Datasets:

---

DanielPFlorian

/

Transformers-Github-Issues

like 0

[]

[]

[]

[]

[]

[ "Hy @sadakmed, \r\n\r\nlet me know if you need help on the issue or if you don't find the time to tackle it. I'll then just make it open to the \"public\" again :-) ", "@patrickvonplaten The loss function it what I stuck on, thank you very much for your guidance.\r\n\r\n\r\n> Let me know if you need help or are stuck :-)\r\n\r\nfor sure I will ;-) ", "Hey @sadakmed,\r\n\r\nDo you have an update on the PR? It's been three weeks now and it would be great to merge this soon. Sorry, we're very fast-moving in this lib and other community contributors have started asking for this feature. By next week, I'll probably have to take a look myself or redistribute the issue. ", "Hi @patrickvonplaten my apologies, \r\n\r\nCould you Please see it now, lemme know if anything is missing. ", "Hi @patrickvonplaten, working on the test 'BartStandaloneCausalLM': I dont know if the `self.model_tester` in 'setUp' should be 'BartDecoderTester' (needed to be implemented), or do u recommend something else. \r\n\r\n ", "Hey @sadakmed, \r\n\r\nThanks a lot for you additions here :-) \r\nYes, we need a new `BartStandaloneDecoderModelTester` analog to how it's done for ProphetNet in `tests/test_modeling_prophetnet.py`. Do you want to give it a try? Otherwise, I can go into your PR and see how to add the tests :-) ", "Hi @patrickvonplaten \r\n\r\n> Do you want to give it a try?\r\n\r\nof course, I'm working on it, ", "Hi @patrickvonplaten, I just pushed the test, could you please check it out! \r\nthaaaanks ", "Hey @sadakmed, \r\n\r\nI corrected `BartForCausalLM` and also added `MBartForCausalLM`. It would be awesome if you could take care of adding `MarianForCausalLM`, `PegasusForCausalLM`, `BlenderbotForCausalLM`, and `BlenderbotSmallForCausalLM`. \r\n\r\nTo do so you can simply copy everything that was done for `MBart` in this PR 1-to-1 to the mentioned models above. Let me know if sounds feasible for you :-) \r\n\r\nThanks a lot for your help so far!", "> adding `MarianForCausalLM`, `PegasusForCausalLM`, `BlenderbotForCausalLM`, and `BlenderbotSmallForCausalLM`.\r\n\r\nimplementing it for use with EncoderDecoder or just the test? \r\n\r\nYes I would like to do it, with all pleasure.", "> > adding `MarianForCausalLM`, `PegasusForCausalLM`, `BlenderbotForCausalLM`, and `BlenderbotSmallForCausalLM`.\r\n> \r\n> implementing it for use with EncoderDecoder or just the test?\r\n> \r\n> Yes I would like to do it, with all pleasure.\r\n\r\nFor those models, there is no need to add a test to `EncoderDecoderModel`. We should only copy-paste the code that was added to MBart to those models and also copy-paste the test in `test_modeling_marian.py` e.g.", "@patrickvonplaten Could you please check the test if it well, and about the test of `Decoder only` I didn't get what do you mean!!", "It would be nice to fix the tests and also add tests for `Pegasus`, `Blenderbot`, and `BlenderbotSmall`", "> It would be nice to fix the tests and also add tests for `Pegasus`, `Blenderbot`, and `BlenderbotSmall`\r\n\r\n@patrickvonplaten, exactly like the one was for Marian?", "@patrickvonplaten could you check please!", "**UPDATE:**\r\n\r\n@LysandreJik @sgugger \r\nThis PR enables all Bart-like models to be used in combination with the Encoder-Decoder framework. The model `BartForCausalLM` is added for Bart and then copied to all other models via the copying mechanism. Also, a new model tester is added for all those models. \r\nWhile working on this I found a small bug for a very edge-case scenario for Bart and corrected it here: https://github.com/huggingface/transformers/pull/9128/files#r569436360 . The newly added tests were failing, which made me aware of the bug. \r\nAlso, I had to slightly change the `check_repo.py` file so that it counts both classes from `all_model_classes` with 1 and 2 paratheses. \r\n", "Great job @sadakmed ", "> Great job @sadakmed\r\n\r\nwouldn't happen without you, thank you very much. see u in the next PR ;) " ]

[]

[]

[ "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." ]

[]

[]

[ "Make sure you install a recent version of transformers, `ParallelMode` was added to the master branch some ~2 weeks ago.", "Yes, as @KDercksen points out, you need an up-to-date install from source to be able to run the examples (as mentioned in the main examples folder README)." ]

[]

[]

[ "Here is how I managed to do it. I have considered a `pandas` dataframe, but you can easily extend it to predict individual sentence too.\r\n\r\n```\r\nimport pandas as pd \r\nimport numpy as np\r\nfrom datasets import Dataset, load_dataset\r\nfrom scipy.special import softmax\r\nfrom transformers import Trainer\r\nfrom transformers import BertTokenizer, BertForSequenceClassification\r\n\r\ntokenizer = BertTokenizer.from_pretrained(<model_name_or_path>)\r\nmodel = BertForSequenceClassification.from_pretrained(<model_name_or_path>)\r\n\r\ndef preprocess_function(examples):\r\n # Tokenize the texts\r\n result = tokenizer(examples['sentence'], padding=False, max_length=None, truncation=True, verbose=False)\r\n return result\r\n\r\ndef predict(dataframe):\r\n eval_dataset = Dataset.from_pandas(dataframe)\r\n eval_dataset = eval_dataset.map(preprocess_function, batched=False, load_from_cache_file=True)\r\n # Initialize our Trainer\r\n trainer = Trainer(model=model, tokenizer=tokenizer)\r\n predictions = trainer.predict(test_dataset=eval_dataset).predictions\r\n # Adding a softmax layer to get probabilities. If you want class labels instead - np.argmax(predictions, axis=1)\r\n predictions = np.array([softmax(element) for element in predictions])[:, 1]\r\n return predictions\r\n```", "@Nickil21 \r\nI used the torch model in the trainer. It's much faster than using pandas and creating a dataset.\r\n\r\n`import torch\r\ndef test_2(trainer, sentence1, sentence2):\r\n id_tolabel = {0:'negative', 1: 'positive'}\r\n model = trainer.model.eval()\r\n tokenized = tokenizer(sentence1, sentence2, return_tensors='pt').to(model.device)\r\n with torch.no_grad():\r\n label = torch.argmax(trainer.model.forward(**tokenized).logits, dim=1)[0].cpu().item()\r\n return id_tolabel[label]\r\nprint(test_2(trainer, 'it is not possible', 'this is impossible'))`" ]

[]

[]

[ "A Python profiling call gives the following improvements:\r\n```\r\nmodel = TFBertModel.from_pretrained(\"bert-base-cased\")\r\n\r\n# With the improvements\r\ncProfile.run(\"model(model.dummy_inputs)\") \r\n54591 function calls (53774 primitive calls) in 0.064 seconds\r\n\r\n# Currently on master\r\ncProfile.run(\"model(model.dummy_inputs)\")\r\n76166 function calls (75204 primitive calls) in 0.095 seconds\r\n```", "Thanks @patrickvonplaten !!\r\n\r\n1. Slow tests are passing for these models\r\n2. I confirm that \"Old\" pre-trained models `tf_model.h5` files that were saved with tf < 2.3 can be loaded into the new layer design\r\n\r\nI haven't tested the tf1 models, you mean testing the `load_tf_weights_in_bert` in the `modeling_bert.py` file?", "@jlei2 has confirmed that now everything works as expected in the profiler and benchmark 👍 https://github.com/huggingface/transformers/issues/6771#issuecomment-745786314", "> 2\\. \"Old\" pre-trained models `tf_model.h5` files that were saved with tf < 2.3 can be loaded into the new layer des\r\n\r\nYeah I mean loading a tf `.ckpt` file using the `from_pretrained(...)` method. The `from_pretrained(...)` method automatically uses the correct functions to load `.ckpt`. I think the easiest way would be to download one of the zips of the official google bert: https://github.com/google-research/bert#bert and quickly check that it can be loaded and that the output on this branch and on master is the same.", "> > 2. \"Old\" pre-trained models `tf_model.h5` files that were saved with tf < 2.3 can be loaded into the new layer des\r\n> \r\n> Yeah I mean loading a tf `.ckpt` file using the `from_pretrained(...)` method. The `from_pretrained(...)` method automatically uses the correct functions to load `.ckpt`. I think the easiest way would be to download one of the zips of the official google bert: https://github.com/google-research/bert#bert and quickly check that it can be loaded and that the output on this branch and on master is the same.\r\n\r\nOk as discussed offline TF1 checkpoints cannot even be loaded into TF2 at the moment (only if one goes through PT), so this PR is good to go for me!" ]

[]

[ { "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 } ]

[]

[]

[]

[ "Could you please also try with the most recent transformers release and report what happens?", "Version 4.0.1 prints: ['Hug', 'ging', 'face'] ", "Explicitly setting the attribute 'do_lower_case' to True solves the problem.\r\n\r\n```python\r\nfrom transformers import RobertaTokenizer\r\ntokenizer = RobertaTokenizer.from_pretrained(\"roberta-base\", do_lower_case=True)\r\ntokenizer.do_lower_case = True\r\nprint(tokenizer.tokenize(\"Huggingface\"))\r\n```\r\n\r\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 we use the AutoTokenizer library, this still does not work. \r\n\r\n```\r\nfrom transformers import AutoTokenizer\r\ntokenizer = AutoTokenizer.from_pretrained(\"roberta-base\", do_lower_case=True)\r\ntokenizer.do_lower_case = True\r\nprint(tokenizer.tokenize(\"Huggingface\"))\r\n```" ]

[]

[ { "login": "SBrandeis", "id": 33657802, "node_id": "MDQ6VXNlcjMzNjU3ODAy", "avatar_url": "https://avatars.githubusercontent.com/u/33657802?v=4", "gravatar_id": "", "url": "https://api.github.com/users/SBrandeis", "html_url": "https://github.com/SBrandeis", "followers_url": "https://api.github.com/users/SBrandeis/followers", "following_url": "https://api.github.com/users/SBrandeis/following{/other_user}", "gists_url": "https://api.github.com/users/SBrandeis/gists{/gist_id}", "starred_url": "https://api.github.com/users/SBrandeis/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/SBrandeis/subscriptions", "organizations_url": "https://api.github.com/users/SBrandeis/orgs", "repos_url": "https://api.github.com/users/SBrandeis/repos", "events_url": "https://api.github.com/users/SBrandeis/events{/privacy}", "received_events_url": "https://api.github.com/users/SBrandeis/received_events", "type": "User", "site_admin": false } ]

[]

[]

[]

[ "LGTM!" ]

[]

[]

[ "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\nThanks!" ]

[]

[]

[ "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\nThanks!" ]

[ { "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": 2669577093, "node_id": "MDU6TGFiZWwyNjY5NTc3MDkz", "url": "https://api.github.com/repos/huggingface/transformers/labels/PR%20for%20Model%20Addition", "name": "PR for Model Addition", "color": "5319e7", "default": false, "description": "" } ]

[]

[ "Don't worry about the TF tests, these are because of TF2.4 which are fixed on `master`.", "The conversion script currently includes a line in which I'm importing a local `vocab.txt`, I know this should be removed in the future." ]

[]

[]

[ "One potential fix is to not use the same padding_idx for the position_ids embedding, why not just use 0? The least actual unpadded value of incremental_indices will be one so 0 is a valid pad.\r\n\r\nIn the `__init__` of `RobertaEmbeddings` (note this code is also duplicated, self.position_embeddings is intiialised twice!)\r\n\r\n self.position_embeddings = nn.Embedding(\r\n config.max_position_embeddings, config.hidden_size, padding_idx=0 # replaced self.padding_idx with 0\r\n )\r\n\r\nAnd then modify create_position_ids_from_input_ids\r\n\r\n```\r\ndef create_position_ids_from_input_ids(input_ids, padding_idx):\r\n \"\"\"\r\n Replace non-padding symbols with their position numbers. Position numbers begin at padding_idx+1. Padding symbols\r\n are ignored. This is modified from fairseq's `utils.make_positions`.\r\n\r\n Args:\r\n x: torch.Tensor x:\r\n\r\n Returns: torch.Tensor\r\n \"\"\"\r\n # The series of casts and type-conversions here are carefully balanced to both work with ONNX export and XLA.\r\n mask = input_ids.ne(padding_idx).int()\r\n incremental_indices = torch.cumsum(mask, dim=1).type_as(mask) * mask\r\n return incremental_indices.long() # removed + padding_idx\r\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." ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "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": "" } ]

[]

[ "🔥 Looking forward to taking a look at the conversion script from GluonNLP/mxnet!", "@patrickvonplaten I added some examples for both `modeling_bort.py` and modeling_tf_bort.py` :hugs: \r\n\r\n@julien-c The conversion script is also added - you just need to install `gluonnlp==0.8.3` and `mxnet==1.5.0`.\r\n\r\nThese versions are defined in the BORT [requirements file](https://github.com/alexa/bort/blob/master/requirements.txt). The conversion script also performs a version check.", "We'll have to think a bit how to advertise this. Let me draft up a \"Contribution Proposal\" for the fine-tuning algorithm.", "Hey @stefan-it,\r\n\r\nI've discussed a bit with @LysandreJik and @sgugger offline and I do agree with @LysandreJik after having thought about it again. I think it's better if we actually don't add any new code (besides the conversion script) that should be added to `src/transformers/models/bert/` and the docs page. I'm very sorry to have you asked to go down this road! I think however it does make more sense to not add any \"tokenizer\" or \"model\" code as those are exact copies of the `RobertaTokenizer` and `BertModel`. It's probably most efficient to open a new PR and only add the required files. Super sorry again!", "Are we planning to implement the architectural optimization (FPTAS) or just the pre-trained models?", "> Are we planning to implement the architectural optimization (FPTAS) or just the pre-trained models?\r\n\r\nGreat question! For now, we'll just add the model weights - see: #9813. A community contribution showing how to do FPTAS in a notebook would be extremely valuable though.", "Closing in favor of #9813" ]

[]

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

[ "It also might be a good idea to catch this error somewhere before `IndexError: index out of range in self` because that is not descriptive and makes debugging difficult.", "You're right @HHousen - thanks for the note! Do you want to open a PR to fix the docs? That would be awesome :-) Otherwise, I can do it as well ", "We are using longformer and we are passing (input_ids, attention_mask , global_attention_mask ,token_type_ids) as input. if we are passing token_type_ids as 0's we are not having any issues but when we try to pass token_type_ids as 1's, or 0's afor segments within the sequence it is throwing following error.\r\n**IndexError: index out of range in self**\r\n**C:/w/b/windows/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: block: [52,0,0], thread: [32,0,0] Assertion `srcIndex < srcSelectDimSize` failed.**", "i have created separate issue for **index out of range in self ** (https://github.com/huggingface/transformers/issues/9162) while using token_type_ids, from the above comment by @HHousen should i remove token_type_ids as parameter while passing it to model ?", "@yuvarajvc Correct. The Longformer doesn't support `token_type_ids`, so you should not pass them to the model." ]

[]

[]

[ "The `run_mlm` script trains a `RobertaForMaskedLM` model, which does not have a pooler layer. That's why you get this warning when using this pretrained model to initialize a `RobertaModel`.", "Thanks @sgugger! Now I get where the problem was. \r\nMoreover, I found some good tutorials on it. Here are the links for those who need it.\r\nhttps://zablo.net/blog/post/training-roberta-from-scratch-the-missing-guide-polish-language-model/\r\nhttps://colab.research.google.com/github/huggingface/blog/blob/master/notebooks/01_how_to_train.ipynb" ]

[]

[]

[ "Could you elaborate on the script you're using. `run_ner.py` does not report `eval_class_report` or `eval_predictions` and there are no print statements in it, nor are there in `Trainer.train` method.", "Yes, it's a custom script. The script creates an `AutoModelForTokenClassification,` passes it to the `Trainer` and calls the `Trainer.train` method.\r\n\r\nWe are using WandB to plot a confusion matrix, so we define our own `compute_metrics` function which we also pass to the `Trainer` (sorry should have stated this earlier). `compute metrics` does return a dictionary with the keys `'predictions', 'class_report' and 'target'`. It looks like one that gets output from `Trainer` has the prefix `'eval_'` in front of each key produced by `compute_metrics`. \r\n\r\nI can't find anywhere in our script or our custom dependencies where we print or log this dictionary, and when I suppress console output from `Trainer` then the problem stops. \r\n\r\n", "Is it just a matter of changing the log level? `run_ner.py` sets it to `INFO` for the main process (it does it twice - once for the root logger and another for transformers' logger:\r\n\r\nhttps://github.com/huggingface/transformers/blob/251eb70c979d74d3823e999236ff3621b07510a1/examples/token-classification/run_ner.py#L158-L168\r\n\r\n", "Maybe your `Trainer` ends up with a `PrinterCallback` that prints all the logs. You can remove this with\r\n```\r\nform transformers.trainer_callback import PrinterCallback\r\ntrainer.remove_callback(PrinterCallback)\r\n```", "Fixed! I had to upgrade to 4.0. Ended up having to upgrade to import the PrinterCalback, but I believe the upgrade itself fixed the problem. ", "Even better if the upgrade fixes the problem! There were printing statements in older versions indeed.", "Same issue here - How do I stop all prints coming from trainer.predict() ?" ]

[]

[]

[ "Could you show the entire code, including how you instantiate the model, as well as your environment information, as mentioned in the template? Thank you.", "that's ok?", "With the following code:\r\n\r\n```py\r\nfrom transformers import TFBertModel, BertTokenizer\r\nfrom time import time\r\nimport tensorflow as tf\r\n\r\nprint(\"GPU Available:\", tf.test.is_gpu_available())\r\n\r\ntokenizer = BertTokenizer.from_pretrained(\"bert-base-cased\")\r\nbert_model = TFBertModel.from_pretrained('bert-base-cased',output_hidden_states=True)\r\n\r\ninput = tokenizer(\"Hey is this slow?\" * 100 , max_length=512,padding=\"max_length\",truncation=True, return_tensors=\"tf\")\r\n\r\nfor i in range(100):\r\n start = time()\r\n outputs = bert_model(input)\r\n print(time() - start)\r\n```\r\n\r\nRunning on CPU doesn't increase the time for me:\r\n\r\n```\r\nGPU Available: False\r\n0.5066382884979248\r\n0.5038580894470215\r\n0.5125613212585449\r\n0.5018391609191895\r\n0.4927494525909424\r\n0.5066125392913818\r\n0.49803781509399414\r\n0.5140326023101807\r\n0.501518726348877\r\n0.49771928787231445\r\n0.5038976669311523\r\n```\r\n\r\n[Running on GPU, no problem either:](https://colab.research.google.com/drive/1tfynzpOiQJKkEi0vkpKaDXTwhB5-k6Td?usp=sharing)\r\n\r\n```\r\nGPU Available: True\r\n0.09349918365478516\r\n0.09653115272521973\r\n0.09893131256103516\r\n0.10591268539428711\r\n0.09297466278076172\r\n0.09105610847473145\r\n0.10088920593261719\r\n0.0935661792755127\r\n0.09639692306518555\r\n0.10130929946899414\r\n0.0947415828704834\r\n0.09380221366882324\r\n```", "Thanks. If I do exactly your code I observe an increasing time! Very strange, but I assume that this has something to do with the gpu memory not releasing?\r\n\r\n```\r\n0.07779383659362793\r\n0.20029330253601074\r\n0.2085282802581787\r\n0.22140789031982422\r\n0.23041844367980957\r\n0.22839117050170898\r\n0.23337340354919434\r\n0.22336935997009277\r\n0.22971582412719727\r\n0.22768259048461914\r\n0.22839140892028809\r\n0.22934865951538086\r\n0.23038363456726074\r\n0.22646212577819824\r\n0.23062443733215332\r\n0.22713351249694824\r\n0.24032235145568848\r\n0.24936795234680176\r\n0.24984216690063477\r\n0.2523007392883301\r\n0.2481672763824463\r\n0.2532966136932373\r\n0.24833273887634277\r\n0.2513241767883301\r\n0.2522923946380615\r\n0.2536492347717285\r\n0.25013017654418945\r\n0.25212621688842773\r\n0.24585843086242676\r\n0.25535058975219727\r\n0.2563152313232422\r\n0.2423419952392578\r\n0.6144394874572754\r\n0.647824764251709\r\n0.6494302749633789\r\n0.6406776905059814\r\n0.6507377624511719\r\n0.6411724090576172\r\n0.6513652801513672\r\n0.6484384536743164\r\n0.6489207744598389\r\n0.6405856609344482\r\n0.6493120193481445\r\n0.6484384536743164\r\n0.6372919082641602\r\n0.6494011878967285\r\n0.6433298587799072\r\n0.65077805519104\r\n0.6475985050201416\r\n0.6383304595947266\r\n0.6525297164916992\r\n0.6413178443908691\r\n0.6475212574005127\r\n0.6485188007354736\r\n0.64430832862854\r\n0.6478779315948486\r\n0.6457436084747314\r\n0.7288320064544678\r\n0.6573460102081299\r\n0.6572368144989014\r\n0.5861053466796875\r\n0.6324939727783203\r\n0.722456693649292\r\n0.6353938579559326\r\n0.6324222087860107\r\n0.6373186111450195\r\n0.6216456890106201\r\n0.6627655029296875\r\n0.7275354862213135\r\n0.6035926342010498\r\n0.6590445041656494\r\n0.5936176776885986\r\n0.6416335105895996\r\n0.6400752067565918\r\n1.1317992210388184\r\n1.2438006401062012\r\n1.2430295944213867\r\n1.2435650825500488\r\n1.2585129737854004\r\n1.2704930305480957\r\n1.2204067707061768\r\n1.2424969673156738\r\n1.2366819381713867\r\n1.2533769607543945\r\n1.2510595321655273\r\n1.2426464557647705\r\n1.2566087245941162\r\n1.2392685413360596\r\n```\r\n\r\nIf you don't mind, my issue regarding the usage for the input of the tokenizer is still open. :)\r\nhttps://github.com/huggingface/transformers/issues/7674", "Indeed, I'll try and check the issue ASAP. Thanks for the reminder!", "@LysandreJik . Thank you! But, do you have any idea for my issue? It seems it is a gpu issue?", "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! I believe this is so because this model uses the new weights system, which was introduced in v3.5.1. Please upgrade your transformers version to at least v3.5.1, we recommand the latest (v4.0.1):\r\n\r\n```\r\npip install -U transformers==4.0.1\r\n```", "@LysandreJik Thanks for the quick reply! It works now 👍 " ]

[]

[]

[ "Not sure, why test cases are failing for 'run_tests_tf'. @LysandreJik . \r\nLet me know any action require from my side.", "Can you rebase on master, a fix has been recently merged.", "The tests have already been fixed on `master`, merging! Thanks a lot @spatil6" ]

[]

[]

[ "After playing with it more and changing files, names etc... I managed to made it work with Roberta. I guess it was a stupid name error... Sorry for taking 5mn of you time reading this. \r\n \r\nI realize it couldn't work with distilbert (bert) as the tokenizers are differents. \r\nIn the end, the model is training. \r\n\r\nMaybe it will help someone else one day. \r\nHave a good day. ", "Glad you could resolve your issue!" ]

[]

[]

[ "Hi there. Please note that this script is not maintained anymore and is provided as is. We only maintain the `finetune_trainer.py` script now.", "Ok, I will switch to that one. Thank you" ]

[]

[]

[ "Hello! The main issue here is that your arrays are of shape `(seq_length)`, whereas they should be of shape `(batch_size, seq_length)`, even if the batch size is 1.\r\n\r\nUpdating your code to reflect that:\r\n\r\n```py\r\nfrom transformers import TFRobertaForSequenceClassification, RobertaConfig\r\nimport numpy as np\r\n\r\nbs = 1\r\nseq_len = 510\r\n\r\nclassifier = TFRobertaForSequenceClassification(RobertaConfig())\r\n\r\n#create random inputs for demo\r\ninput_ids = np.random.randint(0,10000, size=(bs, seq_len,))\r\nattention_mask = np.random.randint(0,2, size=(bs, seq_len,))\r\ntoken_type_ids = np.random.randint(0,2, size=(bs, seq_len,))\r\n\r\n#make a prediction with batch_size of 1\r\noutput = classifier.predict([input_ids, attention_mask, token_type_ids])\r\n\r\nprint(output.logits.shape) # -> outputs (1, 2)\r\n```\r\n\r\nHowever, there seems to be an error as the model cannot handle a sequence length of 512 when used this way. @jplu running the above code with a sequence length of 512 results in the following error:\r\n\r\n```\r\ntensorflow.python.framework.errors_impl.InvalidArgumentError: indices[0,510] = 512 is not in [0, 512)\r\n\t [[node tf_roberta_for_sequence_classification/roberta/embeddings/position_embeddings/embedding_lookup (defined at /home/lysandre/Workspaces/Python/transformers/src/transformers/models/roberta/modeling_tf_roberta.py:199) ]] [Op:__inference_predict_function_8030]\r\n\r\nErrors may have originated from an input operation.\r\nInput Source operations connected to node tf_roberta_for_sequence_classification/roberta/embeddings/position_embeddings/embedding_lookup:\r\n tf_roberta_for_sequence_classification/roberta/embeddings/add (defined at /home/lysandre/Workspaces/Python/transformers/src/transformers/models/roberta/modeling_tf_roberta.py:122)\r\n\r\nFunction call stack:\r\npredict_function\r\n```\r\n\r\nUsing a smaller sequence length doesn't raise the error. Do you mind weighing in on the issue?", "Yep, you are limited to 510 tokens + 2 extra tokens (beginning + end)", "After talking about it a bit offline with @jplu we realize there might be an issue with the `predict` method when passing in the values as a list. Could you try passing them as a dictionary instead?\r\n\r\nDoing this instead:\r\n\r\n```py\r\noutput = classifier.predict({\"input_ids\": input_ids, \"attention_mask\": attention_mask, \"token_type_ids\": token_type_ids})\r\n```", "Hei! Thank you for the feedback. I passed the parameters as a dict with everything else unchanged but still get the output as (seq_len, num_labels) unfortunately. ", "Can you try this:\r\n```\r\nfrom transformers import TFRobertaForSequenceClassification, RobertaConfig\r\nimport numpy as np\r\n\r\nbs = 1\r\nseq_len = 510\r\n\r\nclassifier = TFRobertaForSequenceClassification(RobertaConfig())\r\ninput_ids = np.random.randint(0,10000, size=(bs, seq_len,))\r\nattention_mask = np.random.randint(0,2, size=(bs, seq_len,))\r\ntoken_type_ids = np.zeros(shape=(bs, seq_len,))\r\nclassifier.predict({\"input_ids\": input_ids, \"attention_mask\": attention_mask, \"token_type_ids\": token_type_ids})\r\n```", "Yes, this works with seq_len = 510. It might help stating this behaviour in the docs or perhaps raise an error or show a warning when one tries to input an unbatched sample. Also a bit confusing that seq_len needs to be 510 and not 512 to account for the extra tokens (and the error received when one tries with 512 is a bit murky). Anyway, thanks for the help. I'll go ahead and close this." ]

[ { "id": 1834067346, "node_id": "MDU6TGFiZWwxODM0MDY3MzQ2", "url": "https://api.github.com/repos/huggingface/transformers/labels/Documentation", "name": "Documentation", "color": "77cc3b", "default": false, "description": "" } ]

[]

[]

[ { "id": 1834067346, "node_id": "MDU6TGFiZWwxODM0MDY3MzQ2", "url": "https://api.github.com/repos/huggingface/transformers/labels/Documentation", "name": "Documentation", "color": "77cc3b", "default": false, "description": "" } ]

[ { "login": "SBrandeis", "id": 33657802, "node_id": "MDQ6VXNlcjMzNjU3ODAy", "avatar_url": "https://avatars.githubusercontent.com/u/33657802?v=4", "gravatar_id": "", "url": "https://api.github.com/users/SBrandeis", "html_url": "https://github.com/SBrandeis", "followers_url": "https://api.github.com/users/SBrandeis/followers", "following_url": "https://api.github.com/users/SBrandeis/following{/other_user}", "gists_url": "https://api.github.com/users/SBrandeis/gists{/gist_id}", "starred_url": "https://api.github.com/users/SBrandeis/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/SBrandeis/subscriptions", "organizations_url": "https://api.github.com/users/SBrandeis/orgs", "repos_url": "https://api.github.com/users/SBrandeis/repos", "events_url": "https://api.github.com/users/SBrandeis/events{/privacy}", "received_events_url": "https://api.github.com/users/SBrandeis/received_events", "type": "User", "site_admin": false } ]

[]

[]

[]

[ "Hi @rabeehk ,\r\n\r\nIn `_load_optimizer_and_scheduler` if `model_path` exists and `optimizer` and `scheduler` `state_dict` is found then that means you are loading from a saved checkpoint and continue training from there, so the `lr` is read from the `scheduler` and used instead of the set LR. This is expected behaviour.", "Hi Suraj,\nthanks for the reply. I have a couple of questions on this, 1) I see this\nis ignoring the training epochs in when loading from the saved checkpoints,\nso it does not train for the epochs set, how could I resolve it? Also, if I\nwant to change the lr, could I load from checkpoint but change the lr?could\nyou give me some information how loading from trained optimizer, could\nhelp?\n\nto explain better, I train a model for X epochs, then I want to finetune it\non other datasets with extra Y epochs with different learning rate, for\nthis I pass the updated model to trainer, but then should I pass the\nmodel_path so it loads from the saved checkpoint of optimizer? and why this\nis ignoring the set number of epochs?\nthanks\n\nOn Mon, Dec 14, 2020 at 11:07 AM Suraj Patil <[email protected]>\nwrote:\n\n> Hi @rabeehk <https://github.com/rabeehk> ,\n>\n> In _load_optimizer_and_scheduler if model_path exists and optimizer and\n> scheduler state_dict is found then that means you are loading from a\n> saved checkpoint and continue training from there, so the lr is read from\n> the scheduler and used instead of the set LR. This is expected behaviour.\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/9099#issuecomment-744366176>,\n> or unsubscribe\n> <https://github.com/notifications/unsubscribe-auth/ABP4ZCDDDMJGNROAD3AKZFTSUXWVJANCNFSM4U2SCXGA>\n> .\n>\n", "If you want to fine-tune the saved checkpoint on another dataset then you could save it in diff path or remove the saved `optimizer` and `scheduler` files.\r\n\r\nAlso @sgugger might have a better answer here.", "Hi Suraj,\r\n\r\nI am having a similar problem here.\r\nWhen the trainer continues from a checkpoint, i.e. `trainer.train(check_point_path)`, I notice a peek in the learning curve. I suspect that is related to what Rabeeh has mentioned.\r\n\r\nPlease have a look at the learning curve I got after I had to resume the training twice.\r\n\r\n\r\n\r\nAny ideas?", "> this I pass the updated model to trainer, but then should I pass the model_path so it loads from the saved checkpoint of optimizer? and why this is ignoring the set number of epochs?\r\n\r\nPassing a `model_path` to the train method is done when you want to resume an interrupted training, which is why it does not do all epochs (it resumes the training from where you where). If you want to do a new training, you should not use that argument and manually pass the optimizer/scheduler you want to use at init.\r\n\r\n@abdullah-alnahas I have no idea what your plot is since you haven't told us how you generated it. ", "Thanks for your response @sgugger , and sorry for not making myself clear.\r\n\r\nI am training an Electra model from scratch using the [`Trainer`] API.(https://huggingface.co/transformers/main_classes/trainer.html). I have interrupted the trainer twice, then resumed training by `trainer.train(latest_checkpoint_path)`.\r\nAfter that, I have generated the learning curve plot from `{latest_checkpoint_path}/trainer_state.json`'s `log_history` using `step` as the x axis, and `loss` as the y axis.\r\n\r\nMy question: Is it normal that the learning curve peaks after resuming the training from a checkpoint after an interruption?", "The loss is reinitialized to 0 (it's not saved with the checkpoints) so it could come from this. There were also some recent changes in how the loss is logged so having your transformers version would help. The CI tests the final values of the weights of a (small) model are the same with a full training or resumed training, so I think this is just some weird reporting of the loss.", "thanks Suraj and everyone, makes sense not to initialize the optimizers.", "Hi @sgugger, \r\n\r\nI encountered the same issue on Transformers 4.3.0. I think the problem is not the loss being reinitialized as 0, but that the model is not being loaded from model_path. Only `TrainerState` is loaded but not the model weights. I looked through the code before concluding this, but as a sanity check, the current code will run even if `pytorch_model.bin` is not in the checkpoint directory, confirming that its not being loaded at all. It's odd that the CI tests are passing....\r\n\r\nAnyway I modified `trainer.py:train()`under the code block:\r\n\r\n```\r\n# Check if continuing training from a checkpoint\r\nif model_path and os.path.isfile(os.path.join(model_path, \"trainer_state.json\")):\r\n...\r\n self._globalstep_last_logged = self.state.global_step \r\n\r\n if isinstance(self.model, PreTrainedModel):\r\n model = model.from_pretrained(model_path)\r\n if not self.is_model_parallel:\r\n model = model.to(self.args.device)\r\n else:\r\n state_dict = torch.load(os.path.join(model_path, WEIGHTS_NAME))\r\n model.load_state_dict(state_dict)\r\n```\r\n\r\n`self._globalstep_last_logged = self.state.global_step` ensures the first logging of the loss is correct. `self._globalstep_last_logged` should not be 0 (that line is removed in the later part of the code)\r\n\r\nThe training is properly resumed after this. \r\n\r\n", "`Trainer` does not handle the reloading of the model indeed, which can be confusing. So l'll add that functionality this afternoon!", "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 @shaonanqinghuaizongshishi \r\n\r\nCould you please post the code snippet, stack trace and your env info so that we can take a look ?", "I am working on:\r\nubuntu 16.04\r\ntorch 1.5.0\r\ntransformers 3.4.0\r\n\r\n\r\n```\r\ndevice = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\r\ntokenizer = LayoutLMTokenizer.from_pretrained(model_path)\r\nmodel = LayoutLMForTokenClassification.from_pretrained(model_path, num_labels=config.num_labels).to(device)\r\noutputs = model(b_input_ids, bbox=b_boxes, token_type_ids=None,\r\n attention_mask=b_input_mask, labels=b_labels)\r\n```\r\n\r\n\r\nThen I run CUDA_LAUNCH_BLOCKING=1 python layoutLM.py, and got the following error:\r\n\r\n```\r\nTraceback (most recent call last):\r\n File \"layoutLM.py\", line 275, in <module>\r\n train(train_dataloader, validation_dataloader)\r\n File \"layoutLM.py\", line 162, in train\r\n attention_mask=b_input_mask, labels=b_labels)\r\n File \"/Classification/lib/python3.6/site-packages/torch/nn/modules/module.py\", line 550, in __call__\r\n result = self.forward(*input, **kwargs)\r\n File \"/Classification/lib/python3.6/site-packages/transformers/modeling_layoutlm.py\", line 864, in forward\r\n return_dict=return_dict,\r\n File \"/Classification/lib/python3.6/site-packages/torch/nn/modules/module.py\", line 550, in __call__\r\n result = self.forward(*input, **kwargs)\r\n File \"/Classification/lib/python3.6/site-packages/transformers/modeling_layoutlm.py\", line 701, in forward\r\n inputs_embeds=inputs_embeds,\r\n File \"/Classification/lib/python3.6/site-packages/torch/nn/modules/module.py\", line 550, in __call__\r\n result = self.forward(*input, **kwargs)\r\n File \"/Classification/lib/python3.6/site-packages/transformers/modeling_layoutlm.py\", line 118, in forward\r\n + token_type_embeddings\r\nRuntimeError: CUDA error: an illegal memory access was encountered\r\n```\r\n\r\n", "Hi there!\r\n\r\nI have been investigating the model by making [integration tests](https://github.com/NielsRogge/transformers/blob/e5431da34ab2d03d6114303f18fd70192c880913/tests/test_modeling_layoutlm.py#L318), and turns out it outputs the same tensors as the original repository on the same input data, so there are no issues (tested this both for the base model - `LayoutLMModel` as well as the models with heads on top - `LayoutLMForTokenClassification` and `LayoutLMForSequenceClassification`).\r\n\r\nHowever, the model is poorly documented in my opinion, I needed to first look at the original repository to understand everything. I made a demo notebook that showcases how to fine-tune HuggingFace's `LayoutLMForTokenClassification` on the FUNSD dataset (a sequence labeling task): https://github.com/NielsRogge/Transformers-Tutorials/blob/master/LayoutLM/Fine_tuning_LayoutLMForTokenClassification_on_FUNSD.ipynb\r\n\r\nLet me know if this helps you!\r\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." ]

[]

[]

[]

[]

[]

[ "Have you tried removing the `strict=False`, and instead specify `return_dict=False` when you initialize the model with `from_pretrained`? Can you let me know if this fixes your issue?", "> Have you tried removing the `strict=False`, and instead specify `return_dict=False` when you initialize the model with `from_pretrained`? Can you let me know if this fixes your issue?\r\n\r\nThanks. It seemed the error message is gone. However, I still receive the warning messages. Is there anyway I can modify the script and make it work without warning?", "Usually these do not impact the result, as they are python values that do not change over time. Have you seen an error in prediction?", "@LysandreJik Sounds good. Haven't seen anything wrong yet. :)", "I got this error too, when converting .safetensors to TorchScript. It's a model called ced( https://github.com/jimbozhang/hf_transformers_custom_model_ced.git )\r\n\r\n```\r\n device = torch.device(\"cpu\")\r\n example_input = torch.rand(1, 64, 301).to(device)\r\n model.eval() \r\n traced_script_module = torch.jit.trace(model, example_input)\r\n traced_script_module.save(\"ced_tiny_trace_torch_script.pt\")\r\n print('TorchScript trace saved')\r\n```\r\n\r\n\r\n```\r\nTraceback (most recent call last):\r\n File \"/Users/luis/PycharmProjects/hf_transformers_custom_model_ced/convert_pytorch_mobile.py\", line 72, in <module>\r\n convert2(model)\r\n File \"/Users/luis/PycharmProjects/hf_transformers_custom_model_ced/convert_pytorch_mobile.py\", line 44, in convert2\r\n traced_script_module = torch.jit.trace(model, example_input)\r\n File \"/Users/luis/PycharmProjects/hf_transformers_custom_model_ced/venv/lib/python3.9/site-packages/torch/jit/_trace.py\", line 794, in trace\r\n return trace_module(\r\n File \"/Users/luis/PycharmProjects/hf_transformers_custom_model_ced/venv/lib/python3.9/site-packages/torch/jit/_trace.py\", line 1056, in trace_module\r\n module._c._create_method_from_trace(\r\nRuntimeError: Encountering a dict at the output of the tracer might cause the trace to be incorrect, this is only valid if the container structure does not change based on the module's inputs. Consider using a constant container instead (e.g. for `list`, use a `tuple` instead. for `dict`, use a `NamedTuple` instead). If you absolutely need this and know the side effects, pass strict=False to trace() to allow this behavior.\r\n\r\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." ]

[]

[]

[ "Hey @adithyaan-creator, \r\n\r\nas the error message says you need to install the sentence piece library :-) \r\n\r\nIf you run:\r\n\r\n```\r\npip install sentencepiece==0.1.91\r\n```\r\n\r\nbefore, it should work.", "Thanks @patrickvonplaten . ", "Hi @patrickvonplaten \r\nI installed sentencepiece, but it still doesnt seem to be working for me. Please see the snapshot below. Please help.\r\n\r\n\r\n", "@DesiKeki try sentencepiece version 0.1.94.", "Thanks @adithyaan-creator , it worked!", "Hello @patrickvonplaten \r\nI have gone through the issue and the suggestions given above. However, I am facing the same issue and for some reason, none of the above solutions are proving fruitful.\r\n\r\nThe issue I am facing is exactly the same as the one stated above:\r\n`from transformers import T5Tokenizer,T5ForConditionalGeneration,Adafactor`\r\n`!pip install sentencepiece==0.1.91`\r\n`tokenizer = T5Tokenizer.from_pretrained(\"t5-base\")`\r\n`print(tokenizer)`\r\n\r\nThe output of the above code is: None. \r\nI tried using other versions of sentencepiece as well (as the one suggested above 0.1.94 and others as well). But it is still not working.\r\n\r\n\r\n", "Did you restart your kernel after installing `sentencepiece`? See conversation in https://github.com/huggingface/transformers/issues/10797", "> Did you restart your kernel after installing `sentencepiece`? See conversation in #10797\r\n\r\nit works for me, thank you", "> Did you restart your kernel after installing `sentencepiece`? See conversation in #10797\r\n\r\nIt works for me. Thanks a lot." ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "You should try to reduce the batch size. This will reduce the memory usage.", "Yup. What @LysandreJik said is correct. Use the following:\r\n`--per_device_train_batch_size x \\`\r\n`--per_device_eval_batch_size x \\`\r\nReplace x with your preferred batch size, I would recommend the highest power of 2 your GPU memory allows.", "It worked! With the Colab's GPU memory size of 12.72GB, the batch size worked at:\r\n\r\n`--per_device_train_batch_size 2 \\`\r\n`--per_device_eval_batch_size 16 \\`\r\n\r\nThanks for the quick response guys." ]

[]

[]

[ "@lhoestq - feel free to merge if you're ok with the PR" ]

[]

[]

[ "^C means you have hit Ctrl + C on your machine and stops the command running. You should re-run the command without hitting Ctrl + C.", "Yup I am aware that ^C is a halt command. I am running this on colab and I have tried to run this 5-7 times now, not hitting Ctrl+C once. For some reason it appears itself and halts the execution. ", "There might be something in colab that aborts bash command after some time then, or it happens when the session disconnects. But there is absolutely nothing in the script that triggers a cancel like this, so there is nothing we can do to fix this.\r\n\r\nNote that the scripts are not meant to be run on Colab, we have [notebook versions](https://github.com/huggingface/notebooks/tree/master/examples) of them for that.", "I think I have figured out the issue. This is happening because the dataset is large and when the full thing is loaded, colab crashes.", "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, thanks for opening an issue! We try to keep the github issues for bugs/feature requests, rather than help with training.\r\nCould you ask your question on the [forum](https://discusss.huggingface.co) instead?\r\n\r\nThanks!" ]

[]

[]

[ "Could you try with the flag `from_pt=True` when using `from_pretrained`? ", "That worked, thanks" ]

[]

[]

[ "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.", "Did anyone figure out how to do this?" ]

[]

[]

[ "Hello, thanks for opening an issue! We try to keep the github issues for bugs/feature requests, rather than help with training-related issues.\r\nCould you ask your question on the [forum](https://discusss.huggingface.co) instead? You'll have better answers over there.\r\n\r\nThanks!" ]

[]

[]

[]

[]

[]

[ "Very cool! Could you also enable the parallelization tests for these models? You can check how it was done in the initial model parallel PR, [here's the commit](https://github.com/huggingface/transformers/pull/8696/commits/cde47f0d110176d3834b736dac27bc9bc2a4de43) related to the tests. You can just add the `T5EncoderModel` to the `all_parallelizable_model_classes` attribute of the `T5ModelTester` class.", "> Very cool! Could you also enable the parallelization tests for these models? You can check how it was done in the initial model parallel PR, [here's the commit](https://github.com/huggingface/transformers/pull/8696/commits/cde47f0d110176d3834b736dac27bc9bc2a4de43) related to the tests. You can just add the `T5EncoderModel` to the `all_parallelizable_model_classes` attribute of the `T5ModelTester` class.\r\n\r\nThanks for the tip.\r\nDone, please let me know if anything else is needed from my side.", "Also it would be great if you could run `make style && make quality` or `make fixup` to solve the quality issues.", "> This LGTM. Looking into it it seems we have an error in `T5Stask` as it is creating the device map with `torch.cuda.device_count()`, rather than the `range` of that value like you're doing it here. Since we're always passing the device map to `T5Stack` (it's never used as a standalone model) we don't see it, but it doesn't seem correct.\r\n> \r\n> What do you think? If you think this is true, do you mind adding a `range` in `T5Stack` so that we can merge it together? Thanks!\r\n\r\nYes, you are correct, T5Stack should also use range. Since \"get_device_map\" function apply len to it .\r\nI have updated T5Stack using range.", "> Also it would be great if you could run `make style && make quality` or `make fixup` to solve the quality issues.\r\n\r\nDone and passed the code quality testing.", "Wonderful!" ]

[]

[]

[ "same problem here? what is going on because I run glue smoothly.it seems that the problem related to the script itself.", "The problem is in your JSON file. The squad v2 JSON file is not in a format the datasets library can directly preprocess, so you need to make it compliant with it. You should take this issue to the [`datasets`](https://github.com/huggingface/datasets) library and explain what your needs is.\r\n\r\nYou can also check the mock data file used in the [tests](https://github.com/huggingface/transformers/blob/master/tests/fixtures/tests_samples/SQUAD/sample.json) to see the expected format. A datasets expert would know better than me but I think the problem is that the squad JSON file has lists of dicts for the \"answers\" field when datasets expects a dictionary keys to list.", "The problem is not the JSON file that I have and I was able to solve it by using Transformers 3.x with no issues.", "Transformers v4 does not support training on SQuAD v2 via its example training script. For now, you have to use Transformers v3.", "Yes you could run it with the older script which was parsing the JSON differently. The new version uses the datasets library and requires the JSON to be organized differently (for compatibility with Arrow). ", "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." ]

[]

[]

[ "This repo(https://github.com/kifish/GPT4NLG) will help you out.", "Have you taken a look at the [`run_clm.py` script](https://github.com/huggingface/transformers/blob/master/examples/language-modeling/run_clm.py) in the examples? It seems to be doing exactly what you're looking for.\r\n\r\nErratum: I just realized that your link to `library` was actually a link to `run_clm.py`. What do you mean you didn't find the training script? Do you mean to say you would like an understandable guide on how to fine-tune GPT-2? Then [this guide](https://github.com/ncoop57/i-am-a-nerd/blob/master/_notebooks/2020-05-12-chatbot-part-1.ipynb) on DialoGPT, which is based off of the GPT-2 architecture, may be helpful.", "@kifish thank you for your reply, it is an interesting repository.\r\n\r\nI am doing this for the first time so I am looking for something more simple, although I got some ideas reading the code which I will use.", "@LysandreJik thank you , for the helpful tutorial.\r\n\r\nWhen I mentioned that I didn't find the training script, I meant that in line 327 of the `run_clm.py`, the `main` method calls the `train` method of a `Trainer`, but I couldn't find the code of the `train` method.\r\n\r\nIn addition, for the data preparation as input for the model, except for the extra tokens that I mentioned in my initial post, I thought to also add a `bos` and `eos` token at the begging and at the end of each text respectively, so that the model learns when a text starts and ends. The GPT-2 Tokenizer has already these tokens but they have the same id, they both have the id 50256. What is the reason behind this? \r\nIn order to deal with this, another way to prepare data is to use just the `eos` token to denote the end of a text, since the model should basically learn when a text ends. Can you please explain briefly?\r\n\r\nThank you in advance.\r\n\r\n", "The code for the `train` method is in the `Trainer` class that you can find [here](https://github.com/huggingface/transformers/blob/master/src/transformers/trainer.py#L565). It's a bit encapsulated so it might be a bit hard to follow - but we're working on simpler examples that show a basic PyTorch training loop and which should be out in a few months.\r\n\r\nYou may also find [this notebook interesting, which goes into finer detail on how to train a language model.](https://github.com/huggingface/notebooks/blob/master/examples/language_modeling.ipynb)\r\n\r\nRegarding your question on the extra tokens, adding a `bos` and `eos` token depends on how the model was pre-trained. BERT requires these as they were used during its pre-training. However, GPT-2 has very few special tokens during pre-training: a single `<|endoftext|>` token that was placed between sequences. I recommend you read the GPT-2 paper to get an idea of their pre-processing; we try to stay as close to the original implementation as possible.", "@LysandreJik thank you for your reply.\r\n\r\nI am writing the code for training GPT-2 and firstly, as you suggested, I concatenated the input texts separated by `<|endoftext|>` token and then split it into fixed lengths, which is the model input.\r\n\r\nIn the GPT-2 casual language modelling, the model input is also the labels of the model, so to compute the loss, the input and labels arguments of the loss function are the same. The [tutorial](https://github.com/huggingface/notebooks/blob/master/examples/language_modeling.ipynb) that you mentioned in the previous post, when references the model input and labels talks about a shifting on the left or on the right.\r\n\r\nTo calculate the loss correctly, are the model input and labels arguments of the loss function the same, or am I missing something?", "If you pass the labels to the model, you should pass exactly the same value as the input IDs. The model then shifts the label IDs and calculates the loss on its own.", "@LysandreJik I have a question regarding training. \r\n\r\nFor a Language Model which are indicative metrics to monitor training? Because for now I just consider the loss. For example, in classification it is usually considered the F1 score.\r\nAdditionally, in the tutorial that you shared, no early stopping was used in training, and the model was trained for three epochs. I have a small dataset of ~8K small texts (2MB). Do you have any suggestions on how to train the model, i.e. if there is no early stopping, how to decide when to stop training and how can I evaluate training?\r\n\r\nYou have been very helpful,\r\nregards.", "Hello @kifish \r\n\r\nthe GPT4NLG repo that you have shared back then was very helpful but I am no longer able to see it. Can you do anything about that?\r\n\r\nThank you in advance.", "> Hello @kifish\r\n> \r\n> the GPT4NLG repo that you have shared back then was very helpful but I am no longer able to see it. Can you do anything about that?\r\n> \r\n> Thank you in advance.\r\n\r\nhttps://github.com/kifish/GPT4NLG/tree/github", " https://github.com/kifish/GPT4NLG/tree/github\r\n\r\nThank you @kifish. If possible, let it visible for some days.", "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\r\nI traced the error and this is happening in this line:\r\n\r\nlabel_str = tokenizer.batch_decode(pred.label_ids, skip_special_tokens=True)\r\nhttps://github.com/rabeehk/debug-seq2seq/blob/1bcadb4b5497a0cbab6c2778e87335c5edcbd0a2/seq2seq/metrics/metrics.py#L99\r\n\r\nhere is the format of label_ids\r\n\r\n```\r\npred.label_ids [[10747 7 15 1]\r\n [10998 1 0 0]\r\n [10998 1 0 0]\r\n ...\r\n [10998 1 0 -100]\r\n [10998 1 0 -100]\r\n [10998 1 0 -100]] (3269, 4)\r\n\r\n```\r\ncould you please have a look? this is really blocking me, as T5 tokenizer fails for many datasets. thanks ", "I understood the issue now, previously boolq dataset had labels of 0/1 => max_decoding length of 3, now they changed it to True/False => max decoding length of 4, which causes the bug in my codes for decoding since max_decoding length was set to 3. this is solved now. thanks @lhoestq ", "Glad you resolved your issue." ]

[]

[]

[ "Are you looking for https://huggingface.co/transformers/model_summary.html ?", "@cronoik thanks for taking a look! Wasn't looking for that page, was trying to find just a general definition of what a transformer is in terms of the general concept", "Do you have something like this in mind: https://github.com/huggingface/transformers/blob/master/notebooks/02-transformers.ipynb ?", "No, wasn't looking for a notebook, just a one line description/explanation of what a transformer is\r\n\r\nHow would you describe it to someone who doesn't know about transformers?", "Maybe simply as self-attention based deep learning model architecture.", "@cronoik That's a good start, `self-attention` & `deep learning` aren't yet defined in the glossary\r\n\r\nHow would you define those?", "`self-attention`: each element of the input finds out which other elements of the input they should attend to.\r\n`deep learning`: machine learning algorithms which uses NN which several layers.\r\n\r\n@darigovresearch ", "@cronoik thanks for that! Would you like to put in a pull request so that your definitions go into the transformers glossary and the set of flashcards that we built or would you like us to do it?\n\nI'm sure those definitions would be welcome and easily merged by the maintainers\n\nhttps://www.darigovresearch.com/huggingface-transformers-glossary-flashcards", "Thanks @sgugger for merging the pull request, @cronoik your definitions are now on the glossary page and I have also added them to the flashcards so this issue can now be closed. Thank you both for your help!\n\nGlossary https://huggingface.co/transformers/glossary.html\n\nFlashcards https://www.darigovresearch.com/huggingface-transformers-glossary-flashcards", "Thanks for your commits ;)." ]

[]

[]

[ "Thanks for the suggestions! I've included them so should be good to go :)", "Thanks a lot!" ]

[]

[]

[ "Maybe @joeddav has an idea!", "The pipeline output is sorted from highest to lowest scores, so in your code `pred_idx` will always be `0` and `pred_cls` will always be `\"Single Patient\"`. Instead you want,\r\n\r\n```python\r\npred_cls = results['labels'][0]\r\npred_idx = labels.index(pred_cls)\r\n```", "Oh lol, I didn't know it was that simple xD. Thanks @joeddav that increased the accuracy to 73% (though less than online demo) which is good enough. Thank you so much!" ]

[]

[]

[ "Hi, could you provide the full error, as well as the command you use to launch the script? Thank you!", "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." ]

[]

[]

[ "have implemented that by myself.\r\nIf needed, I can make a pull request", "Hi, I'm also trying to utilize such a customized mask. Would you mind sharing your implementation? Thank you!" ]

[]

[]

[]

[ { "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" } ]

[]

[ "@patil-suraj this pull request can also be closed! \r\nMost of the typos were already fixed, the remaining ones were fixed in [this pull request](https://github.com/huggingface/transformers/pull/10989)\r\n\r\n" ]

[]

[]

[ "It is clear now - ready to document how to do it right: https://github.com/huggingface/transformers/pull/9070\r\n" ]

[]

[]

[]

[ { "id": 1990918270, "node_id": "MDU6TGFiZWwxOTkwOTE4Mjcw", "url": "https://api.github.com/repos/huggingface/transformers/labels/Good%20First%20Issue", "name": "Good First Issue", "color": "bbf794", "default": false, "description": "" }, { "id": 2392046359, "node_id": "MDU6TGFiZWwyMzkyMDQ2MzU5", "url": "https://api.github.com/repos/huggingface/transformers/labels/Good%20Second%20Issue", "name": "Good Second Issue", "color": "dd935a", "default": false, "description": "Issues that are more difficult to do than \"Good First\" issues - give it a try if you want!" } ]

[]

[ "I dont know if @MeRajat claimed this issue, However if not **I want to take this issue** ", "Usually one opens a PR to claim the issue (the PR does not have to be finished) - so I think it's still open. ", "@patrickvonplaten This FR is opened for some time,so thought of working on it, almost completed development, should I raise PR?\nAs @sadakmed is also working on it, so thought of asking.", "Hey @spatil6,\r\n\r\nI think the PR is already in an advanced stage, so I hope the PR is finished by next week. If not, I'll ping you again :-) " ]

[]

[]

[]

[]

[]

[ "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 should have addressed everybody's comments :)" ]

[ { "id": 1905493434, "node_id": "MDU6TGFiZWwxOTA1NDkzNDM0", "url": "https://api.github.com/repos/huggingface/transformers/labels/dependencies", "name": "dependencies", "color": "0366d6", "default": false, "description": "Pull requests that update a dependency file" } ]

[]

[]

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

[]

[ "After reading the paper again, I'm really excited to pre-train models for another domain 🤗 do you know when the pre-training code will be available 🤔", "@stefan-it glad to hear that you enjoyed our work. I haven't released the pre-training code yet as it is not as user friendly as I would want it to be. But it just happens that I'm planning to work on releasing a first version some time **this week**, so good timing 😊.\r\n\r\nYou can subscribe to the following issue if you want to be notified: https://github.com/helboukkouri/character-bert/issues/4\r\n\r\nCheers!", "Sounds great @helboukkouri! Let us know if we can help in any way, we'd love to see character BERT in `transformers`!", "Hey @helboukkouri , really cool PR for the upcoming model integration :hugs: \r\n\r\nI've already looked at it, and have a question about the `CharacterMapper` implementation. So in the current implementation it supports a maximum word length of 50 (so all word representations are padded to this length, if I'm correctly reading it). Do you think it would decrease training (and later fine-tuning) time, when using a smaller value :thinking: \r\n\r\nSo e.g. in German we could have really long words such as \"Bezirksschornsteinfegermeister\", but 50 is really long (but I think this is [language-dependend](https://arxiv.org/pdf/1207.2334.pdf)).", "Hey @stefan-it, thanks! 😊\r\n\r\n> Do you think it would decrease training (and later fine-tuning) time, when using a smaller value 🤔\r\n\r\nWhen we compute some stats around model speed, we find that while CharacterBERT is twice as slow as BERT during pre-training (108% slower), it is not as slow during downstream task fine-tuning (19% on avg.) This means that most of the \"slowness\" happens during pre-training, which makes us think that the Masked Language Modeling output layer is at fault here. In particular, the differences with BERT are: (1) no parameter sharing between the wordpiece embedding matrix and the output layer and (2) a larger output layer (we use top 100k tokens in the pre-training corpus as a vocabulary) since we want to be able to predict a reasonably high number of tokens so that MLM can be beneficial.\r\n\r\nSo to answer your question: reducing the maximum word length might reduce overall speeds but this change will probably negligible when compared to the effects listed above.\r\n\r\nYou may wonder why we used 50 character long representations. To be honest, we didn't want to tweak this `CharacterCNN` too much as it is originally the same layer that is used in ELMo. We just trusted the guys from AllenAI to have done a good work choosing the architecture and just re-used it 😄", "Hi @helboukkouri thanks for your detailed answer! This explains the whole training time/speed topic really great :hugs: ", "> After reading the paper again, I'm really excited to pre-train models for another domain 🤗 do you know when the pre-training code will be available 🤔\r\n\r\nCode is out! Feel free to open issues if you have any problems using it.", "Hi @helboukkouri, I have read the paper with great interest. I am currently working on the same topic. I tried to reproduce the result with our custom data. We could complete phase 1. Now we are heading towards fine-tuning of pretrained model for MLM and NSP tasks. Would you consider sharing research materials for the same. ", "Hi @pradeepsinghnitk, thanks for your interest.\r\n\r\nCould you be more specific about what you mean by `phase 1` and also if by `fine-tuning of pretrained model for MLM and NSP tasks` you mean pre-training or actual task-specific finetuning (e.g. on text classification tasks)?\r\n\r\nIn any case, check this code as it gives basic context for loading a model and running an inference. Fine-tuning it on any task should be straightforward (as you would with BERT basically) : https://github.com/helboukkouri/character-bert\r\n\r\nAnd for NSP and MLM (which is usually what is called `pre-training`), the code is here: https://github.com/helboukkouri/character-bert-pretraining\r\n\r\nUnfortunately, the import of CharacterBERT in the `transformers` library did not really succeed. It's been a while but if I remember well the issues were related to the different tests failing due to character-based tokenization being not very well supported at the time. \r\n\r\nI'll notify everybody if I ever go back to working on this again.\r\n\r\nCheers!", "Thank you for your response. \r\nTo be specific about phase 1; bash $WORKDIR/bash_scripts/run_pretraining.character_bert.step_1.sh (phase 1: maximum input length of 128 and maximum number of masked tokens per input of 20.) we could successfully execute this for char_bert pertaining and also for bert_based pretraining. \r\nNow, we would like to reproduce https://github.com/helboukkouri/character-bert-finetuning. But there was no code uploaded here. \r\n\r\n\r\n\"And for NSP and MLM (which is usually what is called pre-training), the code is here: https://github.com/helboukkouri/character-bert-pretraining\". this part of the scripts we have already executed\r\n", "Looking forward to this integration since December 2020!", "@stefan-it Hi Stefan, I saw it on your twitter account that you finished training German version of CharacterBERT. It is not on Huggingface yet, but I am writing my master thesis on OCR post correction on historical german corpus, and can really use it! Can you tell me how I can have access to your model? Thank you so much! Greetings from Stuttgart!", "Is it still not supported by transformers?" ]