the `run_clm` script as this option does not make sense for causal language models such as GPT-2, which are pretrained by concatenating all available texts separated by a special token, not by using individual sentences with padding (like masked language models).\r\n\r\nTo list all available arguments, just use -h or --help as an option for the script. ", "> The correct argument name is `--per_device_train_batch_size` or `--per_device_eval_batch_size`.\r\n> \r\n> Thee is no `--line_by_line` argument to the `run_clm` script as this option does not make sense for causal language models such as GPT-2, which are pretrained by concatenating all available texts separated by a special token, not by using individual sentences with padding (like masked language models).\r\n> \r\n> To list all available arguments, just use -h or --help as an option for the script.\r\n\r\nThanks, I figured out the ``--per_device_train_batch_size `` parameter and got ``run_clm`` to work.\r\n\r\nI really need ``--line-by-line`` for my dataset as the training dataset is just individual sentences, where the next sentence has no connection with the previous.\r\n\r\nIs there any way to get ``line by line `` to work with ``run_clm``?\r\n\r\nThanks", "As I said, this makes no sense for those types of models so this won't be in our official examples. You can adapt the part that does this in `run_mlm` for your own need.", "I get the same out of memory error because it tries to run this on my 1050 ti instead of my k80. I exported CUDA_VISIBLE_DEVICES=1,2 which is my k80, but this script always runs on my tiny 1050ti. Is there a switch to set which gpu to use?", "@LysandreJik @sgugger can we load data in RAM in batches i.e lazy loading of data in RAM from disks and delete it after training on specific data?\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": 1897896961, "node_id": "MDU6TGFiZWwxODk3ODk2OTYx", "url": "https://api.github.com/repos/huggingface/transformers/labels/Migration", "name": "Migration", "color": "e99695", "default": false, "description": "" } ]

[]

[ "If I understand correctly, you're trying to load a pytorch model from a `pytorch_model.bin`? If so, have you taken a look at the [quickstart](https://huggingface.co/transformers/v2.4.0/quickstart.html#main-concepts)? The `from_pretrained` method is probably what you're looking for.", "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" ]

[]

[]

[ "Oops realized it was because this weight tying doesn't work across different architectures." ]

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

[]

[ "Hi there. There is little anyone can do to help without knowing the actual command you are running.\r\n", "Hi\nI am running\nhttps://github.com/huggingface/transformers/blob/master/examples/seq2seq/finetune_trainer.py\nonce\nwith multiple gpu machines once with 1 machine,\nI have adapted it for my usecase, this is hard for me to provide the exact\ncommand, as I need to share the whole codebase.\nMy question is though more general I observe performance differences if you\nrun this code on 1 dataset with multiple gpu/gpu, any thoughts on this?\nthanks\nRabeeh\n\nOn Mon, Nov 23, 2020 at 1:16 AM Sylvain Gugger <[email protected]>\nwrote:\n\n> Hi there. There is little anyone can do to help without knowing the actual\n> command you are running.\n>\n> —\n> You are receiving this because you authored the thread.\n> Reply to this email directly, view it on GitHub\n> <https://github.com/huggingface/transformers/issues/8718#issuecomment-731871920>,\n> or unsubscribe\n> <https://github.com/notifications/unsubscribe-auth/ABP4ZCHAF5CQRJP2LDI2XQDSRGSUPANCNFSM4T6X5WFQ>\n> .\n>\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", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

[]

[]

[ "I like it! ", "Great, I am glad that you did like it.\r\n\r\nThanks @patrickvonplaten and @jplu for your feedback.\r\n\r\n@patrickvonplaten :\r\nI have adjusted all your code review, and also add it to the T5 documentation.\r\nThe only missing part is the tests, it will be great if you could add it.\r\n\r\n@jplu :\r\nI have removed the unnecessary parameters from the TF model.\r\n\r\nIs there anything else needed from my side to merge the pull request ?", "> Great, I am glad that you did like it.\r\n> \r\n> Thanks @patrickvonplaten and @jplu for your feedback.\r\n> \r\n> @patrickvonplaten :\r\n> I have adjusted all your code review, and also add it to the T5 documentation.\r\n> The only missing part is the tests, it will be great if you could add it.\r\n> \r\n> @jplu :\r\n> I have removed the unnecessary parameters from the TF model.\r\n> \r\n> Is there anything else needed from my side to merge the pull request ?\r\n\r\nI think that's great! I'll fix the tests and merge :-) ", "## Update:\r\n\r\nPR is ready for review IMO. Would be great if @LysandreJik @jplu and @sgugger you can take a look :-) ", "> ## Update:\r\n> PR is ready for review IMO. Would be great if @LysandreJik @jplu and @sgugger you can take a look :-)\r\n\r\nThanks a lot @patrickvonplaten ^_^", "> Very clean implementation, thanks a lot @agemagician!\r\n\r\nYou are welcome.\r\nI am glad that I could help making the library better, even with a small contribution.\r\nI have to say without @patrickvonplaten help, I could not make it ^_^" ]

[]

[]

[ "> Normally the model attribute of the Trainer is always a reference to the real model (without the module from DataParallel and the likes), so using self.model here should prevent this error.\r\n\r\nIt did - thank you!\r\n\r\nThis is a very ambiguous situation for a user who wants to use HF trainer in their code. When to use `model` the argument and when `self.model`.\r\n\r\nWhat happens here is `model = torch.nn.DataParallel(self.model)` in the previous frame (`src/transformers/trainer.py:prediction_loop`), so `model` no longer has its normal methods accessible.\r\n\r\nHere are some possible solutions to resolve this ambiguity:\r\n\r\n1. monkeypatch `torch.nn.DataParallel` to expand its API to support all the methods of the original model transparently by installing a catch all `__getattr__` and remap all the failed method look ups to delegate to `self.module`.\r\n\r\n2. not to call the function argument `model` anymore, since it isn't under multi gpu, but is something else. \r\n\r\n3. remove the `model` argument completely + document to always use `self.model` - currently in `seq2seq_trainer.py `once we switch to `self.model`, `prediction_step()` no longer needs `model` as an argument (but is it always the case?)\r\n\r\n", "We can certainly improve the documentation and the debugging experience. I think I prefer the solution 2 since 1. is too magic (so will probably make things harder to debug) and 3 is not compatible with the regular `Trainer` (that needs the unwrapped model though I'd need to check to be sure).\r\n\r\nDoing `model` -> `wrapped_model` should be enough to clarify things? Wdyt\r\n", "> [...] 3 is not compatible with the regular `Trainer` (that needs the unwrapped model though I'd need to check to be sure).\r\n\r\nDid you mean to say \"needs the wrapped model\"?\r\n\r\nUnless I'm misreading what you wrote 3rd solution is the right one, since the Trainer doesn't do anything with the wrapped model. I don't know though whether this is so everywhere.\r\n\r\nThe 4th solution is passing `self.model `as the `model` arg, and making the wrapped model available via `self.wrapped_model` if the user needs it.\r\n\r\n> Doing `model` -> `wrapped_model` should be enough to clarify things? Wdyt\r\n\r\nExcept it won't be wrapped per se most of the time - very confusing to the user. Currently it should be called `may_be_wrapped_model_use_self_model_instead` variable ;)", "I meant the wrapped model, sorry.", " I'm getting this issue too using a T5 Model on multiple gpus\r\n\r\n`AttributeError: 'DataParallel' object has no attribute 'generate'`\r\n\r\nIs this supposed to be resolved? I've never seen this before. I've tried with 4.10.0 as well as current master branch", "@JamesDeAntonis \r\n\r\nIs it possible you somehow have a really old `transformers` in your `sys.path`?\r\n\r\nIf not, as always we need a way to reproduce the problem as the first step. And ideally in a new issue so that it can be tracked.\r\n\r\nBut you can also see the fix in this PR and try to trace it to where the `generate` call is made. Clearly it's not calling it on the correct object.\r\n\r\nThank you." ]

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

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

[]

[]

[ "Thanks for review @jplu . I'll update my code with review comments and new input processing.\r\n", "> Thank you very much for this very nice addition!!\r\n> \r\n> I left few comments on it. Also can you run the following piece of code and tell me if it works properly:\r\n> \r\n> ```\r\n> import tensorflow as tf\r\n> from transformers import GPT2Tokenizer, TFGPT2ForSequenceClassification\r\n> \r\n> model = tf.function(TFGPT2ForSequenceClassification.from_pretrained(\"microsoft/dialogrpt\"))\r\n> tokenizer = GPT2Tokenizer.from_pretrained(\"microsoft/dialogrpt\")\r\n> inputs = tokenizer(\"Hello\", return_tensors=\"tf\")\r\n> model(inputs)\r\n> ```\r\n> \r\n> @LysandreJik I would recommend as well to wait a bit that the new input processing to be merged.\r\n\r\n<img width=\"1176\" alt=\"output\" src=\"https://user-images.githubusercontent.com/6419011/100481765-2c2da900-311b-11eb-8fdd-15762f7d43df.png\">\r\n", "Hello @jplu and @LysandreJik ,\r\nI have refactored code as per review comments and added new input processing as well.\r\n\r\nKindly review.", "> Much better!! Thanks for the updates.\r\n> \r\n> There is still one comment to be addressed and the tests to fix.\r\n\r\n@jplu tests are also fixed now.", "@spatil6 we have merged today a PR that updates the way the booleans are processed. You can see an example in the TF BERT file for example, can you rebase and proceed to the same changes please. It would be awesome if you could do it!", "> @spatil6 we have merged today a PR that updates the way the booleans are processed. You can see an example in the TF BERT file for example, can you rebase and proceed to the same changes please. It would be awesome if you could do it!\r\n\r\nSure, will do that." ]

[]

[]

[ "seems to be related to #8613 ", "Fix https://github.com/huggingface/transformers/pull/8716\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 } ]

[]

[ "with t5-base seems to run fine, so perhaps the specific model in the README is not available. Thank you. ", "Hello! Indeed, it seems this model does not exist. Do you want to open a PR with a model that works? Thanks!", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

[]

[]

[ "If you have an accuracy near 10% on a two-label sequence classification task - does that mean it gets 90% of the results wrong? If so, you might just have switched the labels.", "Hi, no the problem is not related to what you said. I tried also to perform one hot encoding on the labels and change the loss function to \"categorical_crossentropy\" but the results are the same.\r\nI tried to use the official pre trained english model (**https://github.com/google-research/bert**) with another module and I don't have this problem (the keras model is the same).", "Hello!\r\n\r\nCan you try with `TFBertForSequenceClassification`?", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

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

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

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

[]

[]

[ { "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" } ]

[]

[]

[]

[]

[ "Hello, thanks for opening an issue! We try to keep the github issues for bugs/feature requests.\r\nCould you ask your question on the [forum](https://discusss.huggingface.co) instead?\r\n\r\nIt would be great if you could also include both training scripts, so that we may compare. There should be no difference between the PyTorch training or the TensorFlow training.\r\n\r\nThanks!" ]

[]

[ { "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 } ]

[ "duplicate of https://github.com/huggingface/transformers/issues/8643. \r\n\r\nThis is indeed a big problem. I'll try to get to it this week!", "Hey @FL33TW00D - actually I cannot reproduce your error....Can you try to update to the `tokenizers` version as well?", "I can correctly shorten T5's embedding matrix...", "Hi @patrickvonplaten,\r\nAppreciate you looking at this. \r\n\r\nI suspect that in this case it's user error.\r\nI am attempting to add the special tokens like so prior to pretraining:\r\n```python\r\nfrom transformers import T5TokenizerFast, T5ForConditionalGeneration\r\nMODEL_NAME = 'google/t5-v1_1-base'\r\nspecial_tokens = [\"<ORG>\",\r\n \"<PERSON>\"]\r\ntokenizer = T5TokenizerFast.from_pretrained('t5-base')\r\nspecial_tokens_dict = {'additional_special_tokens': ['<ORG>','<PERSON>']}\r\nnum_added_tokens = tokenizer.add_special_tokens(special_tokens_dict)\r\nprint(f'ADDED TOKENS: {num_added_tokens}')\r\nmodel = T5ForConditionalGeneration.from_pretrained(MODEL_NAME)\r\nmodel.resize_token_embeddings(len(tokenizer))\r\nmodel.to(\"cuda\")\r\n```\r\nI then pretrain the model, and save like so:\r\n`model.save_pretrained('t5_base_test')`\r\n\r\nIt is upon model loading that I receive the error:\r\n```\r\nT5ForConditionalGeneration.from_pretrained('./t5_base_test')\r\n```\r\n```\r\nsize mismatch for lm_head.weight: copying a param with shape torch.Size([32128, 768]) from checkpoint, the shape in current model is torch.Size([32102, 768]).\r\n```\r\n\r\nFrom the config.json, it looks like the rest of the layers are being scaled to the len(tokenizer) of 32102, and only the language modelling head on the final layer remaining as 32128.\r\n\r\nAny insight into this?\r\n\r\nMany Thanks,\r\nChris", "I can reproduce - will fix it! Thanks for the detailed error description ", "BTW, it's recommend to always use the same model identifier for model and tokenizer, even though in this case it would not have made a difference. So:\r\n\r\n```python\r\ntokenizer = T5TokenizerFast.from_pretrained('google/t5-v1_1-base')\r\n```", "> \r\n> \r\n> I can reproduce - will fix it! Thanks for the detailed error description\r\n\r\nMassive thanks for fixing this. Really appreciate it.", "Will try to have it merged into master by tomorrow", "> Will try to have it merged into master by tomorrow\r\n\r\n@patrickvonplaten No worries already forked and working great! :+1: ", "Hi @patrickvonplaten \r\n\r\nI have faced the same situation with `MT5ForConditionalGeneration` when I have reproduced the [question_generation](https://github.com/patil-suraj/question_generation) with Thai language data (I have prepared the dataset from ['xquad.th'](https://huggingface.co/datasets/viewer/?dataset=xquad&config=xquad.th) ) by @patil-suraj \r\n\r\nThis is my error messages\r\n```bash\r\n>>> model = MT5ForConditionalGeneration.from_pretrained('./mt5-base-qg-hl-xquad-th-6-epochs')\r\nTraceback (most recent call last):\r\n File \"<stdin>\", line 1, in <module>\r\n File \"/Users/sakares/transformers/src/transformers/modeling_utils.py\", line 1144, in from_pretrained\r\n raise RuntimeError(\r\nRuntimeError: Error(s) in loading state_dict for MT5ForConditionalGeneration:\r\n\tsize mismatch for lm_head.weight: copying a param with shape torch.Size([250112, 768]) from checkpoint, \r\nthe shape in current model is torch.Size([250102, 768]).\r\n```\r\n\r\nI have dive into the file src/transformers/models/mt5/modeling_mt5.py and found that MT5ForConditionalGeneration \r\n just overrode the T5ForConditionalGeneration, and it should not be a problem.\r\n\r\nSorry to bring you here @patil-suraj . I am just curious since I have modified the script run_qg.py for MT5\r\nand, according to this [discussion](https://github.com/huggingface/transformers/pull/8880#issuecomment-737113053), I found the script did not have things like \r\n`model.resize_token_embeddings(len(tokenizer))`\r\n\r\nMy question: should I run method resize_token_embeddings before start the training model.\r\n\r\n", "also facing same issue as @sakares ..Have you solved it?", "Can you guys try again on master -> this should have been fixed by now: https://github.com/huggingface/transformers/issues/9055#issuecomment-745450713", "@acul3 No luck yet. But I found the alternative solution with PyTorch Lightning script [\"Finetune MT5 for Question Generation in Hindi\"](https://www.kaggle.com/parthplc/finetune-mt5-for-question-generation-in-hindi/) and it works as expected", " i manage to solve my problem by changing tokenizer to `google/mt5-base' instead of 't5-base'(my mistake) and install transformers from source(master) as @patrickvonplaten told\r\n\r\nwill try to look that script @sakares..thank you" ]

[]

[]

[ "Hello, \r\nA little update, I think I fixed this issue. \r\nThe tokenizer returns a tensor of shape (n_passages, n_sequence_length) but only because I have duplicated the question like [questions] * n_passages. It was not clear on the documentation since I thought it was automatically done. " ]

[]

[]

[ "mT5 is not pretrained on downstream tasks like T5 was - see: https://huggingface.co/transformers/master/model_summary.html#mt5 \r\n\r\nSo it not surprising that mT5 won't work well out-of-the-box without fine-tuning.", "Ah, I hadn't realised that. But in that case, wouldn't the expected output be a reconstruction of the input?", "> Ah, I hadn't realised that. But in that case, wouldn't the expected output be a reconstruction of the input?\r\n\r\nHard to say if the model does not include any sentinel tokens (`<extra_id_1>`) and if one uses `generate()` instead of just the forward pass.... . Wolud be interesting to play around with the two pre-trained model variants though and see what differences they show...", "I agree that I would only get reconstruction if the decoding setup matched training :) Can you point me at any documentation that describes what special tokens are expected? I dug around in your implementation and the official repo but couldn't see anything. The output of `tokenizer.prepare_seq2seq_batch()` is the same for src and tgt as well (presumably because it uses the T5 tokenizer - does it not need its own?)\r\n\r\nEdit: Looking again, it seems like the sentinel tokens are just the equivalent of `[MASK]`? In which case the model should be able to reconstruct the input if it has access to the full (un-noised) sequence.", "Maybe these pointers help:\r\n- https://github.com/huggingface/transformers/issues/7451\r\n- https://github.com/huggingface/transformers/issues/7910\r\n- https://github.com/huggingface/transformers/issues/3985\r\n\r\nmT5 is pretrained exactly like T5 only without the downstream supersived training mixin. I think the T5 paper should explain in detail how this in done.", "Does anybody have some more pointers on how to use (train) the mT5 model that has been added to master for text generation? Anything explaining how the finetuning is done in practice using Huggingface Transformers would be greatly appreciated!", "Hey @Rijgersberg, what exactly do you mean by text generation ? GPT2-like open-end text generation?", "Well not open-end text generation in the sense of \"writing\", but using text-to-text generation to perform all types of different NLP tasks with little to no training. Basically what the GPT-3-paper calls \"few shot learning\".\r\n\r\nSpecifically, I would be interested in replicating the [WT5?! Training Text-to-Text Models to Explain their Predictions](https://arxiv.org/abs/2004.14546) results in languages other than English. But I'm having some trouble understanding what the differences between the T5 and mT5 models in Transformers mean for accomplishing that task.", "Hey @tomhosking how did you use MT5ForConditionalGeneration, T5Tokenizer\r\nI used \r\n```\r\npip install transformers\r\n```\r\nBut it is showing \r\n```\r\nImportError: cannot import name 'MT5ForConditionalGeneration'\r\n```\r\nHow can we install it?🤔\r\n", "@parthplc You can specify version of package You would like to install. For me it was experimental: `transformers==4.0.0rc1` and it works fine.\r\n\r\nFor training mT5 model for generating summary You can check out [this](https://towardsdatascience.com/fine-tuning-a-t5-transformer-for-any-summarization-task-82334c64c81) post. It worked for me.\r\n\r\n[edit]\r\nI forgot to mention, the only modification You have to make is to replace `T5ForConditionalGeneration` with `MT5ForConditionalGeneration`.", "> Well not open-end text generation in the sense of \"writing\", but using text-to-text generation to perform all types of different NLP tasks with little to no training. Basically what the GPT-3-paper calls \"few shot learning\".\r\n> \r\n> Specifically, I would be interested in replicating the [WT5?! Training Text-to-Text Models to Explain their Predictions](https://arxiv.org/abs/2004.14546) results in languages other than English. But I'm having some trouble understanding what the differences between the T5 and mT5 models in Transformers mean for accomplishing that task.\r\n\r\nIn this case, I would just fine-tune mT5 with the normal causal language modeling objective meaning:\r\n\r\n```python\r\nfrom transformers import MT5ForConditionalGeneration, T5Tokenizer \r\nmt5 = MT5ForConditionalGeneration.from_pretrained(\"google/mt5-base\")\r\nmt5_tok = T5Tokenizer.from_pretrained(\"google/mt5-base\")\r\n\r\ninput_ids = mt5_tok(\"explain sentiment: I went to see this movie with my husband, and we both thought the acting was terrible!\", return_tensors=\"pt\").input_ids # in the language of your choice\r\nlabels = mt5_tok(\"negative explanation: the acting was terrible.\", return_tensors=\"pt\").input_ids # in the language of your choice\r\n\r\nloss = mt5(input_ids=input_ids, labels=labels).loss\r\n```\r\n\r\nI took one of the visual examples of the paper you mentioned.\r\n\r\nIn short, there is no difference in how mt5 and t5 should be fine-tuned.\r\n\r\nAlso, @mrm8488 already successfully fine-tuned an mT5 model: https://twitter.com/mrm8488/status/1329478063768350723\r\nsorry to ping you here @mrm8488 - but maybe you have some tips/tricks for mt5 fine-tuning? \r\n\r\nAlso pinging our T5 fine-tuning expert @patil-suraj ", "> Well not open-end text generation in the sense of \"writing\", but using text-to-text generation to perform all types of different NLP tasks with little to no training. Basically what the GPT-3-paper calls \"few shot learning\".\r\n\r\nI'm not sure if you can use mT5 with no training (fine-tuning), since it was not pre-trained with any supervised objective like `T5`. \r\n\r\nOne experiment to try is to fine-tune `mT5` on the english data and see if it works for your language without any language specific fine-tuning (In my experiments, `T5` trained on English SQuAD for que gen gave interesting results for French and German without any language specific fine-tuning).\r\n\r\nBut for better results you should fine-tune `mT5` on the language specific dataset.\r\n\r\nAnd also as Patrick said, you can fine-tune `mT5` and `T5` the same way. \r\nThe major differences between `mT5` and `T5` are\r\n- `mT5` is based on [`T51.1`](https://github.com/google-research/text-to-text-transfer-transformer/blob/master/released_checkpoints.md#t511)\r\n- pre-trained on 101 languages\r\n- no supervised pre-training", "Hi, I slightly modified the script provided by @patil-suraj to fine-tune [`T5` on SQUAD] (https://github.com/patil-suraj/exploring-T5/blob/master/T5_on_TPU.ipynb) and after many epochs (I think I am missing anything/doing something wrong) I got 'decent' results fine-tuning mT5-small on tydiQA for multilingual QA https://huggingface.co/mrm8488/mT5-small-finetuned-tydiqa-for-xqa. The [PR with the model card](https://github.com/huggingface/transformers/pull/8729) for more details is not approved yet.", "> Hi, I slightly modified the script provided by @ patil-suraj to fine-tune [`T5` on SQUAD] (https://github.com/patil-suraj/exploring-T5/blob/master/T5_on_TPU.ipynb) and after many epochs (I think I am missing anything/doing something wrong) I got 'decent' results fine-tuning mT5-small on tydiQA for multilingual QA https://huggingface.co/mrm8488/mT5-small-finetuned-tydiqa-for-xqa. The [PR with the model card](https://github.com/huggingface/transformers/pull/8729) for more details is not approved yet.\r\n\r\njust merged it :-) BTW, you can now directly create the model cards online - no need for PRs anymore ;-)", "> > Well not open-end text generation in the sense of \"writing\", but using text-to-text generation to perform all types of different NLP tasks with little to no training. Basically what the GPT-3-paper calls \"few shot learning\".\r\n> \r\n> I'm not sure if you can use mT5 with no training (fine-tuning), since it was not pre-trained with any supervised objective like `T5`.\r\n> \r\n> One experiment to try is to fine-tune `mT5` on the english data and see if it works for your language without any language specific fine-tuning (In my experiments, `T5` trained on English SQuAD for que gen gave interesting results for French and German without any language specific fine-tuning).\r\n> \r\n> But for better results you should fine-tune `mT5` on the language specific dataset.\r\n> \r\n> And also as Patrick said, you can fine-tune `mT5` and `T5` the same way.\r\n> The major differences between `mT5` and `T5` are\r\n> \r\n> * `mT5` is based on [`T51.1`](https://github.com/google-research/text-to-text-transfer-transformer/blob/master/released_checkpoints.md#t511)\r\n> * pre-trained on 101 languages\r\n> * no supervised pre-training\r\n\r\nhey @patil-suraj @mrm8488 how can we finetune mT5 for other languages. Let's suppose we have language translation problem for any language other than English and if we finetune using T5 tokenizer we would be replacing each word with unk tokens. how will it be fine-tuned? eg.\r\n```\r\nprint(tokenizer.decode(data['source_ids']))\r\nprint(tokenizer.decode(data['target_ids']))\r\n```\r\n```\r\nEnglish to Hindi: Tell me the name of the ninth month.</s> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad>\r\n<unk> <unk> <unk> <unk> <unk> <unk> </s> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad> <pad>\r\n```\r\n", "@parthplc - I don't really understand your question. Since mT5 was trained on 101 languages it's tokenizer can obviously handle all those languages, *e.g.*:\r\n\r\n```python \r\nfrom transformers import AutoTokenizer\r\n\r\ntok = AutoTokenizer.from_pretrained(\"google/mt5-small\")\r\ntok.decode(tok(\"Der Satz wird auch definiert als sprachliche Einheit, die aus Subjekt und Prädikat besteht. Dies soll auf Aristoteles zurückgehen. Entsprechend definiert die traditionelle Grammatik den Satz als bestehend aus: Satzaussage (Prädikat), Satzergänzung (Objekt) und Satzgegenstand (Subjekt).\").input_ids) # gives no <unk> symbols\r\n```\r\n\r\nHopefully, this makes more sense now", "> ## Environment info\r\n> * `transformers` version: #9c0afdaf7b091c341072b432ad6ee17ba7a5016b\r\n> * Platform: Google colab\r\n> * Python version: 3.6.9\r\n> * PyTorch version (GPU?): 1.7.0\r\n> No GPU\r\n> \r\n> ### Who can help\r\n> mT5: @patrickvonplaten\r\n> \r\n> ## Information\r\n> Generating from `mT5-small` gives (nearly) empty output:\r\n> \r\n> ```\r\n> from transformers import MT5ForConditionalGeneration, T5Tokenizer\r\n> model = MT5ForConditionalGeneration.from_pretrained(\"google/mt5-small\")\r\n> tokenizer = T5Tokenizer.from_pretrained(\"google/mt5-small\")\r\n> article = \"translate to french: The capital of France is Paris.\"\r\n> batch = tokenizer.prepare_seq2seq_batch(src_texts=[article], return_tensors=\"pt\")\r\n> output_ids = model.generate(input_ids=batch.input_ids, num_return_sequences=1, num_beams=8, length_penalty=0.1)\r\n> tokenizer.decode(output_ids[0])\r\n> ```\r\n> \r\n> `>>> <pad> <extra_id_0></s>`\r\n> \r\n> Using the same input for T5 gives reasonable output:\r\n> \r\n> ```\r\n> from transformers import T5ForConditionalGeneration, T5Tokenizer\r\n> model = T5ForConditionalGeneration.from_pretrained(\"t5-small\")\r\n> tokenizer = T5Tokenizer.from_pretrained(\"t5-small\")\r\n> article = \"translate to french: The capital of France is Paris.\"\r\n> batch = tokenizer.prepare_seq2seq_batch(src_texts=[article], return_tensors=\"pt\")\r\n> output_ids = model.generate(input_ids=batch.input_ids, num_return_sequences=1, num_beams=8, length_penalty=0.1)\r\n> tokenizer.decode(output_ids[0])\r\n> ```\r\n> \r\n> `>>> <pad> La capitale de la France est Paris.</s>`\r\n> \r\n> My understanding is that mT5 is trained in the same way as T5, and should work in a very similar way?\r\n\r\nHi, I met the same problem when fine-tuning mt5 to a Chinese QG environment. I'm wondering if you have solved this issue?", "hi @nomoreoneday \r\n\r\n`mT5` is a pre-trained model and it's not finetuned on any downstream task, whereas T5 was already trained on translation task as part of its supervised pre-training mixture, which could explain the empty output of mT5.\r\n\r\nYou should fine-tune the model on your task, to use for generation. \r\n\r\n> I met the same problem when fine-tuning mt5 to a Chinese QG environment\r\n\r\nAnd if you are having trouble with fine-tuning then please post a shot code snippet so we can reproduce your issue.", "> hi @nomoreoneday\r\n> \r\n> `mT5` is a pre-trained model and it's not finetuned on any downstream task, whereas T5 was already trained on translation task as part of its supervised pre-training mixture, which could explain the empty output of mT5.\r\n> \r\n> You should fine-tune the model on your task, to use for generation.\r\n> \r\n> > I met the same problem when fine-tuning mt5 to a Chinese QG environment\r\n> \r\n> And if you are having trouble with fine-tuning then please post a shot code snippet so we can reproduce your issue.\r\n\r\nhi @patil-suraj \r\n\r\nthanks for replying. I'm trying to replicate your project(https://github.com/patil-suraj/question_generation) on a Chinese QG task. I got decent results when I run \r\n```\r\nexport CUDA_VISIBLE_DEVICES=0\r\npython3 eval.py \\\r\n --model_name_or_path mt5-small-ncp-qg-hl-base_epoch30 \\\r\n --valid_file_path data/valid_data_qa_hl_mt5_ncp_all_task.pt \\\r\n --model_type mt5 \\\r\n --num_beams 4 \\\r\n --max_decoding_length 32 \\\r\n --output_path hypothesis_mt5-small-ncp-qg-hl-base_epoch30_ncp_all_task.txt\r\n```\r\nBut when I trying to construct the pipeline and run:\r\n\r\n`\r\ndef _extract_answers(self,context):\r\n\r\n sents,inputs = self._prepare_inputs_for_ans_extraction(context)\r\n inputs = self._tokenize(inputs,padding = True,truncation = True) #encoding\r\n print(\"inputs after encoding:\",inputs)\r\n\r\n outs = self.ans_model.generate(\r\n input_ids = inputs['input_ids'].to(self.device),\r\n attention_mask = inputs['attention_mask'].to(self.device),\r\n max_length = 32,\r\n )\r\n\r\n dec = [self.ans_tokenizer.decode(ids,skip_special_tokens=False) for ids in outs] #decoding\r\n print(\"dec:\", dec)\r\n answers = [item.split('<sep>') for item in dec]\r\n print(\"answers1:\",answers)\r\n answers = [i[:-1] for i in answers]\r\n print(\"answers2:\",answers)\r\n\r\n return sents, answers\r\n`\r\n\r\nI got the empty answers. like this\r\n\r\n`dec: ['<pad> <extra_id_0></s>']\r\nanswers1: [['<pad> <extra_id_0></s>']]\r\nanswers2: [[]]`", "I wondering if there is any difference in data preprocessing between t5 and mt5. ", "> > hi @nomoreoneday\r\n> > `mT5` is a pre-trained model and it's not finetuned on any downstream task, whereas T5 was already trained on translation task as part of its supervised pre-training mixture, which could explain the empty output of mT5.\r\n> > You should fine-tune the model on your task, to use for generation.\r\n> > > I met the same problem when fine-tuning mt5 to a Chinese QG environment\r\n> > \r\n> > \r\n> > And if you are having trouble with fine-tuning then please post a shot code snippet so we can reproduce your issue.\r\n> \r\n> hi @patil-suraj\r\n> \r\n> thanks for replying. I'm trying to replicate your project(https://github.com/patil-suraj/question_generation) on a Chinese QG task. I got decent results when I run\r\n> \r\n> ```\r\n> export CUDA_VISIBLE_DEVICES=0\r\n> python3 eval.py \\\r\n> --model_name_or_path mt5-small-ncp-qg-hl-base_epoch30 \\\r\n> --valid_file_path data/valid_data_qa_hl_mt5_ncp_all_task.pt \\\r\n> --model_type mt5 \\\r\n> --num_beams 4 \\\r\n> --max_decoding_length 32 \\\r\n> --output_path hypothesis_mt5-small-ncp-qg-hl-base_epoch30_ncp_all_task.txt\r\n> ```\r\n> \r\n> But when I trying to construct the pipeline and run:\r\n> \r\n> `\r\n> def _extract_answers(self,context):\r\n> \r\n> ```\r\n> sents,inputs = self._prepare_inputs_for_ans_extraction(context)\r\n> inputs = self._tokenize(inputs,padding = True,truncation = True) #encoding\r\n> print(\"inputs after encoding:\",inputs)\r\n> \r\n> outs = self.ans_model.generate(\r\n> input_ids = inputs['input_ids'].to(self.device),\r\n> attention_mask = inputs['attention_mask'].to(self.device),\r\n> max_length = 32,\r\n> )\r\n> \r\n> dec = [self.ans_tokenizer.decode(ids,skip_special_tokens=False) for ids in outs] #decoding\r\n> print(\"dec:\", dec)\r\n> answers = [item.split('<sep>') for item in dec]\r\n> print(\"answers1:\",answers)\r\n> answers = [i[:-1] for i in answers]\r\n> print(\"answers2:\",answers)\r\n> \r\n> return sents, answers\r\n> ```\r\n> \r\n> `\r\n> \r\n> I got the empty answers. like this\r\n> \r\n> `dec: ['<pad> <extra_id_0></s>'] answers1: [['<pad> <extra_id_0></s>']] answers2: [[]]`\r\n\r\nhaving the same issue here", "I have finally overcome the ['<pad> <extra_id_0></s>'] issue and obtained decent post-training predictions with MT5, just had to \r\n1. lower the lr (set to 0.001, as indicated in mt5 paper) \r\nand \r\n2. train for a lot more epochs in comparison with T5 for the same task (60 epochs for MT5 vs 10 for T5 for a simple text style transfer task fine-tuning). \r\n\r\n@nomoreoneday I have no touched anything but the model name when switching between t5 and mt5 in my training pipeline, wonder if I should?", "> Hi, I slightly modified the script provided by @patil-suraj to fine-tune [`T5` on SQUAD] (https://github.com/patil-suraj/exploring-T5/blob/master/T5_on_TPU.ipynb) and after many epochs (I think I am missing anything/doing something wrong) I got 'decent' results fine-tuning mT5-small on tydiQA for multilingual QA https://huggingface.co/mrm8488/mT5-small-finetuned-tydiqa-for-xqa. The [PR with the model card](https://github.com/huggingface/transformers/pull/8729) for more details is not approved yet.\r\n\r\n@mrm8488 Hi, for this model mrm8488/mT5-small-finetuned-tydiqa-for-xqa , I tried to run your demo script, but failed with error loading the tokenizer. And the Hosted inference API on this page doesn't work as well.\r\n hope to see your feedback.", "> > hi @nomoreoneday\r\n> > `mT5` is a pre-trained model and it's not finetuned on any downstream task, whereas T5 was already trained on translation task as part of its supervised pre-training mixture, which could explain the empty output of mT5.\r\n> > You should fine-tune the model on your task, to use for generation.\r\n> > > I met the same problem when fine-tuning mt5 to a Chinese QG environment\r\n> > \r\n> > \r\n> > And if you are having trouble with fine-tuning then please post a shot code snippet so we can reproduce your issue.\r\n> \r\n> hi @patil-suraj\r\n> \r\n> thanks for replying. I'm trying to replicate your project(https://github.com/patil-suraj/question_generation) on a Chinese QG task. I got decent results when I run\r\n> \r\n> ```\r\n> export CUDA_VISIBLE_DEVICES=0\r\n> python3 eval.py \\\r\n> --model_name_or_path mt5-small-ncp-qg-hl-base_epoch30 \\\r\n> --valid_file_path data/valid_data_qa_hl_mt5_ncp_all_task.pt \\\r\n> --model_type mt5 \\\r\n> --num_beams 4 \\\r\n> --max_decoding_length 32 \\\r\n> --output_path hypothesis_mt5-small-ncp-qg-hl-base_epoch30_ncp_all_task.txt\r\n> ```\r\n> \r\n> But when I trying to construct the pipeline and run:\r\n> \r\n> `\r\n> def _extract_answers(self,context):\r\n> \r\n> ```\r\n> sents,inputs = self._prepare_inputs_for_ans_extraction(context)\r\n> inputs = self._tokenize(inputs,padding = True,truncation = True) #encoding\r\n> print(\"inputs after encoding:\",inputs)\r\n> \r\n> outs = self.ans_model.generate(\r\n> input_ids = inputs['input_ids'].to(self.device),\r\n> attention_mask = inputs['attention_mask'].to(self.device),\r\n> max_length = 32,\r\n> )\r\n> \r\n> dec = [self.ans_tokenizer.decode(ids,skip_special_tokens=False) for ids in outs] #decoding\r\n> print(\"dec:\", dec)\r\n> answers = [item.split('<sep>') for item in dec]\r\n> print(\"answers1:\",answers)\r\n> answers = [i[:-1] for i in answers]\r\n> print(\"answers2:\",answers)\r\n> \r\n> return sents, answers\r\n> ```\r\n> \r\n> `\r\n> \r\n> I got the empty answers. like this\r\n> \r\n> `dec: ['<pad> <extra_id_0></s>'] answers1: [['<pad> <extra_id_0></s>']] answers2: [[]]`\r\n\r\nSame too ", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored.", "> I have finally overcome the [' <extra_id_0>'] issue and obtained decent post-training predictions with MT5, just had to\r\n> \r\n> 1. lower the lr (set to 0.001, as indicated in mt5 paper)\r\n> and\r\n> 2. train for a lot more epochs in comparison with T5 for the same task (60 epochs for MT5 vs 10 for T5 for a simple text style transfer task fine-tuning).\r\n> \r\n> @nomoreoneday I have no touched anything but the model name when switching between t5 and mt5 in my training pipeline, wonder if I should?\r\n\r\nThankyou，I have same problem， I am trying train more epochs, to see if it can be correct", "> > I have finally overcome the [' <extra_id_0>'] issue and obtained decent post-training predictions with MT5, just had to\r\n> > \r\n> > 1. lower the lr (set to 0.001, as indicated in mt5 paper)\r\n> > and\r\n> > 2. train for a lot more epochs in comparison with T5 for the same task (60 epochs for MT5 vs 10 for T5 for a simple text style transfer task fine-tuning).\r\n> > \r\n> > @nomoreoneday I have no touched anything but the model name when switching between t5 and mt5 in my training pipeline, wonder if I should?\r\n> \r\n> Thankyou，I have same problem， I am trying train more epochs, to see if it can be correct\r\n\r\nDo you have any newer ideas about this problem?", "> > > hi @nomoreoneday\r\n> > > `mT5` is a pre-trained model and it's not finetuned on any downstream task, whereas T5 was already trained on translation task as part of its supervised pre-training mixture, which could explain the empty output of mT5.\r\n> > > You should fine-tune the model on your task, to use for generation.\r\n> > > > I met the same problem when fine-tuning mt5 to a Chinese QG environment\r\n> > > \r\n> > > \r\n> > > And if you are having trouble with fine-tuning then please post a shot code snippet so we can reproduce your issue.\r\n> > \r\n> > \r\n> > hi @patil-suraj\r\n> > thanks for replying. I'm trying to replicate your project(https://github.com/patil-suraj/question_generation) on a Chinese QG task. I got decent results when I run\r\n> > ```\r\n> > export CUDA_VISIBLE_DEVICES=0\r\n> > python3 eval.py \\\r\n> > --model_name_or_path mt5-small-ncp-qg-hl-base_epoch30 \\\r\n> > --valid_file_path data/valid_data_qa_hl_mt5_ncp_all_task.pt \\\r\n> > --model_type mt5 \\\r\n> > --num_beams 4 \\\r\n> > --max_decoding_length 32 \\\r\n> > --output_path hypothesis_mt5-small-ncp-qg-hl-base_epoch30_ncp_all_task.txt\r\n> > ```\r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > But when I trying to construct the pipeline and run:\r\n> > `\r\n> > def _extract_answers(self,context):\r\n> > ```\r\n> > sents,inputs = self._prepare_inputs_for_ans_extraction(context)\r\n> > inputs = self._tokenize(inputs,padding = True,truncation = True) #encoding\r\n> > print(\"inputs after encoding:\",inputs)\r\n> > \r\n> > outs = self.ans_model.generate(\r\n> > input_ids = inputs['input_ids'].to(self.device),\r\n> > attention_mask = inputs['attention_mask'].to(self.device),\r\n> > max_length = 32,\r\n> > )\r\n> > \r\n> > dec = [self.ans_tokenizer.decode(ids,skip_special_tokens=False) for ids in outs] #decoding\r\n> > print(\"dec:\", dec)\r\n> > answers = [item.split('<sep>') for item in dec]\r\n> > print(\"answers1:\",answers)\r\n> > answers = [i[:-1] for i in answers]\r\n> > print(\"answers2:\",answers)\r\n> > \r\n> > return sents, answers\r\n> > ```\r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > \r\n> > `\r\n> > I got the empty answers. like this\r\n> > `dec: ['<pad> <extra_id_0></s>'] answers1: [['<pad> <extra_id_0></s>']] answers2: [[]]`\r\n> \r\n> having the same issue here\r\n\r\nI am also having the same issue", "Sorry I'm loosing a bit track of what the problem is here. Note that `mt5` cannot generate coherent sentences out-of-the-box because it's only be pretrained on the span-mask filling task and not on any down-stream tasks." ]

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

[]

[ "This path is the default `cache_path` of the datasets library, not transformers. You can change it by setting an environment variable named `HF_HOME` to the path you want, the datasets will then be cached in this path suffixed with \"/datasets/\"", "Hi\nthank you, the huggingface codes gebnerally also create an empty folder\ntitled ' ' when I run it, which is specifies the caching folder address,\ncould it be possible not to create this folder?\nthanks\nBest\nRabeeh\n\nOn Sat, Nov 21, 2020 at 6:44 PM Sylvain Gugger <[email protected]>\nwrote:\n\n> This path is the default cache_path of the datasets library, not\n> transformers. You can change it by setting an environment variable named\n> HF_HOME to the path you want, the datasets will then be cached in this\n> path suffixed with \"/datasets/\"\n>\n> —\n> You are receiving this because you authored the thread.\n> Reply to this email directly, view it on GitHub\n> <https://github.com/huggingface/transformers/issues/8703#issuecomment-731611519>,\n> or unsubscribe\n> <https://github.com/notifications/unsubscribe-auth/ABP4ZCDCFJDCXZAPZ73ORHDSQ74AZANCNFSM4T5ZWXDA>\n> .\n>\n", "We would need to see the code you are running that creates this empty folder named `\" \"` to be able to help.", "Hi there\nI am training seq2seq_trainer codes. I have adapted it for my use case but\nin the original version of codes should also happen.\nthanks\nBest\nRabeeh\n\nOn Sun, Nov 22, 2020, 4:46 AM Sylvain Gugger <[email protected]>\nwrote:\n\n> We would need to see the code you are running that creates this empty\n> folder named \" \" to be able to help.\n>\n> —\n> You are receiving this because you authored the thread.\n> Reply to this email directly, view it on GitHub\n> <https://github.com/huggingface/transformers/issues/8703#issuecomment-731693963>,\n> or unsubscribe\n> <https://github.com/notifications/unsubscribe-auth/ABP4ZCCKYJ5TFIKGT2N7LTLSRCCP5ANCNFSM4T5ZWXDA>\n> .\n>\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", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

[]

[]

[ "After submitting I realized `F.softmax(F.log_softmax())` is equivalent to `F.softmax()`. The different probability values I noticed comes from the normalization of multiple beams in one dimension (line 1169). That doesn't change the behaviour of the sampling." ]

[]

[]

[ "Hey @parthplc,\r\n\r\nCould you try to post a very short code snippet that can reproduce the error. It's too time-consuming to go through such a big notebook sadly", "@parthplc Please post a solution if you close the issue. There is none in your colab as far as I can see (`model.resize_token_embeddings(len(tokenizer))` is the only relevant code I can see and it doesn't fix the problem)." ]

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

[]

[ "Hello! I'm not sure I completely understand the issue. Are you saying you do not obtain the results you expected when using the 8 cores of the TPU, vs using a single TPU core?", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

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

[]

[ "Yes, that model from @sshleifer does not bundle its own tokenizer, as you can see in the list of files: https://huggingface.co/sshleifer/t5-base-cnn/tree/main\r\n\r\nWe'll add this info to the model card, but you can just use the one from t5: `T5Tokenizer.from_pretrained(\"t5-base\")`", "@julien-c Thank you for your help", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

[]

[]

[ "Hi @ganeshjawahar , please have a look at the `run_NER_old.py` script! It should handle custom files 🤗 ", "Usage and more examples are documented here:\r\n\r\nhttps://github.com/huggingface/transformers/tree/master/examples/token-classification#old-version-of-the-script", "Thanks for the quick response. I'm able to make use of `run_ner_old.py` with my custom dataset. Is there a similar documentation to use `run_ner.py` with custom dataset? \r\n\r\nP.S.: `run_ner_old.py` loads all examples into RAM and that's a problem for me as my custom dataset is very large. I was thinking of getting around this issue by using `run_ner.py` which uses datasets library. ", "If you can provide a tiny example for csv or json format, that should be very helpful. 🤗", "Ah, I see, an example for a json-based file format can be found here:\r\n\r\nhttps://github.com/huggingface/transformers/blob/master/tests/fixtures/tests_samples/conll/sample.json\r\n\r\nAnother possibility would be, that you write a custom recipe with Hugging Face datasets library. Then you can run the `run_NER.py` script by passing the (local) path name of your recipe to the script. Just have a look at the CoNNL dataset/recipe:\r\n\r\nhttps://github.com/huggingface/datasets/blob/master/datasets/conll2003/conll2003.py\r\n\r\nYou could usw it as a template and modify it for your needs 🤗 ", "I think the JSON sample should be in the [token-classification README](https://github.com/huggingface/transformers/blob/master/examples/token-classification/README.md) for people trying to use `run_ner.py` from local files. Would you also be willing to provide a CSV sample? So far, I have found through trial, error, and code deciphering that:\r\n\r\n- The CSV needs to start with column names (not respecting this causes `ValueError: External features info don't match the dataset`)\r\n- The column separator should be a comma (`,`)\r\n- Text containing commas should be in double quotes (like this `\",\"`) to disambiguate columns\r\n- Literal double quotes should be escaped with `\\`\r\n\r\nRight now, my CSV file looks like this:\r\n```\r\ntoken,label\r\nDC,M \r\n##T,M \r\n##N,M \r\n##4,M \r\nas,O \r\na,O \r\nm,O \r\n##od,O \r\n##ifier,O\r\n...\r\n```\r\n\r\nI get the following error:\r\n```\r\nFile \"projects/github/transformers/examples/token-classification/run_ner.py\", line 221, in main\r\n if isinstance(features[label_column_name].feature, ClassLabel):\r\nAttributeError: 'Value' object has no attribute 'feature'\r\n```\r\n\r\nUsing the python debugger, I've found that `features[label_column_name] = Value(dtype='string', id=None)` but I don't know if this is expected behavior. I can only assume that it isn't, but I can't seem to figure out what else `features[label_column_name]` could or should be.\r\n\r\nI'm pretty much stuck, and knowing if the issue comes from the structure of my CSV would be very helpful.\r\n\r\nFurthermore, I've tried formatting my data as close as I could to the [JSON conll sample](https://github.com/huggingface/transformers/blob/master/tests/fixtures/tests_samples/conll/sample.json), but I get the following error:\r\n```\r\njson.decoder.JSONDecodeError: Extra data: line 2 column 1\r\n```\r\nAfter a little bit of googling, as I suspected it turns out one cannot have multiple JSON objects in one file. So if the intended JSON format for `run_ner.py` requires one JSON object per sequence but JSON files can't contain more than one JSON object, how can we get `run_ner.py` to work with several sequences in JSON mode?", "Exact same process/issue/errors as @gpiat. Would be very helpful if the format for the csv option for run_ner.py was explicitly defined in the readme. If there was a sample input for the csv option that is fully functional with the script it would be much more simple to modify our custom data to match the sample as opposed to writing a custom recipe.", "Same problem as @gpiat with CSV.\r\n@stefan-it And it seems the old script is no longer available?", "I believe I've solved the same problem as @gpiat , @millanbatra1234 and @AleksandrsBerdicevskis have had:\r\n\r\nReplace the `if isinstance(features[label_column_name].feature, ClassLabel):` in run_ner.py with `if hasattr(features[label_column_name], 'feature') and isinstance(features[label_column_name].feature, ClassLabel):`. \r\n\r\nI tried @gpiat's CSV format and that doesn't work. Instead, I used the JSON format, which looks like this:\r\n\r\n```\r\n{\"tokens\": [\"APPLICATION\", \"and\", \"Affidavit\", \"for\", \"Search\", \"Warrant\", \"as\", \"to\", \"The\", \"Matter\", \"of\", \"the\", \"Search\", \"of\", \"9\", \"Granite\", \"Street\", \",\", \"#\", \"5\", \"(\", \"Attachments\", \":\", \"#\", \"1\", \"Affidavit\", \"of\", \"James\", \"Keczkemethy)(Belpedio\", \",\", \"Lisa\", \")\", \"(\", \"Entered\", \":\", \"12/15/2020\", \")\"], \"tags\": [\"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"B-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"L-MISC\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\"]}\r\n{\"tokens\": [\"APPLICATION\", \"for\", \"Search\", \"Warrant\", \"by\", \"USA\", \"as\", \"to\", \"702\", \"-\", \"517\", \"-\", \"7282\", \"(\", \"KM\", \",\", \"ilcd\", \")\", \"(\", \"Entered\", \":\", \"12/10/2020\", \")\"], \"tags\": [\"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"B-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"L-MISC\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\"]}\r\n{\"tokens\": [\"APPLICATION\", \"AND\", \"AFFIDAVIT\", \"by\", \"USA\", \"as\", \"to\", \"4\", \"CELLULAR\", \"TELEPHONES\", \"SEIZED\", \"FROM\", \"THE\", \"FDC\", \"IN\", \"PHILADELPHIA\", \"AND\", \"CURRENTLY\", \"HELD\", \"BY\", \"THE\", \"FBI\", \"PHILADELPHIA\", \"DIVISION\", \"Re\", \":\", \"Search\", \"Warrant\", \"Issued\", \".\", \"(\", \"mac\", \",\", \")\", \"(\", \"Entered\", \":\", \"12/09/2020\", \")\"], \"tags\": [\"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"B-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"I-MISC\", \"L-MISC\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\", \"O\"]}\r\n```\r\n\r\nSo, yes, you can have more than one JSON object in the file. Each JSON object goes on its own line. This is sometimes called JSONL or JSONLINES format.\r\n\r\n", "@jeremybmerrill Thanks! Yes, JSON does work, I should have mentioned that (it actually does even without changing the code as you suggest). \r\n\r\n(With JSON, I run into another input problem (#9660), but I guess that's a different story.)", "In my case the json format didn't work due to this issue [github.com/huggingface/datasets/issues/2181](https://github.com/huggingface/datasets/issues/2181). Pyarrow can't handle json if the line size is too big. So I had to split large lines into smaller ones.", "JSON works, but CSV still does not work now", "CSV file input does not work! I converted it into JSON, so It works now. \r\n" ]

[]

[]

[]

[ { "id": 2627272588, "node_id": "MDU6TGFiZWwyNjI3MjcyNTg4", "url": "https://api.github.com/repos/huggingface/transformers/labels/Model%20Parallel", "name": "Model Parallel", "color": "8B66A5", "default": false, "description": "Model Parallelilsm Implementations" } ]

[]

[ "Would it be a good idea to support a less painful way of writing a device_map? This is because as the developer experiments with the mapping, the current method is very inefficient to modify the layer maps.\r\n\r\nPerhaps there could be more than one way to do it?\r\n\r\nInstead of:\r\n```\r\n device_map = {0: [0, 1, 2, 3, 4, 5, 6, 7, 8],\r\n 1: [9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21],\r\n 2: [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34],\r\n 3: [35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47]}\r\n\r\n```\r\n\r\ne.g. some ideas for perhaps much simpler ways to create such a map:\r\n\r\n* remap the layers as follows:\r\n```\r\ndevice_map = {\r\n devices: [0, 3, 4, 5],\r\n layer_split: [8, 12, 12, 12],\r\n}\r\n```\r\n* a simple string:\r\n```\r\ndevice_map = {\r\n devices: [0, 1, 2, 3],\r\n layer_map: \"1-10, 9-21, 22-34, 35-47\",\r\n}\r\n```\r\n* a simple string using slice notation\r\n```\r\ndevice_map = {\r\n devices: [0, 1, 2, 3],\r\n layer_slice: \"1:10, 9:21, 22:34, 35:47\",\r\n}\r\n```\r\n\r\nprobably several ways can be supported and a wrapper expand them into the explicit version used now based on the keys of the map in the argument.\r\n\r\nin either case, changing the map is much easier then...\r\n", "Not a bad idea. Adding to that: create a mapping utility that's been tested for larger model types like `device_map = get_device_map(machine = 4, model = \"gpt2-xl\")`. The first device should have fewer layers because it has the embedding and head. This is what I was using for testing:\r\n\r\n```\r\ndef get_device_map(machine: str, model_name: str) -> dict:\r\n \"\"\"Returns a dictionary optimized for distributing a model across\r\n several devices in a model parallel manner.\"\"\"\r\n if machine in [\"TeslaV100x4\", \"p3.8xlarge\", 4]:\r\n device_dict = {\r\n \"gpt2-xl\": {\r\n 0: [0, 1, 2, 3, 4, 5, 6, 7, 8],\r\n 1: [9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21],\r\n 2: [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34],\r\n 3: [35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47],\r\n },\r\n \"t5-large\": {\r\n 0: [0, 1, 2],\r\n 1: [3, 4, 5, 6, 7, 8, 9],\r\n 2: [10, 11, 12, 13, 14, 15, 16],\r\n 3: [17, 18, 19, 20, 21, 22, 23],\r\n },\r\n \"t5-3b\": {\r\n 0: [0, 1, 2],\r\n 1: [3, 4, 5, 6, 7, 8, 9],\r\n 2: [10, 11, 12, 13, 14, 15, 16],\r\n 3: [17, 18, 19, 20, 21, 22, 23],\r\n },\r\n \"t5-11b\": {\r\n 0: [0, 1, 2],\r\n 1: [3, 4, 5, 6, 7, 8, 9],\r\n 2: [10, 11, 12, 13, 14, 15, 16],\r\n 3: [17, 18, 19, 20, 21, 22, 23],\r\n },\r\n }\r\n return device_dict[model_name]\r\n```", "This is definitely a goodness to have in the library as well as it will save the developer start up time! I'd even extend it to a specific card size in the argument to `get_device_map`, since the map would be different depending on the size of the cards.\r\n\r\nBut this won't work for cards of different sizes, e.g. at the moment I have 1x 22GB + 1x 8GB cards. But perhaps this is an odd case and most serious setups have identical cards. I don't know. \r\n", "@stas00 Yeah, I think you're right. Could do something simple like `device_map` dictionary should be ranges like this:\r\n\r\n```\r\ndevice_map = {0: range(0, 10),\r\n 1: range(11, 24),\r\n ...}\r\n```\r\nSimpler than creating a list.", "well, it's the same just using python to save on typing ;) this is still awkward a bit as you have to count ;)\r\n\r\nhere there is less counting: I want you to use devices `[0,1,2,3]` and slice the layers as `[8, 12, 12, 12]` :)\r\n\r\nThat's why I'm suggesting to support more than one way.", "But most likely any of these custom ways can be easily delegated to a helper util, so the end result is the `device_map` as you implemented it. e..g:\r\n```\r\ndevice_map=device_map_make_by_partition([0,1,2,3], [8, 12, 12, 12])\r\ndevice_map=device_map_make_by_slice([0,1,2,3], \"1:10, 9:21, 22:34, 35:47\")\r\n```\r\n" ]

[]

[]

[ "@patil-suraj this pull request can be closed the typo has already been fixed. ", "Thanks for letting me know." ]

[]

[]

[]

[]

[]

[ "This was solved by https://github.com/huggingface/transformers/pull/8616\r\n\r\nThank you for your contribution!" ]

[]

[]

[ "There is an issue here because your sequence is now to long for your model. The model only supports sequences of size 512 or less, but this code:\r\n\r\n```py\r\nsequence_0 = \"The company HuggingFace is based in New York City\" * 100\r\nsequence_1 = \"Apples are especially bad for your health\"\r\nsequence_2 = \"HuggingFace's headquarters are situated in Manhattan\"\r\nparaphrase = tokenizer(sequence_0, sequence_2, return_tensors=\"pt\")\r\nnot_paraphrase = tokenizer(sequence_0, sequence_1, return_tensors=\"pt\")\r\n```\r\ncreates tensors of length `1310` and `1313`, which is too long for your model. You should enable the truncation parameter on your tokenizer to ensure that the length is correct:\r\n```py\r\nparaphrase = tokenizer(sequence_0, sequence_2, return_tensors=\"pt\", truncation=True)\r\nnot_paraphrase = tokenizer(sequence_0, sequence_1, return_tensors=\"pt\", truncation=True)\r\n```" ]

[]

[]

[ "@greenstars having the same issue - How did you resolve this?", "> @greenstars having the same issue - How did you resolve this?\r\n\r\n@EliaKunz I changed \"!pip install git+https://github.com/huggingface/transformers.git\" to \"!pip install transformers\". ", "Thx! Was on datalore with the latest transformers 4 - downgraded to 3.5 and everything is working now.", "I had the same issue with the latest transformers 4.1 (pip installed). It's fixed after adding return_tensors point. \r\n\r\nFrom\r\n\r\n`batch = tokenizer.prepare_seq2seq_batch(src_text, truncation=True, padding='longest').to(torch_device)`\r\n\r\nto\r\n\r\n`batch = tokenizer.prepare_seq2seq_batch(src_text, truncation=True, padding='longest', return_tensors='pt').to(torch_device)`\r\n\r\ndid the job for me.\r\n", "On running \r\n`batch = tokenizer(src_text, truncation=True, padding='longest', return_tensors=\"pt\").to(device)`\r\nI am getting the error\r\n```\r\n---------------------------------------------------------------------------\r\nTypeError Traceback (most recent call last)\r\n<ipython-input-49-e6e55e18a32c> in <module>()\r\n----> 1 batch = tokenizer(src_text, truncation=True, padding='longest', return_tensors=\"pt\").to(device)\r\n 2 translated = model.generate(**batch)\r\n 3 tgt_text = tokenizer.batch_decode(translated, skip_special_tokens=True)\r\n\r\nTypeError: 'NoneType' object is not callable\r\n```\r\nand on running `batch = tokenizer.prepare_seq2seq_batch(src_text, truncation=True, padding='longest', return_tensors='pt').to(torch_device)`\r\nI am getting the error\r\n```\r\n---------------------------------------------------------------------------\r\nAttributeError Traceback (most recent call last)\r\n<ipython-input-50-b7183fa2a37c> in <module>()\r\n----> 1 batch = tokenizer.prepare_seq2seq_batch(src_text, truncation=True, padding='longest', return_tensors='pt').to(torch_device)\r\n 2 translated = model.generate(**batch)\r\n 3 tgt_text = tokenizer.batch_decode(translated, skip_special_tokens=True)\r\n\r\nAttributeError: 'NoneType' object has no attribute 'prepare_seq2seq_batch'\r\n```\r\n\r\nAny help would be greatly appreciated. ", "@YatinKapoor your tokenizer seems to be `None`", "Need to replace PegasusTokenizer with AutoTokenizer\r\n```\r\nfrom transformers import PegasusForConditionalGeneration, AutoTokenizer\r\ntokenizer = AutoTokenizer.from_pretrained(model_name)\r\n````\r\n", "@maifeng thanks! AutoTokenizer did the job for me! ", "> I had the same issue with the latest transformers 4.1 (pip installed). It's fixed after adding return_tensors point.\r\n> \r\n> From\r\n> \r\n> `batch = tokenizer.prepare_seq2seq_batch(src_text, truncation=True, padding='longest').to(torch_device)`\r\n> \r\n> to\r\n> \r\n> `batch = tokenizer.prepare_seq2seq_batch(src_text, truncation=True, padding='longest', return_tensors='pt').to(torch_device)`\r\n> \r\n> did the job for me.\r\n\r\nWorked for me" ]

[]

[]

[ "Having a similar issue while running Multi class classification model", "@patrickvonplaten @sumyuck @sgugger ", "Hi\r\nI am constantly getting this erorr, looks like a bug to me since sometimes it appears sometimes not, could you please help me, this is expensive experiments I am trying on TPUs and I appreciate your help to fix it, it just many times fails due to this error\r\n\r\ngetting this erorr Exception in device=TPU:0: Connection error, and we cannot find the requested files in the cached path. Please try again or make sure your Internet connection is on.\r\nel/0 I1124 07:19:52.663760 424494 main shadow.py:87 > Traceback (most recent call last):\r\n File \"/root/anaconda3/envs/pytorch/lib/python3.6/site-packages/torch_xla/distributed/xla_multiprocessing.py\", line 330, in _mp_start_fn\r\n _start_fn(index, pf_cfg, fn, args)\r\n File \"/root/anaconda3/envs/pytorch/lib/python3.6/site-packages/torch_xla/distributed/xla_multiprocessing.py\", line 324, in _start_fn\r\n fn(gindex, *args)\r\n File \"/workdir/seq2seq/finetune_t5_trainer.py\", line 230, in _mp_fn\r\n main()\r\n File \"/workdir/seq2seq/finetune_t5_trainer.py\", line 71, in main\r\n cache_dir=model_args.cache_dir,\r\n File \"/root/anaconda3/envs/pytorch/lib/python3.6/site-packages/transformers/configuration_utils.py\", line 347, in from_pretrained\r\n config_dict, kwargs = cls.get_config_dict(pretrained_model_name_or_path, **kwargs)\r\n File \"/root/anaconda3/envs/pytorch/lib/python3.6/site-packages/transformers/configuration_utils.py\", line 388, in get_config_dict\r\n local_files_only=local_files_only,\r\n File \"/root/anaconda3/envs/pytorch/lib/python3.6/site-packages/transformers/file_utils.py\", line 955, in cached_path\r\n local_files_only=local_files_only,\r\n File \"/root/anaconda3/envs/pytorch/lib/python3.6/site-packages/transformers/file_utils.py\", line 1125, in get_from_cache\r\n \"Connection error, and we cannot find the requested files in the cached path.\"", "@sumyuck", "@thomwolf ", "this is with transformer 3.5.1, pytorch 1.6, on TPU v3-8, and I am using xla_spawn to launch the jobs, looks like a general issue with caching part. ", "Same for me. Getting this error while trying to execute following line:\r\n tokenizer = LxmertTokenizer.from_pretrained('unc-nlp/lxmert-base-uncased')\r\n\r\n File \"/Users/xxx/anaconda3/envs/test/lib/python3.7/site-packages/transformers/tokenization_utils_base.py\", line 1629, in from_pretrained\r\n local_files_only=local_files_only,\r\n File \"/Users/xxx/anaconda3/envs/test/lib/python3.7/site-packages/transformers/file_utils.py\", line 955, in cached_path\r\n local_files_only=local_files_only,\r\n File \"/Users/xxx/anaconda3/envs/test/lib/python3.7/site-packages/transformers/file_utils.py\", line 1125, in get_from_cache\r\n \"Connection error, and we cannot find the requested files in the cached path.\"\r\nValueError: Connection error, and we cannot find the requested files in the cached path. Please try again or make sure your Internet connection is on.\r\n", "to me this is not a connection issue. i do have connection but an issue in\ncaching mechanism.\n\nOn Wed, Nov 25, 2020, 2:33 AM Alkesh <[email protected]> wrote:\n\n> Same for me. Getting this error while trying to execute following line:\n> tokenizer = LxmertTokenizer.from_pretrained('unc-nlp/lxmert-base-uncased')\n>\n> File\n> \"/Users/xxx/anaconda3/envs/test/lib/python3.7/site-packages/transformers/tokenization_utils_base.py\",\n> line 1629, in from_pretrained\n> local_files_only=local_files_only,\n> File\n> \"/Users/xxx/anaconda3/envs/test/lib/python3.7/site-packages/transformers/file_utils.py\",\n> line 955, in cached_path\n> local_files_only=local_files_only,\n> File\n> \"/Users/xxx/anaconda3/envs/test/lib/python3.7/site-packages/transformers/file_utils.py\",\n> line 1125, in get_from_cache\n> \"Connection error, and we cannot find the requested files in the cached\n> path.\"\n> ValueError: Connection error, and we cannot find the requested files in\n> the cached path. Please try again or make sure your Internet connection is\n> on.\n>\n> —\n> You are receiving this because you authored the thread.\n> Reply to this email directly, view it on GitHub\n> <https://github.com/huggingface/transformers/issues/8690#issuecomment-733405868>,\n> or unsubscribe\n> <https://github.com/notifications/unsubscribe-auth/ABP4ZCGOHVMHGA33EGSQ6UTSRRNGTANCNFSM4T5CBSUA>\n> .\n>\n", "I am having the same issue too. I am pointing to the cache directory where pytorch is saving the models:\r\n`cache_dir = '/home/me/.cache/torch/transformers/' \r\n\r\nmodelpath = \"bert-base-uncased\" \r\n\r\nmodel = AutoModel.from_pretrained(modelpath, cache_dir=cache_dir) \r\n\r\ntokenizer = AutoTokenizer.from_pretrained(modelpath, cache_dir=cache_dir) \r\n`\r\nAnd I am getting a connection error. pytorch: 1.7.0, transformers: 3.5.1.", "Working on a fix, hopefully fixed for good today.\r\n\r\nMeanwhile as a workaround please retry a couple minutes later should do the trick", " I deleted all cache, redownloaded all modes and ran again. It seems to be working as of now. ", "Scaling of connectivity for model hosting should be way improved now. Please comment here if you still experience connectivity issues from now on.\r\n\r\nThanks!", "I am still getting this error with transformers version - 3.5.1 and torch - 1.7.0 on python 3.6.9. Please check. I have tried deleting all cache, installing transformers using pip and source code both. But still getting the same issue again and again.", "@AshishDuhan Are you loading a model in particular? Do you have a code snippet that consistently fails for you? ", "_import torch\r\nfrom transformers import PegasusForConditionalGeneration, PegasusTokenizer\r\n\r\nsrc_text = [\"\"\"<TEXT-HERE>\"\"\"]\r\nmodel_name='google/pegasus-cnn_dailymail'\r\ntorch_device='cuda' if torch.cuda.is_available() else 'cpu'\r\ntokenizer=PegasusTokenizer.from_pretrained(model_name)\r\nmodel=PegasusForConditionalGeneration.from_pretrained(model_name).to(torch_device)\r\nbatch=tokenizer.prepare_seq2seq_batch(src_text, truncation=True, padding='longest').to(torch_device)\r\ntranslated=model.generate(**batch)\r\ntgt_text=tokenizer.batch_decode(translated, skip_special_tokens=True)\r\nprint('Summary:', tgt_text[0])_\r\n\r\n\r\n**This is one of the models I am trying to load. Although I have tried other models too and nothing works. Even the basic command fail with following error:**\r\n\r\n**python -c \"from transformers import pipeline; print(pipeline('sentiment-analysis')('we love you'))\"**\r\nTraceback (most recent call last):\r\n File \"<string>\", line 1, in <module>\r\n File \"/opt/app/jupyter/environments/env_summarization/lib/python3.6/site-packages/transformers/pipelines.py\", line 2828, in pipeline\r\n framework = framework or get_framework(model)\r\n File \"/opt/app/jupyter/environments/env_summarization/lib/python3.6/site-packages/transformers/pipelines.py\", line 106, in get_framework\r\n model = AutoModel.from_pretrained(model, revision=revision)\r\n File \"/opt/app/jupyter/environments/env_summarization/lib/python3.6/site-packages/transformers/modeling_auto.py\", line 636, in from_pretrained\r\n pretrained_model_name_or_path, return_unused_kwargs=True, **kwargs\r\n File \"/opt/app/jupyter/environments/env_summarization/lib/python3.6/site-packages/transformers/configuration_auto.py\", line 333, in from_pretrained\r\n config_dict, _ = PretrainedConfig.get_config_dict(pretrained_model_name_or_path, **kwargs)\r\n File \"/opt/app/jupyter/environments/env_summarization/lib/python3.6/site-packages/transformers/configuration_utils.py\", line 388, in get_config_dict\r\n local_files_only=local_files_only,\r\n File \"/opt/app/jupyter/environments/env_summarization/lib/python3.6/site-packages/transformers/file_utils.py\", line 955, in cached_path\r\n local_files_only=local_files_only,\r\n File \"/opt/app/jupyter/environments/env_summarization/lib/python3.6/site-packages/transformers/file_utils.py\", line 1125, in get_from_cache\r\n \"Connection error, and we cannot find the requested files in the cached path.\"\r\nValueError: Connection error, and we cannot find the requested files in the cached path. Please try again or make sure your Internet connection is on.", "Our connectivity has been good these past 24 hours so this might be a different (local) issue, @AshishDuhan.\r\n\r\nAre you behind a proxy by any chance?\r\n\r\nDoes `curl -i https://huggingface.co/google/pegasus-cnn_dailymail/resolve/main/config.json` work from your machine? Can you try what you're doing from a machine in the cloud, like a Google Colab?", "I am facing the same issue still - \r\n\r\nTraceback (most recent call last):\r\n File \"Untitled.py\", line 59, in <module>\r\n tokenizer = AutoTokenizer.from_pretrained(\"emilyalsentzer/Bio_ClinicalBERT\")\r\n File \"/project/6001557/akallada/digipath/lib/python3.7/site-packages/transformers/models/auto/tokenization_auto.py\", line 310, in from_pretrained\r\n config = AutoConfig.from_pretrained(pretrained_model_name_or_path, **kwargs)\r\n File \"/project/6001557/akallada/digipath/lib/python3.7/site-packages/transformers/models/auto/configuration_auto.py\", line 341, in from_pretrained\r\n config_dict, _ = PretrainedConfig.get_config_dict(pretrained_model_name_or_path, **kwargs)\r\n File \"/project/6001557/akallada/digipath/lib/python3.7/site-packages/transformers/configuration_utils.py\", line 386, in get_config_dict\r\n local_files_only=local_files_only,\r\n File \"/project/6001557/akallada/digipath/lib/python3.7/site-packages/transformers/file_utils.py\", line 1007, in cached_path\r\n local_files_only=local_files_only,\r\n File \"/project/6001557/akallada/digipath/lib/python3.7/site-packages/transformers/file_utils.py\", line 1177, in get_from_cache\r\n \"Connection error, and we cannot find the requested files in the cached path.\"\r\nValueError: Connection error, and we cannot find the requested files in the cached path. Please try again or make sure your Internet connection is on.\r\n", "I'm having the same connection issue. I've tried with and without passing my proxies into the BertModel\r\n\r\n---------------------------------------------------------------------------\r\nValueError Traceback (most recent call last)\r\n<ipython-input-19-c8b8c602a810> in <module>\r\n 1 from transformers import BertTokenizer, BertModel\r\n----> 2 model = BertModel.from_pretrained(\"bert-base-uncased\", **proxies)\r\n\r\n~/opt/anaconda3/envs/milglue/lib/python3.8/site-packages/transformers/modeling_utils.py in from_pretrained(cls, pretrained_model_name_or_path, *model_args, **kwargs)\r\n 865 if not isinstance(config, PretrainedConfig):\r\n 866 config_path = config if config is not None else pretrained_model_name_or_path\r\n--> 867 config, model_kwargs = cls.config_class.from_pretrained(\r\n 868 config_path,\r\n 869 *model_args,\r\n\r\n~/opt/anaconda3/envs/milglue/lib/python3.8/site-packages/transformers/configuration_utils.py in from_pretrained(cls, pretrained_model_name_or_path, **kwargs)\r\n 345 \r\n 346 \"\"\"\r\n--> 347 config_dict, kwargs = cls.get_config_dict(pretrained_model_name_or_path, **kwargs)\r\n 348 return cls.from_dict(config_dict, **kwargs)\r\n 349 \r\n\r\n~/opt/anaconda3/envs/milglue/lib/python3.8/site-packages/transformers/configuration_utils.py in get_config_dict(cls, pretrained_model_name_or_path, **kwargs)\r\n 380 try:\r\n 381 # Load from URL or cache if already cached\r\n--> 382 resolved_config_file = cached_path(\r\n 383 config_file,\r\n 384 cache_dir=cache_dir,\r\n\r\n~/opt/anaconda3/envs/milglue/lib/python3.8/site-packages/transformers/file_utils.py in cached_path(url_or_filename, cache_dir, force_download, proxies, resume_download, user_agent, extract_compressed_file, force_extract, local_files_only)\r\n 946 if is_remote_url(url_or_filename):\r\n 947 # URL, so get it from the cache (downloading if necessary)\r\n--> 948 output_path = get_from_cache(\r\n 949 url_or_filename,\r\n 950 cache_dir=cache_dir,\r\n\r\n~/opt/anaconda3/envs/milglue/lib/python3.8/site-packages/transformers/file_utils.py in get_from_cache(url, cache_dir, force_download, proxies, etag_timeout, resume_download, user_agent, local_files_only)\r\n 1122 )\r\n 1123 else:\r\n-> 1124 raise ValueError(\r\n 1125 \"Connection error, and we cannot find the requested files in the cached path.\"\r\n 1126 \" Please try again or make sure your Internet connection is on.\"\r\n\r\nValueError: Connection error, and we cannot find the requested files in the cached path. Please try again or make sure your Internet connection is on.", "Hard to say without seeing your full networking environment.\r\n\r\nIf you try to `curl -I` the URLs that you get on the arrow icons next to files in e.g. https://huggingface.co/bert-base-uncased/tree/main (or equivalent page for the model you try to download), what happens?", "it happened to me too , is there any fix on that ? ", "is it transient or permanent (i.e. if you relaunch the command does it happen again)? You need to give us some more details if we want to help you troubleshoot.", "Hi\r\nI am still getting this issue. see blow. I am using transformer 3.5.1, could you tell me if the issue is fixed in this version? if not which version of transformers library I should use? thanks\r\n@julien-c \r\n\r\n```\r\n 12/13/2020 13:56:10 - INFO - seq2seq.utils.utils - config is reset to the initial values.\r\ntp/0 I1213 06:00:34.060680 252396 main shadow.py:122 > Traceback (most recent call last):\r\n File \"/usr/local/lib/python3.6/dist-packages/urllib3/connection.py\", line 170, in _new_conn\r\n (self._dns_host, self.port), self.timeout, **extra_kw\r\n File \"/usr/local/lib/python3.6/dist-packages/urllib3/util/connection.py\", line 96, in create_connection\r\n raise err\r\n File \"/usr/local/lib/python3.6/dist-packages/urllib3/util/connection.py\", line 86, in create_connection\r\n sock.connect(sa)\r\nsocket.timeout: timed out\r\ntp/0 I1213 06:00:34.060720 252396 main shadow.py:122 > \r\ntp/0 I1213 06:00:34.060759 252396 main shadow.py:122 > During handling of the above exception, another exception occurred:\r\ntp/0 I1213 06:00:34.060825 252396 main shadow.py:122 > \r\ntp/0 I1213 06:00:34.060866 252396 main shadow.py:122 > Traceback (most recent call last):\r\n File \"/usr/local/lib/python3.6/dist-packages/urllib3/connectionpool.py\", line 706, in urlopen\r\n chunked=chunked,\r\n File \"/usr/local/lib/python3.6/dist-packages/urllib3/connectionpool.py\", line 382, in _make_request\r\n self._validate_conn(conn)\r\n File \"/usr/local/lib/python3.6/dist-packages/urllib3/connectionpool.py\", line 1010, in _validate_conn\r\n conn.connect()\r\n File \"/usr/local/lib/python3.6/dist-packages/urllib3/connection.py\", line 353, in connect\r\n conn = self._new_conn()\r\n File \"/usr/local/lib/python3.6/dist-packages/urllib3/connection.py\", line 177, in _new_conn\r\n % (self.host, self.timeout),\r\nurllib3.exceptions.ConnectTimeoutError: (<urllib3.connection.HTTPSConnection object at 0x7f47db511e80>, 'Connection to s3.amazonaws.com timed out. (connect timeout=10)')\r\ntp/0 I1213 06:00:34.060908 252396 main shadow.py:122 > \r\ntp/0 I1213 06:00:34.060970 252396 main shadow.py:122 > During handling of the above exception, another exception occurred:\r\ntp/0 I1213 06:00:34.061113 252396 main shadow.py:122 > \r\ntp/0 I1213 06:00:34.061207 252396 main shadow.py:122 > Traceback (most recent call last):\r\n File \"/usr/local/lib/python3.6/dist-packages/requests/adapters.py\", line 449, in send\r\n timeout=timeout\r\n File \"/usr/local/lib/python3.6/dist-packages/urllib3/connectionpool.py\", line 756, in urlopen\r\n method, url, error=e, _pool=self, _stacktrace=sys.exc_info()[2]\r\n File \"/usr/local/lib/python3.6/dist-packages/urllib3/util/retry.py\", line 573, in increment\r\n raise MaxRetryError(_pool, url, error or ResponseError(cause))\r\nurllib3.exceptions.MaxRetryError: HTTPSConnectionPool(host='s3.amazonaws.com', port=443): Max retries exceeded with url: /datasets.huggingface.co/datasets/datasets/glue/glue.py (Caused by ConnectTimeoutError(<urllib3.connection.HTTPSConnection object at 0x7f47db511e80>, 'Connection to s3.amazonaws.com timed out. (connect timeout=10)'))\r\ntp/0 I1213 06:00:34.061293 252396 main shadow.py:122 > \r\ntp/0 I1213 06:00:34.061372 252396 main shadow.py:122 > During handling of the above exception, another exception occurred:\r\ntp/0 I1213 06:00:34.061421 252396 main shadow.py:122 > \r\ntp/0 I1213 06:00:34.061486 252396 main shadow.py:122 > Traceback (most recent call last):\r\n File \"finetune_t5_trainer.py\", line 361, in <module>\r\n main()\r\n File \"finetune_t5_trainer.py\", line 269, in main\r\n add_prefix=False if training_args.train_adapters else True)\r\n File \"/workdir/seq2seq/data/tasks.py\", line 70, in get_dataset\r\n dataset = self.load_dataset(split=split)\r\n File \"/workdir/seq2seq/data/tasks.py\", line 306, in load_dataset\r\n return datasets.load_dataset('glue', 'cola', split=split)\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/load.py\", line 589, in load_dataset\r\n path, script_version=script_version, download_config=download_config, download_mode=download_mode, dataset=True\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/load.py\", line 263, in prepare_module\r\n head_hf_s3(path, filename=name, dataset=dataset)\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/utils/file_utils.py\", line 200, in head_hf_s3\r\n return http_head(hf_bucket_url(identifier=identifier, filename=filename, use_cdn=use_cdn, dataset=dataset))\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/utils/file_utils.py\", line 403, in http_head\r\n url, proxies=proxies, headers=headers, cookies=cookies, allow_redirects=allow_redirects, timeout=timeout\r\n File \"/usr/local/lib/python3.6/dist-packages/requests/api.py\", line 104, in head\r\n return request('head', url, **kwargs)\r\n File \"/usr/local/lib/python3.6/dist-packages/requests/api.py\", line 61, in request\r\n return session.request(method=method, url=url, **kwargs)\r\n File \"/usr/local/lib/python3.6/dist-packages/requests/sessions.py\", line 542, in request\r\n resp = self.send(prep, **send_kwargs)\r\n File \"/usr/local/lib/python3.6/dist-packages/requests/sessions.py\", line 655, in send\r\n r = adapter.send(request, **kwargs)\r\n File \"/usr/local/lib/python3.6/dist-packages/requests/adapters.py\", line 504, in send\r\n raise ConnectTimeout(e, request=request)\r\nrequests.exceptions.ConnectTimeout: HTTPSConnectionPool(host='s3.amazonaws.com', port=443): Max retries exceeded with url: /datasets.huggingface.co/datasets/datasets/glue/glue.py (Caused by ConnectTimeoutError(<urllib3.connection.HTTPSConnection object at 0x7f47db511e80>, 'Connection to s3.amazonaws.com timed out. (connect timeout=10)'))\r\ntp/0 I1213 06:00:35.237288 252396 main waiter_thread.cc:2652 [tp][0] EndSession for client id 1607864609277665002 (server tpe18:6297)\r\n```", "Looks like you are getting a timeout connecting to `s3.amazonaws.com`. There's not much we can do here.", "Hi,\r\nI am facing the same issue, the code is running fine on colab but while running it on local system i am getting below error.\r\n\r\nfrom transformers import AutoTokenizer, AutoModelForMaskedLM\r\n\r\ntokenizer = AutoTokenizer.from_pretrained(\"bert-base-cased\")\r\n\r\nmodel = AutoModelForMaskedLM.from_pretrained(\"bert-base-cased\")\r\n\r\n---------------------------------------------------------------------------\r\nValueError Traceback (most recent call last)\r\n<ipython-input-9-4dd822b7db9b> in <module>\r\n 1 from transformers import AutoTokenizer, AutoModelForMaskedLM\r\n 2 \r\n----> 3 tokenizer = AutoTokenizer.from_pretrained(\"bert-base-cased\")\r\n 4 \r\n 5 model = AutoModelForMaskedLM.from_pretrained(\"bert-base-cased\")\r\n\r\n~\\Anaconda3\\envs\\bert-test\\lib\\site-packages\\transformers\\models\\auto\\tokenization_auto.py in from_pretrained(cls, pretrained_model_name_or_path, *inputs, **kwargs)\r\n 308 config = kwargs.pop(\"config\", None)\r\n 309 if not isinstance(config, PretrainedConfig):\r\n--> 310 config = AutoConfig.from_pretrained(pretrained_model_name_or_path, **kwargs)\r\n 311 \r\n 312 if \"bert-base-japanese\" in str(pretrained_model_name_or_path):\r\n\r\n~\\Anaconda3\\envs\\bert-test\\lib\\site-packages\\transformers\\models\\auto\\configuration_auto.py in from_pretrained(cls, pretrained_model_name_or_path, **kwargs)\r\n 339 {'foo': False}\r\n 340 \"\"\"\r\n--> 341 config_dict, _ = PretrainedConfig.get_config_dict(pretrained_model_name_or_path, **kwargs)\r\n 342 \r\n 343 if \"model_type\" in config_dict:\r\n\r\n~\\Anaconda3\\envs\\bert-test\\lib\\site-packages\\transformers\\configuration_utils.py in get_config_dict(cls, pretrained_model_name_or_path, **kwargs)\r\n 384 proxies=proxies,\r\n 385 resume_download=resume_download,\r\n--> 386 local_files_only=local_files_only,\r\n 387 )\r\n 388 # Load config dict\r\n\r\n~\\Anaconda3\\envs\\bert-test\\lib\\site-packages\\transformers\\file_utils.py in cached_path(url_or_filename, cache_dir, force_download, proxies, resume_download, user_agent, extract_compressed_file, force_extract, local_files_only)\r\n 1005 resume_download=resume_download,\r\n 1006 user_agent=user_agent,\r\n-> 1007 local_files_only=local_files_only,\r\n 1008 )\r\n 1009 elif os.path.exists(url_or_filename):\r\n\r\n~\\Anaconda3\\envs\\bert-test\\lib\\site-packages\\transformers\\file_utils.py in get_from_cache(url, cache_dir, force_download, proxies, etag_timeout, resume_download, user_agent, local_files_only)\r\n 1175 else:\r\n 1176 raise ValueError(\r\n-> 1177 \"Connection error, and we cannot find the requested files in the cached path.\"\r\n 1178 \" Please try again or make sure your Internet connection is on.\"\r\n 1179 )\r\n\r\nValueError: Connection error, and we cannot find the requested files in the cached path. Please try again or make sure your Internet connection is on.", "Can you try the debugging procedure mentioned in https://github.com/huggingface/transformers/issues/8690#issuecomment-737246999?", "i am able to open 8690 in web browser. but the error still remains:\r\n\r\nqa = text.SimpleQA(INDEXDIR)\r\n\r\n---------------------------------------------------------------------------\r\nValueError Traceback (most recent call last)\r\n~\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\ktrain\\text\\qa\\core.py in __init__(self, bert_squad_model, bert_emb_model)\r\n 67 try:\r\n---> 68 self.model = TFAutoModelForQuestionAnswering.from_pretrained(self.model_name)\r\n 69 except:\r\n\r\n~\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\transformers\\modeling_tf_auto.py in from_pretrained(cls, pretrained_model_name_or_path, *model_args, **kwargs)\r\n 1204 config, kwargs = AutoConfig.from_pretrained(\r\n-> 1205 pretrained_model_name_or_path, return_unused_kwargs=True, **kwargs\r\n 1206 )\r\n\r\n~\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\transformers\\configuration_auto.py in from_pretrained(cls, pretrained_model_name_or_path, **kwargs)\r\n 332 \"\"\"\r\n--> 333 config_dict, _ = PretrainedConfig.get_config_dict(pretrained_model_name_or_path, **kwargs)\r\n 334 \r\n\r\n~\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\transformers\\configuration_utils.py in get_config_dict(cls, pretrained_model_name_or_path, **kwargs)\r\n 387 resume_download=resume_download,\r\n--> 388 local_files_only=local_files_only,\r\n 389 )\r\n\r\n~\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\transformers\\file_utils.py in cached_path(url_or_filename, cache_dir, force_download, proxies, resume_download, user_agent, extract_compressed_file, force_extract, local_files_only)\r\n 954 user_agent=user_agent,\r\n--> 955 local_files_only=local_files_only,\r\n 956 )\r\n\r\n~\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\transformers\\file_utils.py in get_from_cache(url, cache_dir, force_download, proxies, etag_timeout, resume_download, user_agent, local_files_only)\r\n 1124 raise ValueError(\r\n-> 1125 \"Connection error, and we cannot find the requested files in the cached path.\"\r\n 1126 \" Please try again or make sure your Internet connection is on.\"\r\n\r\nValueError: Connection error, and we cannot find the requested files in the cached path. Please try again or make sure your Internet connection is on.\r\n\r\nDuring handling of the above exception, another exception occurred:\r\n\r\nValueError Traceback (most recent call last)\r\n<ipython-input-72-18505d037255> in <module>\r\n 1 # ask questions (setting higher batch size can further speed up answer retrieval)\r\n----> 2 qa = text.SimpleQA(INDEXDIR)\r\n 3 #answers = qa.ask('What is lotus sutra?', batch_size=8)\r\n\r\n~\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\ktrain\\text\\qa\\core.py in __init__(self, index_dir, bert_squad_model, bert_emb_model)\r\n 348 except:\r\n 349 raise ValueError('index_dir has not yet been created - please call SimpleQA.initialize_index(\"%s\")' % (self.index_dir))\r\n--> 350 super().__init__(bert_squad_model=bert_squad_model, bert_emb_model=bert_emb_model)\r\n 351 \r\n 352 \r\n\r\n~\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\ktrain\\text\\qa\\core.py in __init__(self, bert_squad_model, bert_emb_model)\r\n 68 self.model = TFAutoModelForQuestionAnswering.from_pretrained(self.model_name)\r\n 69 except:\r\n---> 70 self.model = TFAutoModelForQuestionAnswering.from_pretrained(self.model_name, from_pt=True)\r\n 71 self.tokenizer = AutoTokenizer.from_pretrained(self.model_name)\r\n 72 self.maxlen = 512\r\n\r\n~\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\transformers\\modeling_tf_auto.py in from_pretrained(cls, pretrained_model_name_or_path, *model_args, **kwargs)\r\n 1203 if not isinstance(config, PretrainedConfig):\r\n 1204 config, kwargs = AutoConfig.from_pretrained(\r\n-> 1205 pretrained_model_name_or_path, return_unused_kwargs=True, **kwargs\r\n 1206 )\r\n 1207 \r\n\r\n~\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\transformers\\configuration_auto.py in from_pretrained(cls, pretrained_model_name_or_path, **kwargs)\r\n 331 {'foo': False}\r\n 332 \"\"\"\r\n--> 333 config_dict, _ = PretrainedConfig.get_config_dict(pretrained_model_name_or_path, **kwargs)\r\n 334 \r\n 335 if \"model_type\" in config_dict:\r\n\r\n~\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\transformers\\configuration_utils.py in get_config_dict(cls, pretrained_model_name_or_path, **kwargs)\r\n 386 proxies=proxies,\r\n 387 resume_download=resume_download,\r\n--> 388 local_files_only=local_files_only,\r\n 389 )\r\n 390 # Load config dict\r\n\r\n~\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\transformers\\file_utils.py in cached_path(url_or_filename, cache_dir, force_download, proxies, resume_download, user_agent, extract_compressed_file, force_extract, local_files_only)\r\n 953 resume_download=resume_download,\r\n 954 user_agent=user_agent,\r\n--> 955 local_files_only=local_files_only,\r\n 956 )\r\n 957 elif os.path.exists(url_or_filename):\r\n\r\n~\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\transformers\\file_utils.py in get_from_cache(url, cache_dir, force_download, proxies, etag_timeout, resume_download, user_agent, local_files_only)\r\n 1123 else:\r\n 1124 raise ValueError(\r\n-> 1125 \"Connection error, and we cannot find the requested files in the cached path.\"\r\n 1126 \" Please try again or make sure your Internet connection is on.\"\r\n 1127 )\r\n\r\nValueError: Connection error, and we cannot find the requested files in the cached path. Please try again or make sure your Internet connection is on.\r\n", "still get this error for transformer 4.1.1 with torch 1.7.1\r\n\r\nerror message here:\r\n```\r\nTraceback (most recent call last):\r\n File \"run_distributed_eval.py\", line 273, in <module>\r\n run_generate()\r\n File \"run_distributed_eval.py\", line 206, in run_generate\r\n **generate_kwargs,\r\n File \"run_distributed_eval.py\", line 88, in eval_data_dir\r\n tokenizer = AutoTokenizer.from_pretrained(model_name)\r\n File \"/data/User/v5/acl/venv/lib/python3.6/site-packages/transformers/models/auto/tokenization_auto.py\", line 378, in from_pretrained\r\n return tokenizer_class_fast.from_pretrained(pretrained_model_name_or_path, *inputs, **kwargs)\r\n File \"/data/User/v5/acl/venv/lib/python3.6/site-packages/transformers/tokenization_utils_base.py\", line 1781, in from_pretrained\r\n use_auth_token=use_auth_token,\r\n File \"/data/User/v5/acl/venv/lib/python3.6/site-packages/transformers/file_utils.py\", line 1085, in cached_path\r\n local_files_only=local_files_only,\r\n File \"/data/User/v5/acl/venv/lib/python3.6/site-packages/transformers/file_utils.py\", line 1264, in get_from_cache\r\n \"Connection error, and we cannot find the requested files in the cached path.\"\r\nValueError: Connection error, and we cannot find the requested files in the cached path. Please try again or make sure your Internet connection is on.\r\n```", "try transformers 4.00 transformers:4.1 \r\nSame error\r\n\r\n#8690 (comment)\r\nThis can be accessed and downloaded\r\n```\r\nTraceback (most recent call last):\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\runpy.py\", line 194, in _run_module_as_main\r\n return _run_code(code, main_globals, None,\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\runpy.py\", line 87, in _run_code\r\n exec(code, run_globals)\r\n File \"F:\\Software\\Anaconda\\envs\\py38\\Scripts\\rasa.exe\\__main__.py\", line 7, in <module>\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\__main__.py\", line 116, in main\r\n cmdline_arguments.func(cmdline_arguments)\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\cli\\train.py\", line 58, in <lambda>\r\n train_parser.set_defaults(func=lambda args: train(args, can_exit=True))\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\cli\\train.py\", line 90, in train\r\n training_result = rasa.train(\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\train.py\", line 94, in train\r\n return rasa.utils.common.run_in_loop(\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\utils\\common.py\", line 308, in run_in_loop\r\n result = loop.run_until_complete(f)\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\asyncio\\base_events.py\", line 616, in run_until_complete\r\n return future.result()\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\train.py\", line 163, in train_async\r\n return await _train_async_internal(\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\train.py\", line 342, in _train_async_internal\r\n await _do_training(\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\train.py\", line 388, in _do_training\r\n model_path = await _train_nlu_with_validated_data(\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\train.py\", line 811, in _train_nlu_with_validated_data\r\n await rasa.nlu.train(\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\nlu\\train.py\", line 97, in train\r\n trainer = Trainer(\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\nlu\\model.py\", line 163, in __init__\r\n self.pipeline = self._build_pipeline(cfg, component_builder)\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\nlu\\model.py\", line 174, in _build_pipeline\r\n component = component_builder.create_component(component_cfg, cfg)\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\nlu\\components.py\", line 852, in create_component\r\n component = registry.create_component_by_config(component_config, cfg)\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\nlu\\registry.py\", line 193, in create_component_by_config\r\n return component_class.create(component_config, config)\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\nlu\\components.py\", line 525, in create\r\n return cls(component_config)\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\nlu\\utils\\hugging_face\\hf_transformers.py\", line 65, in __init__\r\n self._load_model_instance(skip_model_load)\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\rasa\\nlu\\utils\\hugging_face\\hf_transformers.py\", line 121, in _load_model_instance\r\n self.tokenizer = model_tokenizer_dict[self.model_name].from_pretrained(\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\transformers\\tokenization_utils_base.py\", line 1774, in from_pretrained\r\n resolved_vocab_files[file_id] = cached_path(\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\transformers\\file_utils.py\", line 1077, in cached_path\r\n output_path = get_from_cache(\r\n File \"f:\\software\\anaconda\\envs\\py38\\lib\\site-packages\\transformers\\file_utils.py\", line 1263, in get_from_cache\r\n raise ValueError(\r\nValueError: Connection error, and we cannot find the requested files in the cached path. Please try again or make sure your Internet connection is on\r\n```", "I also ran into this error while trying to download any huggingface model. Turns out for me the cause was that I had set an `export REQUESTS_CA_BUNDLE=path/to/some/certificate` in my .bash_profile, which I needed to get some poetry stuff working. Once I removed this line and restarted, the download was working again.", "It appears to be an SSL/TLS certificate error as @robinderat alludes to, but there are several possible reasons. Here's how I've debugged this, hopefully it helps others although your root cause may be different.\r\n\r\n## Debugging\r\n\r\nOriginal error, fetching model from `https://huggingface.co/elastic/distilbert-base-cased-finetuned-conll03-english`:\r\n\r\n```\r\nValueError: Connection error, and we cannot find the requested files in the cached path. Please try again or make sure your Internet connection is on.\r\n```\r\n\r\nCheck with `curl`:\r\n\r\n```\r\n$ curl -I https://huggingface.co/elastic/distilbert-base-cased-finetuned-conll03-english/resolve/main/config.json\r\ncurl: (60) SSL certificate problem: certificate is not yet valid\r\nMore details here: https://curl.haxx.se/docs/sslcerts.html\r\n\r\ncurl failed to verify the legitimacy of the server and therefore could not\r\nestablish a secure connection to it. To learn more about this situation and\r\nhow to fix it, please visit the web page mentioned above.\r\n```\r\n\r\nChecking with `requests`:\r\n\r\n```\r\n$ python -c \"import requests; requests.get('https://huggingface.co/elastic/distilbert-base-cased-finetuned-conll03-english/resolve/main/config.json')\"\r\nTraceback (most recent call last):\r\n <snip>\r\n File \"/usr/lib/python3.7/ssl.py\", line 412, in wrap_socket\r\n session=session\r\n File \"/usr/lib/python3.7/ssl.py\", line 853, in _create\r\n self.do_handshake()\r\n File \"/usr/lib/python3.7/ssl.py\", line 1117, in do_handshake\r\n self._sslobj.do_handshake()\r\nssl.SSLCertVerificationError: [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: certificate is not yet valid (_ssl.c:1056)\r\n```\r\n\r\nDisabling curl's certificate validation with `-k` flag works:\r\n\r\n```\r\n$ curl -k -I https://huggingface.co/elastic/distilbert-base-cased-finetuned-conll03-english/resolve/main/config.json\r\nHTTP/1.1 200 OK\r\n```\r\n\r\nAnd now in Python, using `verify=False`:\r\n\r\n```\r\n$ python -c \"import requests; r = requests.get('https://huggingface.co/elastic/distilbert-base-cased-finetuned-conll03-english/resolve/main/config.json', verify=False); print(r)\"\r\n/home/josh/source/examples/Machine Learning/Query Optimization/venv/lib/python3.7/site-packages/urllib3/connectionpool.py:1020: InsecureRequestWarning: Unverified HTTPS request is being made to host 'huggingface.co'. Adding certificate verification is strongly advised. See: https://urllib3.readthedocs.io/en/latest/advanced-usage.html#ssl-warnings\r\n InsecureRequestWarning,\r\n<Response [200]>\r\n```\r\n\r\n## Resolution\r\n\r\nSo the \"problem\" is in the certificate. Checking in a browser, the root certificate of `huggingface.co` expires 30 April, 2021 but is valid only from 30 January, 2020.\r\n\r\nChecking my server clock shows that it was out of date (27 January 20201) and critically, *before* the certificate is valid *from*, which makes sense that the root error was \"certificate verify failed: certificate is not yet valid\".\r\n\r\nSet the clock to the real time and check again:\r\n\r\n```\r\n$ sudo date -s \"Feb 11 09:34:03 UTC 2021\"\r\n$ python -c \"import requests; r = requests.get('https://huggingface.co/elastic/distilbert-base-cased-finetuned-conll03-english/resolve/main/config.json'); print(r)\"\r\n<Response [200]>\r\n```\r\n\r\nI now suspect that this host in GCP, which was suspended for a while, did not automatically update it's local time causing this specific problem.\r\n\r\n## Conclusion\r\n\r\n@julien-c I would only suggest at this point that making the root cause visible in the error coming out of `transformers` would be really helpful to more immediately see the problem.\r\n\r\n🎉 " ]

[]

[]

[ "No problem at all!\r\n\r\n+ The inheritance is just for the purpose of not duplicating code.\r\n+ You can change to whatever unk you would like, i wasn't at all careful about this stuff since i didn't try to replicate/test pre-training, just fine-tuning and generation.\r\n+ Your changes sound like obvious low risk improvements. \r\n+ I don't know whether standard mask-filling will work for integration testing purposes, given the seq2seq pre-training objective ." ]

[]

[]

[]

[ { "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 } ]

[]

[ "Not an expert in summarization, but abstractive text summarization does not extract sequences/tokens from the initial text to produce a summary. That would be extractive text summarization. Abstractive text summarization instead can be done with rephrasing, as it seems to be the case here.\r\n\r\nOn a second note, I believe the Pegasus checkpoints were trained on very long sequences, so I'm not entirely sure how it would deal with smaller sequences as the one you used here.\r\n\r\nOn a third note, we try to keep the github issues reserved for issues/feature requests; you would have more luck asking this over on the [forum](https://discuss.huggingface.co).\r\n\r\n@patrickvonplaten or @patil-suraj can chime in if I'm wrong.", "The hyperparameters seem very extreme to me... also `temperature=1` does not do anything and `length_penalty=5` is very high - also note that a length_penalty > 1 actually incentivizes longer sequences. @sshleifer 's model already has good hyper-parameters set as default values that you can see here:\r\nhttps://huggingface.co/sshleifer/distill-pegasus-xsum-16-8/blob/main/config.json\r\n\r\nIf you just use those, *e.g.*:\r\n```python\r\ntranslated = model_pegasus_distill_xsum_16_8.generate(**batch)\r\n```\r\n\r\nyou get this summary:\r\n```\r\nEuropean shares fell sharply on Wednesday as investors remained cautious ahead of a speech by France's president later in the day.\r\n```\r\n\r\nYou can try it yourself here:\r\nhttps://huggingface.co/sshleifer/distill-pegasus-xsum-16-8?text=German+shares+suffered+their+weakest+day+since+early+June+on+Wednesday+as+the+government+agreed+on+an+emergency+lockdown+to+combat+surging+COVID-19+cases%2C+with+other+European+markets+following+suit+on+fears+of+more+curbs+around+the+continent.+The+German+DAX+sank+as+much+as+5%25+before+cutting+some+losses+to+close+down+4.2%25+at+its+lowest+in+five+months.+The+precise+measures+were+still+subject+to+negotiation%2C+with+sources+saying+the+government+had+agreed+to+shut+bars+and+restaurants+from+Nov.+2.+The+pan-European+STOXX+600+index+fell+3%25+in+its+sharpest+one-day+drop+in+five+weeks.+France%27s+main+index+dropped+3.4%25+ahead+of+a+televised+address+by+President+Emmanuel+Macron+at+8%3A00+pm+when+he+is+expected+to+issue+stay-at-home+orders.\r\n```\r\n\r\nMy conclusion would be that it's just the hyperparameters that are badly chosen - not sure if @sshleifer has something to add...", "- Lysandre is correct about abstractive vs. extractive.\r\n- Hallucination is a known issue with Neural Text Generation. It will happen more often if you generate summaries that are more than ~30% the length of the input document (which your length_penalty and max_length encourage).\r\n- `\"sshleifer/distill-pegasus-xsum-16-4\"` is better and faster. See Table 6 of the [best paper in AI history](https://arxiv.org/pdf/2010.13002.pdf) ;). \r\n- I would set `num_beams=4` if I cared at all about speed.", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

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

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

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

[]

[ "Hi! TorchScript requires tracing the model beforehand, which slows down the first forward pass through the model. Could you print the timing of the iterations following the initial one?", "Hi, @LysandreJik \r\ndo you mean in the first iteration \"loaded_model(input_ids)\" will slow?\r\nI already traced model before that.\r\n\r\ntraced_model = torch.jit.trace(model, input_ids)\r\ntorch.jit.save(traced_model, 'trace_gpt2.pt')\r\n\r\nloaded_model = torch.jit.load('trace_gpt2.pt').to('cuda')\r\nloaded_model.eval()\r\n\r\n#print (loaded_model)\r\nstart = time.time()\r\nfor i in range(100):\r\n with torch.no_grad():\r\n loaded_model(input_ids)\r\nend = time.time()\r\nprint ('traced model',(end-start))\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", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

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

[]

[]

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

[]

[ "Can you please add metadata as in https://huggingface.co/docs#what-metadata-can-i-add-to-my-model-card?\r\n\r\nThank you!", "Closing this one as duplicate was already merged!\r\n\r\nFor context please also read https://discuss.huggingface.co/t/announcement-all-model-cards-will-be-migrated-to-hf-co-model-repos/2755" ]

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

[]

[ "`attention_mask` is an optional argument, but that doesn't mean that it should not be passed to the function. If `attention_mask` is `None` then it is initialized to attend all tokens (all 1's in the `attention_mask` tensor), which is incorrect if the input is a batch that includes padding tokens. => It's better to simply pass the `attention_mask` to the forward function.", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

[ { "id": 2627272588, "node_id": "MDU6TGFiZWwyNjI3MjcyNTg4", "url": "https://api.github.com/repos/huggingface/transformers/labels/Model%20Parallel", "name": "Model Parallel", "color": "8B66A5", "default": false, "description": "Model Parallelilsm Implementations" } ]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[ "Hi! Could you run `make style` on your branch so that the code quality check passes? Thanks!", "Hi guys, I only have mobile phone until Dec. 1. I will do it as soon as I can access PC.", "@lhoestq @LysandreJik done applying style. Sorry for late!" ]

[]

[]

[ "@mfuntowicz - could you take a look maybe? :-) ", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

[]

[]

[ "It looks like your using causual language modeling (next word prediction) - _File \"run_clm.py\", line 300, in main_, but xlnet does not use that. It uses permutation language modeling.\r\n\r\nHave you tried with the [new scripts](https://github.com/huggingface/transformers/tree/master/examples/language-modeling#xlnet-and-permutation-language-modeling), I believe this will solve it 👍 ", "Thank you, Tim. So I know XLNET and Transformer-XL are fine-tuned the same\r\nway and tried the *run_plm.py* like you suggested and got the error below:\r\n\r\n[INFO|tokenization_utils_base.py:1650] 2020-11-20 23:00:33,440 >> loading\r\nfile https://huggingface.co/transfo-xl-wt103/resolve/main/vocab.pkl from\r\ncache at\r\n/root/.cache/torch/transformers/6860d92833eb9d2a42cf185e974ca967fbf4cd58fa8d3d9298e56b9ef7ff8d5c.56c8ef92e693414ef2313bde4ba3679a404de1edbcd5a5780def3971f9706850\r\n[INFO|modeling_utils.py:940] 2020-11-20 23:00:34,096 >> loading weights\r\nfile https://huggingface.co/transfo-xl-wt103/resolve/main/pytorch_model.bin\r\nfrom cache at\r\n/root/.cache/torch/transformers/891af5f0c8372327a961a768d4ee40b7ca95c428f9384c534e73b9b655c75468.923bd8e0844a782c35f009eddd08a3600739804fbe13bd234f592f36230ab8a9\r\nTraceback (most recent call last): File \"run_plm.py\", line 382, in <module>\r\nmain() File \"run_plm.py\", line 244, in main cache_dir=model_args.cache_dir,\r\nFile\r\n\"/usr/local/lib/python3.6/dist-packages/transformers/modeling_utils.py\",\r\nline 947, in from_pretrained model = cls(config, *model_args,\r\n**model_kwargs) File\r\n\"/usr/local/lib/python3.6/dist-packages/transformers/modeling_xlnet.py\",\r\nline 1294, in __init__ self.transformer = XLNetModel(config) File\r\n\"/usr/local/lib/python3.6/dist-packages/transformers/modeling_xlnet.py\",\r\nline 940, in __init__ self.reuse_len = config.reuse_len AttributeError:\r\n'TransfoXLConfig' object has no attribute 'reuse_len'\r\n\r\nOn Fri, Nov 20, 2020 at 1:52 PM Tim Isbister <[email protected]>\r\nwrote:\r\n\r\n> It looks like your using causual language modeling (next word prediction)\r\n> - *run_clm.py*, but xlnet does not use that. It uses permutation language\r\n> modeling.\r\n>\r\n> Have you tried with the new scripts\r\n> <https://github.com/huggingface/transformers/tree/master/examples/language-modeling#xlnet-and-permutation-language-modeling>,\r\n> I believe this will solve it 👍\r\n>\r\n> —\r\n> You are receiving this because you authored the thread.\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/8674#issuecomment-731348230>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AHYMJAF5X3H4GW7WFSAQEOLSQ23F3ANCNFSM4T4CBVKQ>\r\n> .\r\n>\r\n\r\n", "Hmm strange, well actually I now tried running the provided example that I suggested to you. But also failing with `IndexError: index out of bounds` as you got in the first attempt. \r\n\r\nEdit: Looks like we have problems with loading the data, if I hardcoded a dataset to the `run_plm.py` \r\n`datasets = load_dataset('wikitext', 'wikitext-103-raw-v1’)` it works. \r\n\r\n[Provided example](https://github.com/huggingface/transformers/tree/master/examples/language-modeling#xlnet-and-permutation-language-modeling): \r\n\r\n```\r\npython run_plm.py \\\r\n --model_name_or_path=xlnet-base-cased \\\r\n --dataset_name wikitext \\\r\n --dataset_config_name wikitext-2-raw-v1 \\\r\n --do_train \\\r\n --do_eval \\\r\n --output_dir /tmp/test-plm\r\n```\r\n\r\n```\r\nTraceback (most recent call last):\r\n File \"run_plm.py\", line 382, in <module>\r\n main()\r\n File \"run_plm.py\", line 321, in main\r\n tokenized_datasets = tokenized_datasets.map(\r\n File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/dataset_dict.py\", line 286, in map\r\n {\r\n File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/dataset_dict.py\", line 287, in <dictcomp>\r\n k: dataset.map(\r\n File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/arrow_dataset.py\", line 1243, in map\r\n return self._map_single(\r\n File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/arrow_dataset.py\", line 157, in wrapper\r\n out: Union[\"Dataset\", \"DatasetDict\"] = func(self, *args, **kwargs)\r\n File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/fingerprint.py\", line 163, in wrapper\r\n out = func(self, *args, **kwargs)\r\n File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/arrow_dataset.py\", line 1528, in _map_single\r\n writer.write_batch(batch)\r\n File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/arrow_writer.py\", line 278, in write_batch\r\n pa_table = pa.Table.from_pydict(typed_sequence_examples)\r\n File \"pyarrow/table.pxi\", line 1474, in pyarrow.lib.Table.from_pydict\r\n File \"pyarrow/array.pxi\", line 322, in pyarrow.lib.asarray\r\n File \"pyarrow/array.pxi\", line 222, in pyarrow.lib.array\r\n File \"pyarrow/array.pxi\", line 110, in pyarrow.lib._handle_arrow_array_protocol\r\n File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/arrow_writer.py\", line 100, in __arrow_array__\r\n if trying_type and out[0].as_py() != self.data[0]:\r\n File \"pyarrow/array.pxi\", line 1058, in pyarrow.lib.Array.__getitem__\r\n File \"pyarrow/array.pxi\", line 540, in pyarrow.lib._normalize_index\r\nIndexError: index out of bounds\r\n```", "Thanks, Tim. Do you have any suggestions for how I can load my own train\r\nand validation datasets? And how will it work with the following code\r\nbelow. I am now using the current transformer version.\r\nThanks again, Tim.\r\n\r\n\r\npython run_plm.py \\\r\n --model_name_or_path=transfo-xl-wt103 \\\r\n --train_file='/content/drive/My Drive/finetuned_models/train.txt' \\\r\n --validation_file='/content/drive/My Drive/finetuned_models/valid.txt' \\\r\n --save_total_limit=5 \\\r\n --num_train_epochs=1.0 \\\r\n --do_train \\\r\n --do_eval \\\r\n --per_gpu_train_batch_size=2 \\\r\n --per_gpu_eval_batch_size=2 \\\r\n\r\nOn Fri, Nov 20, 2020 at 7:23 PM Tim Isbister <[email protected]>\r\nwrote:\r\n\r\n> Hmm strange, well actually I now tried running the provided example that I\r\n> suggested to you. But also failing with IndexError: index out of bounds\r\n> as you got in the first attempt. I have the latest versions of all the\r\n> libraries running on Ubuntu 18.04 LTS with Titan RTX.\r\n>\r\n> Provided example\r\n> <https://github.com/huggingface/transformers/tree/master/examples/language-modeling#xlnet-and-permutation-language-modeling>\r\n> :\r\n>\r\n> python run_plm.py \\\r\n> --model_name_or_path=xlnet-base-cased \\\r\n> --dataset_name wikitext \\\r\n> --dataset_config_name wikitext-2-raw-v1 \\\r\n> --do_train \\\r\n> --do_eval \\\r\n> --output_dir /tmp/test-plm\r\n>\r\n> Traceback (most recent call last):\r\n> File \"run_plm.py\", line 382, in <module>\r\n> main()\r\n> File \"run_plm.py\", line 321, in main\r\n> tokenized_datasets = tokenized_datasets.map(\r\n> File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/dataset_dict.py\", line 286, in map\r\n> {\r\n> File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/dataset_dict.py\", line 287, in <dictcomp>\r\n> k: dataset.map(\r\n> File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/arrow_dataset.py\", line 1243, in map\r\n> return self._map_single(\r\n> File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/arrow_dataset.py\", line 157, in wrapper\r\n> out: Union[\"Dataset\", \"DatasetDict\"] = func(self, *args, **kwargs)\r\n> File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/fingerprint.py\", line 163, in wrapper\r\n> out = func(self, *args, **kwargs)\r\n> File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/arrow_dataset.py\", line 1528, in _map_single\r\n> writer.write_batch(batch)\r\n> File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/arrow_writer.py\", line 278, in write_batch\r\n> pa_table = pa.Table.from_pydict(typed_sequence_examples)\r\n> File \"pyarrow/table.pxi\", line 1474, in pyarrow.lib.Table.from_pydict\r\n> File \"pyarrow/array.pxi\", line 322, in pyarrow.lib.asarray\r\n> File \"pyarrow/array.pxi\", line 222, in pyarrow.lib.array\r\n> File \"pyarrow/array.pxi\", line 110, in pyarrow.lib._handle_arrow_array_protocol\r\n> File \"/home/tim/anaconda3/envs/transformers/lib/python3.8/site-packages/datasets/arrow_writer.py\", line 100, in __arrow_array__\r\n> if trying_type and out[0].as_py() != self.data[0]:\r\n> File \"pyarrow/array.pxi\", line 1058, in pyarrow.lib.Array.__getitem__\r\n> File \"pyarrow/array.pxi\", line 540, in pyarrow.lib._normalize_index\r\n> IndexError: index out of bounds\r\n>\r\n> —\r\n> You are receiving this because you authored the thread.\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/8674#issuecomment-731472522>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AHYMJABV347A2NH2SK6NU4LSQ4B6HANCNFSM4T4CBVKQ>\r\n> .\r\n>", "Maybe @sgugger has an idea!", "The problem is that the script uses the tokenizer max length when no `max_seq_length` is passed, and that the XLNet tokenizer has a ridiculously high maximum sequence length. I have suggested a fix in #8738.\r\n\r\nWhile waiting for this PR to be merged, a temporary fix is to just add --max_seq_length 512 (or any value you'd like) to your command.", "I tried using the max_seq_length argument and set it to 400 and got the\r\nfollowing error:\r\n\r\nTraceback (most recent call last):\r\n File \"run_plm.py\", line 379, in <module>\r\n main()\r\n File \"run_plm.py\", line 244, in main\r\n cache_dir=model_args.cache_dir,\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/modeling_utils.py\",\r\nline 947, in from_pretrained\r\n model = cls(config, *model_args, **model_kwargs)\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/modeling_xlnet.py\",\r\nline 1294, in __init__\r\n self.transformer = XLNetModel(config)\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/modeling_xlnet.py\",\r\nline 940, in __init__\r\n self.reuse_len = config.reuse_len\r\nAttributeError: 'TransfoXLConfig' object has no attribute 'reuse_len'\r\n\r\n\r\nOn Mon, Nov 23, 2020 at 4:02 PM Lysandre Debut <[email protected]>\r\nwrote:\r\n\r\n> Closed #8674 <https://github.com/huggingface/transformers/issues/8674>\r\n> via #8738 <https://github.com/huggingface/transformers/pull/8738>.\r\n>\r\n> —\r\n> You are receiving this because you authored the thread.\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/8674#event-4029627504>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AHYMJADWOOMGPRQUSIUPRBDSRLEXRANCNFSM4T4CBVKQ>\r\n> .\r\n>", "The script only works for XLNet models, you will need to tweak it for other models", "I see. Thank you, Sylvain.\r\n\r\nOn Tue, Nov 24, 2020 at 10:46 AM Sylvain Gugger <[email protected]>\r\nwrote:\r\n\r\n> The script only works for XLNet models, you will need to tweak it for\r\n> other models\r\n>\r\n> —\r\n> You are receiving this because you authored the thread.\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/8674#issuecomment-733059560>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AHYMJADG3HM6UCRSXGS56ULSRPIM3ANCNFSM4T4CBVKQ>\r\n> .\r\n>\r\n", "Just to confirm that I am doing the right thing, what is the correct script\r\nin the language-modeling folder for fine-tuning Transformer-XL and CTRL? I\r\nam using run_clm.py for both of them currently but it keeps returning the\r\nsame error message for both. See error below:\r\n\r\nFor Transformer-XL:\r\n\r\n[INFO|modeling_utils.py:1065] 2020-11-24 15:55:30,982 >> All the\r\nweights of TransfoXLLMHeadModel were initialized from the model\r\ncheckpoint at transfo-xl-wt103.\r\nIf your task is similar to the task the model of the checkpoint was\r\ntrained on, you can already use TransfoXLLMHeadModel for predictions\r\nwithout further training.\r\n 2%|▏ | 6/311 [00:13<11:44, 2.31s/ba]Traceback (most recent\r\ncall last):\r\n File \"run_clm.py\", line 351, in <module>\r\n main()\r\n File \"run_clm.py\", line 261, in main\r\n load_from_cache_file=not data_args.overwrite_cache,\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/dataset_dict.py\",\r\nline 303, in map\r\n for k, dataset in self.items()\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/dataset_dict.py\",\r\nline 303, in <dictcomp>\r\n for k, dataset in self.items()\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/arrow_dataset.py\",\r\nline 1259, in map\r\n update_data=update_data,\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/arrow_dataset.py\",\r\nline 157, in wrapper\r\n out: Union[\"Dataset\", \"DatasetDict\"] = func(self, *args, **kwargs)\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/fingerprint.py\",\r\nline 163, in wrapper\r\n out = func(self, *args, **kwargs)\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/arrow_dataset.py\",\r\nline 1520, in _map_single\r\n batch, indices, check_same_num_examples=len(self.list_indexes()) >\r\n0, offset=offset\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/arrow_dataset.py\",\r\nline 1438, in apply_function_on_filtered_inputs\r\n function(*fn_args, effective_indices, **fn_kwargs) if with_indices\r\nelse function(*fn_args, **fn_kwargs)\r\n File \"run_clm.py\", line 254, in tokenize_function\r\n return tokenizer(examples[text_column_name])\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py\",\r\nline 2214, in __call__\r\n **kwargs,\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py\",\r\nline 2399, in batch_encode_plus\r\n **kwargs,\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils.py\",\r\nline 567, in _batch_encode_plus\r\n verbose=verbose,\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils.py\",\r\nline 630, in _batch_prepare_for_model\r\n return_attention_mask=return_attention_mask,\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py\",\r\nline 2531, in pad\r\n f\"type of {first_element} unknown: {type(first_element)}. \"\r\nValueError: type of [] unknown: <class 'list'>. Should be one of a\r\npython, numpy, pytorch or tensorflow object.\r\n 2%|▏ | 6/311 [00:16<13:48, 2.72s/ba]\r\n\r\n\r\nFor CTRL:\r\n\r\n[INFO|modeling_utils.py:1065] 2020-11-24 15:52:33,705 >> All the\r\nweights of CTRLLMHeadModel were initialized from the model checkpoint\r\nat ctrl.\r\nIf your task is similar to the task the model of the checkpoint was\r\ntrained on, you can already use CTRLLMHeadModel for predictions\r\nwithout further training.\r\n 0%| | 0/311 [00:00<?,\r\n?ba/s][WARNING|tokenization_utils_base.py:2736] 2020-11-24\r\n15:52:39,268 >> Token indices sequence length is longer than the\r\nspecified maximum sequence length for this model (293 > 256). Running\r\nthis sequence through the model will result in indexing errors\r\n 2%|▏ | 6/311 [00:05<04:21, 1.17ba/s]Traceback (most recent\r\ncall last):\r\n File \"run_clm.py\", line 351, in <module>\r\n main()\r\n File \"run_clm.py\", line 261, in main\r\n load_from_cache_file=not data_args.overwrite_cache,\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/dataset_dict.py\",\r\nline 303, in map\r\n for k, dataset in self.items()\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/dataset_dict.py\",\r\nline 303, in <dictcomp>\r\n for k, dataset in self.items()\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/arrow_dataset.py\",\r\nline 1259, in map\r\n update_data=update_data,\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/arrow_dataset.py\",\r\nline 157, in wrapper\r\n out: Union[\"Dataset\", \"DatasetDict\"] = func(self, *args, **kwargs)\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/fingerprint.py\",\r\nline 163, in wrapper\r\n out = func(self, *args, **kwargs)\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/arrow_dataset.py\",\r\nline 1520, in _map_single\r\n batch, indices, check_same_num_examples=len(self.list_indexes()) >\r\n0, offset=offset\r\n File \"/usr/local/lib/python3.6/dist-packages/datasets/arrow_dataset.py\",\r\nline 1438, in apply_function_on_filtered_inputs\r\n function(*fn_args, effective_indices, **fn_kwargs) if with_indices\r\nelse function(*fn_args, **fn_kwargs)\r\n File \"run_clm.py\", line 254, in tokenize_function\r\n return tokenizer(examples[text_column_name])\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py\",\r\nline 2214, in __call__\r\n **kwargs,\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py\",\r\nline 2399, in batch_encode_plus\r\n **kwargs,\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils.py\",\r\nline 567, in _batch_encode_plus\r\n verbose=verbose,\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils.py\",\r\nline 630, in _batch_prepare_for_model\r\n return_attention_mask=return_attention_mask,\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py\",\r\nline 2531, in pad\r\n f\"type of {first_element} unknown: {type(first_element)}. \"\r\nValueError: type of [] unknown: <class 'list'>. Should be one of a\r\npython, numpy, pytorch or tensorflow object.\r\n 2%|▏ | 6/311 [00:08<07:26, 1.46s/ba]\r\n\r\n\r\n\r\n\r\nOn Tue, Nov 24, 2020 at 10:50 AM Adaku Uchendu <[email protected]> wrote:\r\n\r\n> I see. Thank you, Sylvain.\r\n>\r\n> On Tue, Nov 24, 2020 at 10:46 AM Sylvain Gugger <[email protected]>\r\n> wrote:\r\n>\r\n>> The script only works for XLNet models, you will need to tweak it for\r\n>> other models\r\n>>\r\n>> —\r\n>> You are receiving this because you authored the thread.\r\n>> Reply to this email directly, view it on GitHub\r\n>> <https://github.com/huggingface/transformers/issues/8674#issuecomment-733059560>,\r\n>> or unsubscribe\r\n>> <https://github.com/notifications/unsubscribe-auth/AHYMJADG3HM6UCRSXGS56ULSRPIM3ANCNFSM4T4CBVKQ>\r\n>> .\r\n>>\r\n>\r\n>\r\n> --\r\n> *Adaku Uchendu*\r\n>\r\n> *McNair Scholar*\r\n> *Mathematics major*\r\n> *Statistic minor *\r\n> *Math Lab Tutor*\r\n> *Pre-Calculus LA*\r\n> *University of Maryland, Baltimore County *\r\n> *Class of 2018*\r\n>", "It looks like it comes from a bug in the slow tokenizers that can't handle an empty sequence at the beginning. We're looking into it. ", "Thank you, Sylvain.\r\n\r\nOn Tue, Nov 24, 2020 at 11:50 AM Sylvain Gugger <[email protected]>\r\nwrote:\r\n\r\n> It looks like it comes from a bug in the slow tokenizers that can't handle\r\n> an empty sequence at the beginning. We're looking into it.\r\n>\r\n> —\r\n> You are receiving this because you authored the thread.\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/8674#issuecomment-733103967>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AHYMJAHQJ4FQ24C6JWITIO3SRPP53ANCNFSM4T4CBVKQ>\r\n> .\r\n>", "Hi Sylvain, \r\n\r\nI just wanted to know if the slow tokenizer bug is no longer a problem?\r\nThank you ", "It seems to be solved on master, I didn't try on the v4 release but is might also be solved there too.", "Hi Sylvain,\r\n\r\nI just tried to re-run my code with the recent changes and I still got the same error. Just to make sure I am doing the correct thing, I have attached my code below. And I am using the transformer version 4.0 with the current github transformer repo.\r\nThank you.\r\n\r\n\r\ncd language-modeling/\r\npython run_clm.py \\\r\n --model_type=transfo-xl \\\r\n --model_name_or_path=transfo-xl-wt103 \\\r\n --train_file='/content/drive/My Drive/finetuned_models/train.txt' \\\r\n --validation_file='/content/drive/My Drive/finetuned_models/valid.txt' \\\r\n --save_total_limit=5 \\\r\n --num_train_epochs=1.0 \\\r\n --do_train \\\r\n --do_eval \\\r\n --per_gpu_train_batch_size=2 \\\r\n --per_gpu_eval_batch_size=2 \\\r\n --output_dir='/content/drive/My Drive/finetuned_models/transformer_xl'", "Indeed, the PR mentioned above will fix that specific issue.\r\n\r\nNote that `TransfoXLLMHeadModel` is not supported by `Trainer` anyway as it does not return the reduced loss. Once the PR is merged it should work with CTRL however." ]

[ { "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", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

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

[]

[ "It's not yet supported sadly", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored." ]

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

[]

[ "> For some models, I had to either pass some extra arguments to perform the metrics calculation or to access remote services to retrieve some additional data.\r\n\r\nHow did you do that in the current `Trainer`? I'm not against making `compute_metrics` more flexible but I don't see how you can add more arguments to its call without subclassing Trainer and overriding certain methods, in which case, you can also override the init.", "In fact I have my own subclassed Trainer, but I wanted to reuse as many as possible of the current `__init__` parameters of the regular Trainer. I know I can define my own `compute_metrics` function with a different signature in the `__init__` method of my Trainer, but I was trying to avoid that and reuse the current `compute_metrics` signature :-)\r\n\r\nMaybe the example above is not clear enough as the the DummyTrainer is not subclassing the standard one. The example was trying to highlight the function signature more than the reuse of the original Trainer.\r\n", "This seems like a very edge case to me, so I would leave the current `Trainer` as is, and adapt the `__init__` in your subclasses of `Trainer`.", "Yes, I agree that is kind of a corner case. In my subclassed Trainer I was trying to reuse the `prediction_loop()` method from the base Trainer but I needed to pass some more parameters to my `compute_metrics` function, apart from the `EvalPrediction` param. So the easiest workaround was of course, to copy the original `prediction_loop()` in my subclassed Trainer and, instead of calling:\r\n\r\nhttps://github.com/huggingface/transformers/blob/8062fa63c564d4cc0d29573acd89092f1eb1df64/src/transformers/trainer.py#L1398\r\n\r\n, I call my version of `compute_metrics` with the extra parameters. \r\n\r\nWith this proposal was trying to avoid that ugly copy-paste I had to do by 1) changing the `compute_metrics` signature as I described above, 2) defining an attribute in Trainer to serve as a placeholder for the extra metrics arguments (and which can be set from subclassed Trainers,) e.g.:\r\n\r\n```python\r\nself.metrics_extra_args: Optional[Dict] = None\r\n```\r\nand 3) changing line 1398 to something like:\r\n\r\n```python\r\nmetrics = self.compute_metrics(EvalPrediction(predictions=preds, label_ids=label_ids), self.metrics_extra_args)\r\n```\r\n\r\nBut I understand that is kind of convoluted. Thanks @sgugger in any case for your consideration!", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md) are likely to be ignored.", "I really need this" ]

[ { "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" } ]

[]

[ "Good catch, thanks!" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "Closing since it was fixed in #8585" ]

[]

[]

[]

[]

[]

[ ":+1: nice job!", "Hi, this command may need a args?\r\n```\r\ntransformers-cli lfs-enable-largefiles repo_path\r\n```\r\n", "> Hi, this command may need a args?\r\n\r\nYes, correct\r\n", "Thanks a lot for the PR, and thanks for letting me know.\r\n\r\nI will give my feedback after testing it with the 11B model soon. ", "Hi, thanks a lot for the PR.\r\nI try to reinstall transformers with this new branch and follow with this comment:\r\n```\r\ngit clone https://huggingface.co/mymusise/CPM-Third-Party\r\ncd CPM-Third-Party\r\ngit lfs install \r\ntransformers-cli lfs-enable-largefiles .\r\n\r\ncp ../models/tf_model.h5 ./\r\ngit add . && git commit -m 'add model'\r\ngit push\r\n```\r\nThen, after half an hour later, it raised an error:\r\n```\r\n$ git push\r\nGit LFS: (0 of 1 files) 0 B / 9.68 GB\r\nGit LFS: (0 of 1 files) 4.66 GB / 9.68 GB\r\nEOFoading LFS objects: 0% (0/1), 9.31 B / 9.68 GB\r\nerror: failed to push some refs to 'https://huggingface.co/mymusise/CPM-Third-Party'\r\n```\r\n\r\nDid I do something wrong?", "@mymusise might have been an intermittent server error. Can you try again?", "> @mymusise might have been an intermittent server error. Can you try again?\r\n\r\nYes, I try again and again. But it always raises this error at `9.31 GB / 9.68 GB`.\r\n\r\n:eyes: But, now `git push` will return a ` 502 ` error :\r\n```\r\nfatal: unable to access 'https://huggingface.co/mymusise/CPM-Third-Party/': The requested URL returned error: 502\r\n```", "@mymusise Yes, looks like this is crashing/locking the server 😱\r\n\r\nDo you mind trying again on Monday? As we'll have more bandwidth to fix then. Sorry about that :/", "It's ok, guy, haha. Waiting for your good news. \r\nHappy weekend!", ">Yes, I try again and again. But it always raises this error at 9.31 GB / 9.68 GB.\r\n\r\nHi, I try again today and push the big model file without any exception! :tada: \r\nThank guys!", "Hi, here I push another model file(4.9GB) again, but this time it gives me a **504** Gateway Time-out error :sweat:\r\n\r\n```\r\nroot@iZt4n9z4x3ph9oc3hhrdneZ:~/CPM-FP16-Third-Party# time git push \r\nUsername for 'https://huggingface.co': mymusise\r\nPassword for 'https://[email protected]': \r\nCounting objects: 3, done.\r\nCompressing objects: 100% (2/2), done.\r\nWriting objects: 66% (2/3), 2.52 GiB | 3.62 MiB/s \r\nWriting objects: 100% (3/3), 4.46 GiB | 3.77 MiB/s, done.\r\nTotal 3 (delta 0), reused 1 (delta 0)\r\nerror: RPC failed; HTTP 504 curl 22 The requested URL returned error: 504 Gateway Time-out\r\nfatal: The remote end hung up unexpectedly\r\nfatal: The remote end hung up unexpectedly\r\nEverything up-to-date\r\n\r\nreal\t22m8.653s\r\nuser\t0m29.758s\r\nsys\t0m17.136s\r\n```\r\n\r\nSeems the big file is uploaded completely, it looks like there is some problem with the server configuration about the timeout. @julien-c \r\n\r\nTried three times with the same result.", "@Pierrci - I ran some tests on ~10 GB files (`t5-3b`) and didn't encounter any problems! However when doing it for ~45GB files (`t5-11b`), I encounter some problems.\r\n\r\n```\r\ncd t5-11b-repo\r\ntransformers-cli lfs-enable-largefiles /path/to/repo\r\ngit add .\r\ngit commit -m \"add\"\r\ngit push # <= this command failse\r\n```\r\n\r\nIn case it is useful, here is a link to the `git trace` error message: https://github.com/patrickvonplaten/files_to_link_to/blob/master/output.txt\r\n\r\nI can always go back to the way of manually uploading to the git-lfs hash path. Maybe the error message is helpful though :-) ", "Thanks @patrickvonplaten, I see where it might be coming from, gonna look at it now!", "I deployed a fix that should address your problem, can you try again @patrickvonplaten?\r\n\r\n@mymusise Is it possible for you to try again with `GIT_CURL_VERBOSE=1 git push` so that we can try to get more information? From what you shared so far, your error seems to be different from Patrick's one.", "Thanks, @Pierrci. Yes, I think my error is different from Patrick's one. Here is the [information](https://gist.github.com/mymusise/bfa331e3effe0876efbc0011334ead96) with `GIT_CURL_VERBOSE=1 git push`. \r\nHope it help.\r\n", "> I deployed a fix that should address your problem, can you try again @patrickvonplaten?\r\n> \r\n> @mymusise Is it possible for you to try again with `GIT_CURL_VERBOSE=1 git push` so that we can try to get more information? From what you shared so far, your error seems to be different from Patrick's one.\r\n\r\nAwesome it works now @Pierrci - thanks! :-) ", "> Thanks, @Pierrci. Yes, I think my error is different from Patrick's one. Here is the [information](https://gist.github.com/mymusise/bfa331e3effe0876efbc0011334ead96) with `GIT_CURL_VERBOSE=1 git push`.\r\n> Hope it help.\r\n\r\n@mymusise Are you sure LFS is properly installed and configured for the repo? From your logs it seems your `git push` command isn't doing any LFS work (like running a `pre-push` hook or calling our LFS endpoint), trying instead to push all the files through the classic git endpoint, which can't work." ]

[]

[]

[ "This is a known issue that's being tracked at #8663" ]

[]

[]

[ "Hello, could you please fill the issue template when opening issues? Otherwise we cannot help you. Thanks.", "## Environment info\r\n \r\n- `transformers` version: 3.5.1\r\n- Platform: cpu \r\n- Python version: 3.7\r\n- PyTorch version (GPU?): 1.0.4\r\n- Tensorflow version (GPU?): \r\n tensorflow-datasets 4.1.0 <pip>\r\ntensorflow-metadata 0.25.0 <pip>\r\n- Using GPU in script?: - \r\n- Using distributed or parallel set-up in script?: - \r\n\r\n### Who can help\r\n\r\n Here is my two lines to get t5-config:\r\n\r\n```\r\nfrom transformers import AutoConfig\r\nconfig = AutoConfig.from_pretrained('t5-base')\r\n```\r\n Model Cards: @julien-c\r\n T5: @patrickvonplaten\r\n\r\n## Information\r\nHere are the errors, is there an issue with t5-base storage? I am really confused, thank you for your help on this.\r\nthanks \r\n\r\n```\r\nfile t5-base/config.json not found\r\nTraceback (most recent call last):\r\n File \"/opt/conda/envs/internship/lib/python3.7/site-packages/transformers/configuration_utils.py\", line 388, in get_config_dict\r\n local_files_only=local_files_only,\r\n File \"/opt/conda/envs/internship/lib/python3.7/site-packages/transformers/file_utils.py\", line 962, in cached_path\r\n raise EnvironmentError(\"file {} not found\".format(url_or_filename))\r\nOSError: file t5-base/config.json not found\r\n\r\nDuring handling of the above exception, another exception occurred:\r\n\r\nTraceback (most recent call last):\r\n File \"test.py\", line 2, in <module>\r\n config = AutoConfig.from_pretrained('t5-base')\r\n File \"/opt/conda/envs/internship/lib/python3.7/site-packages/transformers/configuration_auto.py\", line 333, in from_pretrained\r\n config_dict, _ = PretrainedConfig.get_config_dict(pretrained_model_name_or_path, **kwargs)\r\n File \"/opt/conda/envs/internship/lib/python3.7/site-packages/transformers/configuration_utils.py\", line 400, in get_config_dict\r\n raise EnvironmentError(msg)\r\nOSError: Can't load config for 't5-base'. Make sure that:\r\n\r\n- 't5-base' is a correct model identifier listed on 'https://huggingface.co/models'\r\n\r\n- or 't5-base' is the correct path to a directory containing a config.json file\r\n```\r\n\r\n## To reproduce\r\nplease run the two lines above.\r\n\r\n## Expected behavior\r\nloading the config \r\n", "Hi, sure, done", "Hi @LysandreJik this issues is really weird and really blocking me, I greatly appreciate having a look. thanks ", "Hey @rabeehk - I cannot reproduce the error. I suspect the following:\r\n\r\nYou run the code from a directory that includes a local folder that is called `t5-base` which does not have a `config.json`.\r\nCould you try to run:\r\n\r\n```\r\nfrom transformers import AutoConfig\r\nconfig = AutoConfig.from_pretrained('t5-base')\r\n```\r\n\r\nfrom another folder or delete the `t5-base` folder that might be the reason? ", "Hi patrick\nthank you for the reply, this is what happening when I call this line:\n\n config = T5Config.from_pretrained(\n model_args.config_name if model_args.config_name else\nmodel_args.model_name_or_path,\n )\n\nthe code by itself creates an empty t5-base directory, I delete it and then\nit recreates it.\nDo you have an idea on this?\nthanks\nBest\nRabeeh\n\nOn Thu, Nov 19, 2020 at 9:32 PM Patrick von Platen <[email protected]>\nwrote:\n\n> Hey @rabeehk <https://github.com/rabeehk> - I cannot reproduce the error.\n> I suspect the following:\n>\n> You run the code from a directory that includes a local folder that is\n> called t5-base which does not have a config.json.\n> Could you try to run:\n>\n> from transformers import AutoConfig\n> config = AutoConfig.from_pretrained('t5-base')\n>\n> from another folder or delete the t5-base folder that might be the reason?\n>\n> —\n> You are receiving this because you were mentioned.\n> Reply to this email directly, view it on GitHub\n> <https://github.com/huggingface/transformers/issues/8661#issuecomment-730620761>,\n> or unsubscribe\n> <https://github.com/notifications/unsubscribe-auth/ABP4ZCGIXURMYC7WWVFZPXDSQV6GTANCNFSM4T3SASPQ>\n> .\n>\n", "Could you tell me please how to set the path when using datasets that it\ndoes not cache in the default path?\nto me there might not be space for the model and this is all happening.\nOverall all the caches done in huggingface code, could you tell me how to\nset the different path for them\nthanks\n\nOn Fri, Nov 20, 2020 at 1:12 PM Rabeeh Karimi <[email protected]> wrote:\n\n> Hi patrick\n> thank you for the reply, this is what happening when I call this line:\n>\n> config = T5Config.from_pretrained(\n> model_args.config_name if model_args.config_name else\n> model_args.model_name_or_path,\n> )\n>\n> the code by itself creates an empty t5-base directory, I delete it and\n> then it recreates it.\n> Do you have an idea on this?\n> thanks\n> Best\n> Rabeeh\n>\n> On Thu, Nov 19, 2020 at 9:32 PM Patrick von Platen <\n> [email protected]> wrote:\n>\n>> Hey @rabeehk <https://github.com/rabeehk> - I cannot reproduce the\n>> error. I suspect the following:\n>>\n>> You run the code from a directory that includes a local folder that is\n>> called t5-base which does not have a config.json.\n>> Could you try to run:\n>>\n>> from transformers import AutoConfig\n>> config = AutoConfig.from_pretrained('t5-base')\n>>\n>> from another folder or delete the t5-base folder that might be the\n>> reason?\n>>\n>> —\n>> You are receiving this because you were mentioned.\n>> Reply to this email directly, view it on GitHub\n>> <https://github.com/huggingface/transformers/issues/8661#issuecomment-730620761>,\n>> or unsubscribe\n>> <https://github.com/notifications/unsubscribe-auth/ABP4ZCGIXURMYC7WWVFZPXDSQV6GTANCNFSM4T3SASPQ>\n>> .\n>>\n>\n", "So when I run the codes, there is caching done here\ncahce dir /idiap/home/rkarimi/.cache/huggingface/datasets\ncahce dir /idiap/home/rkarimi/.cache/huggingface/datasets/downloads\n\ncould I change this?\nthanks\nBest\nRabeeh\n\nOn Fri, Nov 20, 2020 at 1:44 PM Rabeeh Karimi <[email protected]> wrote:\n\n> Could you tell me please how to set the path when using datasets that it\n> does not cache in the default path?\n> to me there might not be space for the model and this is all happening.\n> Overall all the caches done in huggingface code, could you tell me how to\n> set the different path for them\n> thanks\n>\n> On Fri, Nov 20, 2020 at 1:12 PM Rabeeh Karimi <[email protected]>\n> wrote:\n>\n>> Hi patrick\n>> thank you for the reply, this is what happening when I call this line:\n>>\n>> config = T5Config.from_pretrained(\n>> model_args.config_name if model_args.config_name else\n>> model_args.model_name_or_path,\n>> )\n>>\n>> the code by itself creates an empty t5-base directory, I delete it and\n>> then it recreates it.\n>> Do you have an idea on this?\n>> thanks\n>> Best\n>> Rabeeh\n>>\n>> On Thu, Nov 19, 2020 at 9:32 PM Patrick von Platen <\n>> [email protected]> wrote:\n>>\n>>> Hey @rabeehk <https://github.com/rabeehk> - I cannot reproduce the\n>>> error. I suspect the following:\n>>>\n>>> You run the code from a directory that includes a local folder that is\n>>> called t5-base which does not have a config.json.\n>>> Could you try to run:\n>>>\n>>> from transformers import AutoConfig\n>>> config = AutoConfig.from_pretrained('t5-base')\n>>>\n>>> from another folder or delete the t5-base folder that might be the\n>>> reason?\n>>>\n>>> —\n>>> You are receiving this because you were mentioned.\n>>> Reply to this email directly, view it on GitHub\n>>> <https://github.com/huggingface/transformers/issues/8661#issuecomment-730620761>,\n>>> or unsubscribe\n>>> <https://github.com/notifications/unsubscribe-auth/ABP4ZCGIXURMYC7WWVFZPXDSQV6GTANCNFSM4T3SASPQ>\n>>> .\n>>>\n>>\n", "> Hi patrick thank you for the reply, this is what happening when I call this line: config = T5Config.from_pretrained( model_args.config_name if model_args.config_name else model_args.model_name_or_path, ) the code by itself creates an empty t5-base directory, I delete it and then it recreates it. Do you have an idea on this? thanks Best Rabeeh\r\n> […](#)\r\n> On Thu, Nov 19, 2020 at 9:32 PM Patrick von Platen ***@***.***> wrote: Hey @rabeehk <https://github.com/rabeehk> - I cannot reproduce the error. I suspect the following: You run the code from a directory that includes a local folder that is called t5-base which does not have a config.json. Could you try to run: from transformers import AutoConfig config = AutoConfig.from_pretrained('t5-base') from another folder or delete the t5-base folder that might be the reason? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <[#8661 (comment)](https://github.com/huggingface/transformers/issues/8661#issuecomment-730620761)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABP4ZCGIXURMYC7WWVFZPXDSQV6GTANCNFSM4T3SASPQ> .\r\n\r\nI don't think `from_pretrained(...)` ever creates a directory -> this should not happen. Not sure what's going on there. Can you maybe add a colab where I can reproduce your error?", "> Could you tell me please how to set the path when using datasets that it does not cache in the default path? to me there might not be space for the model and this is all happening. Overall all the caches done in huggingface code, could you tell me how to set the different path for them thanks\r\n> […](#)\r\n> On Fri, Nov 20, 2020 at 1:12 PM Rabeeh Karimi ***@***.***> wrote: Hi patrick thank you for the reply, this is what happening when I call this line: config = T5Config.from_pretrained( model_args.config_name if model_args.config_name else model_args.model_name_or_path, ) the code by itself creates an empty t5-base directory, I delete it and then it recreates it. Do you have an idea on this? thanks Best Rabeeh On Thu, Nov 19, 2020 at 9:32 PM Patrick von Platen < ***@***.***> wrote: > Hey @rabeehk <https://github.com/rabeehk> - I cannot reproduce the > error. I suspect the following: > > You run the code from a directory that includes a local folder that is > called t5-base which does not have a config.json. > Could you try to run: > > from transformers import AutoConfig > config = AutoConfig.from_pretrained('t5-base') > > from another folder or delete the t5-base folder that might be the > reason? > > — > You are receiving this because you were mentioned. > Reply to this email directly, view it on GitHub > <[#8661 (comment)](https://github.com/huggingface/transformers/issues/8661#issuecomment-730620761)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/ABP4ZCGIXURMYC7WWVFZPXDSQV6GTANCNFSM4T3SASPQ> > . >\r\n\r\nIs this issue about `t5-base` config or about datasets? I don't follow here", "Hi Patrick\nthis is now solved, I was mistakenly choose the output_path as t5-base and\nthis was the reason for the creation of empty t5-base directory.\nI would like to thank you so much for the help.\nBest\nRabeeh\n\nOn Fri, Nov 20, 2020 at 3:01 PM Patrick von Platen <[email protected]>\nwrote:\n\n> Could you tell me please how to set the path when using datasets that it\n> does not cache in the default path? to me there might not be space for the\n> model and this is all happening. Overall all the caches done in huggingface\n> code, could you tell me how to set the different path for them thanks\n> … <#m_3116611502979953478_>\n> On Fri, Nov 20, 2020 at 1:12 PM Rabeeh Karimi *@*.*> wrote: Hi patrick\n> thank you for the reply, this is what happening when I call this line:\n> config = T5Config.from_pretrained( model_args.config_name if\n> model_args.config_name else model_args.model_name_or_path, ) the code by\n> itself creates an empty t5-base directory, I delete it and then it\n> recreates it. Do you have an idea on this? thanks Best Rabeeh On Thu, Nov\n> 19, 2020 at 9:32 PM Patrick von Platen < @.*> wrote: > Hey @rabeehk\n> <https://github.com/rabeehk> https://github.com/rabeehk - I cannot\n> reproduce the > error. I suspect the following: > > You run the code from a\n> directory that includes a local folder that is > called t5-base which does\n> not have a config.json. > Could you try to run: > > from transformers\n> import AutoConfig > config = AutoConfig.from_pretrained('t5-base') > > from\n> another folder or delete the t5-base folder that might be the > reason? > >\n> — > You are receiving this because you were mentioned. > Reply to this\n> email directly, view it on GitHub > <#8661 (comment)\n> <https://github.com/huggingface/transformers/issues/8661#issuecomment-730620761>>,\n> > or unsubscribe >\n> https://github.com/notifications/unsubscribe-auth/ABP4ZCGIXURMYC7WWVFZPXDSQV6GTANCNFSM4T3SASPQ\n> > . >\n>\n> Is this issue about t5-base config or about datasets? I don't follow here\n>\n> —\n> You are receiving this because you were mentioned.\n> Reply to this email directly, view it on GitHub\n> <https://github.com/huggingface/transformers/issues/8661#issuecomment-731187426>,\n> or unsubscribe\n> <https://github.com/notifications/unsubscribe-auth/ABP4ZCATGEHAA7275ZAEFZLSQZZFPANCNFSM4T3SASPQ>\n> .\n>\n" ]

[]

[]

[]