Datasets:

---

DanielPFlorian

/

Transformers-Github-Issues

like 0

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

[]

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

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

[]

[ "What is ELnet model? The list of models that can be used for NER are: BERT, RoBERTa, DistilBERT (only for English text) and CamemBERT (only for French text).\r\n\r\n> ## Questions & Help\r\n> Hello everyone, when I try to use ELnet model for the NER task through run_ner.py, it shows the following problem:\r\n> \r\n> **init**() got an unexpected keyword argument 'do_lower_case'\r\n> \r\n> So is it some problem in the modeling_utils.py? Thanks for someone's response!", "> What is ELnet model? The list of models that can be used for NER are: BERT, RoBERTa, DistilBERT (only for English text) and CamemBERT (only for French text).\r\n> \r\n> > ## Questions & Help\r\n> > Hello everyone, when I try to use ELnet model for the NER task through run_ner.py, it shows the following problem:\r\n> > **init**() got an unexpected keyword argument 'do_lower_case'\r\n> > So is it some problem in the modeling_utils.py? Thanks for someone's response!\r\n\r\nI add XLNet in the run_ner.py:\r\nfrom transformers import XLNetConfig, XLNetTokenizer, XLNetForTokenClassification\r\nMODEL_CLASSES = {\r\n \"bert\": (BertConfig, BertForTokenClassification, BertTokenizer),\r\n \"roberta\": (RobertaConfig, RobertaForTokenClassification, RobertaTokenizer),\r\n \"distilbert\": (DistilBertConfig, DistilBertForTokenClassification, DistilBertTokenizer),\r\n \"camembert\": (CamembertConfig, CamembertForTokenClassification, CamembertTokenizer),\r\n \"xlnet\": (XLNetConfig, XLNetTokenizer, XLNetForTokenClassification),\r\n}", "> > What is ELnet model? The list of models that can be used for NER are: BERT, RoBERTa, DistilBERT (only for English text) and CamemBERT (only for French text).\r\n> > > ## Questions & Help\r\n> > > Hello everyone, when I try to use ELnet model for the NER task through run_ner.py, it shows the following problem:\r\n> > > **init**() got an unexpected keyword argument 'do_lower_case'\r\n> > > So is it some problem in the modeling_utils.py? Thanks for someone's response!\r\n> \r\n> I add XLNet in the run_ner.py:\r\n> from transformers import XLNetConfig, XLNetTokenizer, XLNetForTokenClassification\r\n> MODEL_CLASSES = {\r\n> \"bert\": (BertConfig, BertForTokenClassification, BertTokenizer),\r\n> \"roberta\": (RobertaConfig, RobertaForTokenClassification, RobertaTokenizer),\r\n> \"distilbert\": (DistilBertConfig, DistilBertForTokenClassification, DistilBertTokenizer),\r\n> \"camembert\": (CamembertConfig, CamembertForTokenClassification, CamembertTokenizer),\r\n> \"xlnet\": (XLNetConfig, XLNetTokenizer, XLNetForTokenClassification),\r\n> }\r\n\r\nDid you read #1592 and #2051 and similar?", "> > > What is ELnet model? The list of models that can be used for NER are: BERT, RoBERTa, DistilBERT (only for English text) and CamemBERT (only for French text).\r\n> > > > ## Questions & Help\r\n> > > > Hello everyone, when I try to use ELnet model for the NER task through run_ner.py, it shows the following problem:\r\n> > > > **init**() got an unexpected keyword argument 'do_lower_case'\r\n> > > > So is it some problem in the modeling_utils.py? Thanks for someone's response!\r\n> > \r\n> > \r\n> > I add XLNet in the run_ner.py:\r\n> > from transformers import XLNetConfig, XLNetTokenizer, XLNetForTokenClassification\r\n> > MODEL_CLASSES = {\r\n> > \"bert\": (BertConfig, BertForTokenClassification, BertTokenizer),\r\n> > \"roberta\": (RobertaConfig, RobertaForTokenClassification, RobertaTokenizer),\r\n> > \"distilbert\": (DistilBertConfig, DistilBertForTokenClassification, DistilBertTokenizer),\r\n> > \"camembert\": (CamembertConfig, CamembertForTokenClassification, CamembertTokenizer),\r\n> > \"xlnet\": (XLNetConfig, XLNetTokenizer, XLNetForTokenClassification),\r\n> > }\r\n> \r\n> Did you read #1592 and #2051 and similar?\r\n\r\nI just find it, thanks sir!", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

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

[]

[ "Hi, you're downloading one of the original implementation BERT models, which is in TensorFlow and you are trying to load it into one of our Pytorch models. \r\n\r\nYou can either download one of our checkpoints hosted on our S3 with:\r\n\r\n```py\r\nfrom transformers import BertForMaskedLM\r\n\r\nmodel = BertForMaskedLM.from_pretrained(\"bert-base-cased\")\r\n```\r\n\r\nThis model will now be available offline as it will be saved in your pytorch cache.\r\n\r\nOr you can convert the BERT model you downloaded to a checkpoint readable by our library by using the script [convert_bert_original_tf_checkpoint_to_pytorch.py](https://github.com/huggingface/transformers/blob/master/transformers/convert_bert_original_tf_checkpoint_to_pytorch.py)", "@LysandreJik thank you . \r\n**How to differentiate Bert tensor flow and pytorch models ?**\r\n\r\nI followed the instruction and create a PyTorch model using this pyhton code ->convert_bert_original_tf_checkpoint_to_pytorch.py\r\n\r\nINFO:transformers.modeling_bert:Initialize PyTorch weight ['cls', 'seq_relationship', 'output_weights']\r\nINFO:transformers.modeling_bert:Skipping cls/seq_relationship/output_weights/adam_m\r\nINFO:transformers.modeling_bert:Skipping cls/seq_relationship/output_weights/adam_v\r\nINFO:transformers.modeling_bert:Skipping global_step\r\nSave PyTorch model to /content/drive/My Drive/BMaskLang\r\n\r\nthe BMaskLang file was 402 MB size and it did not have any file extension ,now when i tired to load this pytorch model i get an error\r\n\r\n_from transformers import BertForMaskedLM\r\nmodel = BertForMaskedLM.from_pretrained(\"/content/drive/My Drive/BMaskLang\")\r\nError:\r\nUnicodeDecodeError: 'utf-8' codec can't decode byte 0x80 in position 0: invalid start byte_\r\n\r\n\r\nBasically what im trying to do is train a BertForMaskedLM on a custom corpus .\r\n\r\nwhat are the steps to train BertForMaskedLM model on custom corpus and \r\nAfter train model how to load it ?\r\nAfter loading how to test it on a new sentence ?\r\n\r\nFor example if there was a sentence in sample_text.txt corpus like\r\n\"He went to space.He brought a moon\"\r\n\r\nif i want to test my pretrained BertForMaskedLM to check if it correctly predicts the masked word in sentences\" He went to [Mask] .He brought a gallon [Mask]\r\n\r\nso the model must predict the same words which was in sample_text.txt corpus \"space\",\"moon\" rather than other words like \"store\",\"water\" since it was trained on this sample_text.txt corpus .im expecting this behavior .Is this possible to pretrain and build language model using transformers bert ?\r\n\r\n ", "> @LysandreJik thank you .\r\n> **How to differentiate Bert tensor flow and pytorch models ?**\r\n> \r\n> I followed the instruction and create a PyTorch model using this pyhton code ->convert_bert_original_tf_checkpoint_to_pytorch.py\r\n> \r\n> INFO:transformers.modeling_bert:Initialize PyTorch weight ['cls', 'seq_relationship', 'output_weights']\r\n> INFO:transformers.modeling_bert:Skipping cls/seq_relationship/output_weights/adam_m\r\n> INFO:transformers.modeling_bert:Skipping cls/seq_relationship/output_weights/adam_v\r\n> INFO:transformers.modeling_bert:Skipping global_step\r\n> Save PyTorch model to /content/drive/My Drive/BMaskLang\r\n> \r\n> the BMaskLang file was 402 MB size and it did not have any file extension ,now when i tired to load this pytorch model i get an error\r\n> \r\n> _from transformers import BertForMaskedLM model = BertForMaskedLM.from_pretrained(\"/content/drive/My Drive/BMaskLang\") Error: UnicodeDecodeError: 'utf-8' codec can't decode byte 0x80 in position 0: invalid start byte_\r\n> \r\n> Basically what im trying to do is train a BertForMaskedLM on a custom corpus .\r\n> \r\n> what are the steps to train BertForMaskedLM model on custom corpus and\r\n> After train model how to load it ?\r\n> After loading how to test it on a new sentence ?\r\n> \r\n> For example if there was a sentence in sample_text.txt corpus like\r\n> \"He went to space.He brought a moon\"\r\n> \r\n> if i want to test my pretrained BertForMaskedLM to check if it correctly predicts the masked word in sentences\" He went to [Mask] .He brought a gallon [Mask]\r\n> \r\n> so the model must predict the same words which was in sample_text.txt corpus \"space\",\"moon\" rather than other words like \"store\",\"water\" since it was trained on this sample_text.txt corpus .im expecting this behavior .Is this possible to pretrain and build language model using transformers bert ?\r\n\r\nI haved the same problem that how to load bert model yesterday. And now I found the solution. \r\n1. run convert_bert_original_tf_checkpoint_to_pytorch.py to create pytorch_model.bin\r\n2. rename bert_config.json to config.json\r\n\r\nafter that, the dictionary must have\r\n\r\nconfig.json (BertForMaskedLM.from_pretrained() need it)\r\npytorch_model.bin (BertForMaskedLM.from_pretrained() need it)\r\nvocab.txt (BertTokenizer.from_pretrained() need it)\r\n\r\npython version 3.7\r\npytorch version 1.3.1\r\ntensorflow version 2.0.0", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "> Hi, you're downloading one of the original implementation BERT models, which is in TensorFlow and you are trying to load it into one of our Pytorch models.\r\n> \r\n> You can either download one of our checkpoints hosted on our S3 with:\r\n> \r\n> ```python\r\n> from transformers import BertForMaskedLM\r\n> \r\n> model = BertForMaskedLM.from_pretrained(\"bert-base-cased\")\r\n> ```\r\n> \r\n> This model will now be available offline as it will be saved in your pytorch cache.\r\n> \r\n> Or you can convert the BERT model you downloaded to a checkpoint readable by our library by using the script [convert_bert_original_tf_checkpoint_to_pytorch.py](https://github.com/huggingface/transformers/blob/master/transformers/convert_bert_original_tf_checkpoint_to_pytorch.py)\r\n\r\nhi, I can't open this link https://github.com/huggingface/transformers/blob/master/transformers/convert_bert_original_tf_checkpoint_to_pytorch.py, where can I find this script? Thanks!", "It's [here](https://github.com/huggingface/transformers/blob/master/src/transformers/convert_bert_original_tf_checkpoint_to_pytorch.py)", "Traceback (most recent call last):\r\nFile \"/users/sroychou/BERT_text_summarisation/scripts/train_bert_summarizer.py\", line 12, in\r\nfrom metrics import optimizer, loss_function, label_smoothing, get_loss_and_accuracy, tf_write_summary, monitor_run\r\nFile \"/users/sroychou/BERT_text_summarisation/scripts/metrics.py\", line 16, in\r\n_, _, _ = b_score([\"I'm Batman\"], [\"I'm Spiderman\"], lang='en', model_type='bert-base-uncased')\r\nFile \"/users/sroychou/.local/lib/python3.7/site-packages/bert_score/score.py\", line 105, in score\r\ntokenizer = AutoTokenizer.from_pretrained(model_type)\r\nFile \"/users/sroychou/.local/lib/python3.7/site-packages/transformers/tokenization_auto.py\", line 298, in from_pretrained\r\nconfig = AutoConfig.from_pretrained(pretrained_model_name_or_path, **kwargs)\r\nFile \"/users/sroychou/.local/lib/python3.7/site-packages/transformers/configuration_auto.py\", line 330, in from_pretrained\r\nconfig_dict, _ = PretrainedConfig.get_config_dict(pretrained_model_name_or_path, **kwargs)\r\nFile \"/users/sroychou/.local/lib/python3.7/site-packages/transformers/configuration_utils.py\", line 382, in get_config_dict\r\nraise EnvironmentError(msg)\r\nOSError: Can't load config for 'bert-base-uncased'. Make sure that:\r\n\r\n'bert-base-uncased' is a correct model identifier listed on 'https://huggingface.co/models'\r\n\r\nor 'bert-base-uncased' is the correct path to a directory containing a config.json file" ]

[]

[]

[ "The code line that loads the BERT configuration is surely correct:\r\n```\r\n> config = transformers.BertConfig.from_json_file('./bertlm_model/config.json')\r\n```\r\nBut, for what concern the loading of a fine-tuned BERT model on a custom dataset, I think it's not correct the line you've used. Can you try with the following line suggested by me?\r\n```\r\n> from transformers import TFBertForSequenceClassification\r\n> model = TFBertForSequenceClassification.from_pretrained('bertlm_model', from_pt = True)\r\n```\r\n\r\nI suspect that it doesn't work however. **It's a PyTorch->TF 2.0 conversion problem**. It would be useful to understand that this bug occurs with _only_ BERT model or with _other_ models.\r\n\r\n> ## Bug\r\n> Model I am using (Bert, XLNet....): Bert\r\n> \r\n> Language I am using the model on (English, Chinese....): Multi-lingual\r\n> \r\n> The problem arise when using:\r\n> \r\n> * [x] the official example scripts: (give details)\r\n> * [ ] my own modified scripts: (give details)\r\n> \r\n> The tasks I am working on is:\r\n> \r\n> * [ ] an official GLUE/SQUaD task: (give the name)\r\n> * [x] my own task or dataset: (give details)\r\n> \r\n> ## Expected behavior\r\n> I have fine-tuned a language model using `run_lm_finetuning.py`.\r\n> \r\n> When trying to load it with TFBertForSequenceClassification however, it fails.\r\n> \r\n> ```\r\n> config = transformers.BertConfig.from_json_file('./bertlm_model/config.json')\r\n> model = transformers.TFBertForSequenceClassification.from_pretrained('./bertlm_model/', from_pt = True)\r\n> ```\r\n> \r\n> Showing the following error:\r\n> \r\n> ```\r\n> >>> model = transformers.TFBertForSequenceClassification.from_pretrained('./bertlm_model/', from_pt = True, config = config)\r\n> Traceback (most recent call last):\r\n> File \"<stdin>\", line 1, in <module>\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_utils.py\", line 288, in from_pretrained\r\n> return load_pytorch_checkpoint_in_tf2_model(model, resolved_archive_file)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 82, in load_pytorch_checkpoint_in_tf2_model\r\n> return load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs=tf_inputs, allow_missing_keys=allow_missing_keys)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 145, in load_pytorch_weights_in_tf2_model\r\n> assert name in pt_state_dict, \"{} not found in PyTorch model\".format(name)\r\n> AssertionError: classifier.weight not found in PyTorch model\r\n> ```\r\n> \r\n> If I try to run either `transformers.BertForSequenceClassification.from_pretrained('bertlm_model')` or `transformers.TFBertModel.from_pretrained('bertlm_model', from_pt = True)` all is fine!\r\n> \r\n> ## Environment\r\n> * OS: Ubuntu 18.04\r\n> * Python version: 3.7.5\r\n> * PyTorch version: 1.3.1\r\n> * Transformers version (or branch): Git repo master comit [0cb1638](https://github.com/huggingface/transformers/commit/0cb163865a4c761c226b151283309eedb2b1ca4d)\r\n> * Using GPU: Yes\r\n> * Distributed of parallel setup ?\r\n> * Any other relevant information:", "Thanks for your answer - unfortunately it didn't work..\r\n\r\nAs I'm fine-tuning the LM on bert-multilingual, I can't try it out with other models. However I have tried to load all the different BERT huggingface-sub-models using my fine-tuned language model and it seems it is only TFBertModel and TFBertForMaskedLM it will load?\r\n\r\nHope that can lead you in a direction?\r\n\r\n\r\n```\r\nimport transformers\r\nmodel_dir = 'bertlm_model/'\r\nconfig = transformers.BertConfig.from_json_file(model_dir + 'config.json')\r\n```\r\n\r\n### TFBertModel (works fine)\r\n```\r\n>>> model = transformers.TFBertModel.from_pretrained(model_dir, from_pt = True, config = config)\r\n>>> \r\n```\r\n\r\n### TFBertForPreTraining (won't load)\r\n```\r\n>>> model = transformers.TFBertForPreTraining.from_pretrained(model_dir, from_pt = True, config = config)\r\nTraceback (most recent call last):\r\n File \"<stdin>\", line 1, in <module>\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_utils.py\", line 288, in from_pretrained\r\n return load_pytorch_checkpoint_in_tf2_model(model, resolved_archive_file)\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 82, in load_pytorch_checkpoint_in_tf2_model\r\n return load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs=tf_inputs, allow_missing_keys=allow_missing_keys)\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 145, in load_pytorch_weights_in_tf2_model\r\n assert name in pt_state_dict, \"{} not found in PyTorch model\".format(name)\r\nAssertionError: cls.seq_relationship.weight not found in PyTorch model\r\n>>> \r\n```\r\n\r\n### TFBertForMaskedLM (works fine)\r\n```\r\n>>> model = transformers.TFBertForMaskedLM.from_pretrained(model_dir, from_pt = True, config = config)\r\n>>> \r\n```\r\n\r\n### TFBertForNextSentencePrediction (won't load)\r\n```\r\n>>> model = transformers.TFBertForNextSentencePrediction.from_pretrained(model_dir, from_pt = True, config = config)\r\nTraceback (most recent call last):\r\n File \"<stdin>\", line 1, in <module>\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_utils.py\", line 288, in from_pretrained\r\n return load_pytorch_checkpoint_in_tf2_model(model, resolved_archive_file)\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 82, in load_pytorch_checkpoint_in_tf2_model\r\n return load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs=tf_inputs, allow_missing_keys=allow_missing_keys)\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 145, in load_pytorch_weights_in_tf2_model\r\n assert name in pt_state_dict, \"{} not found in PyTorch model\".format(name)\r\nAssertionError: cls.seq_relationship.weight not found in PyTorch model\r\n>>> \r\n```\r\n\r\n### TFBertForSequenceClassification (won't load)\r\n```\r\n>>> model = transformers.TFBertForSequenceClassification.from_pretrained(model_dir, from_pt = True, config = config)\r\nTraceback (most recent call last):\r\n File \"<stdin>\", line 1, in <module>\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_utils.py\", line 288, in from_pretrained\r\n return load_pytorch_checkpoint_in_tf2_model(model, resolved_archive_file)\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 82, in load_pytorch_checkpoint_in_tf2_model\r\n return load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs=tf_inputs, allow_missing_keys=allow_missing_keys)\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 145, in load_pytorch_weights_in_tf2_model\r\n assert name in pt_state_dict, \"{} not found in PyTorch model\".format(name)\r\nAssertionError: classifier.weight not found in PyTorch model\r\n>>> \r\n```\r\n\r\n### TFBertForMultipleChoice (won't load)\r\n```\r\n>>> model = transformers.TFBertForMultipleChoice.from_pretrained(model_dir, from_pt = True, config = config)\r\nTraceback (most recent call last):\r\n File \"<stdin>\", line 1, in <module>\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_utils.py\", line 288, in from_pretrained\r\n return load_pytorch_checkpoint_in_tf2_model(model, resolved_archive_file)\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 82, in load_pytorch_checkpoint_in_tf2_model\r\n return load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs=tf_inputs, allow_missing_keys=allow_missing_keys)\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 109, in load_pytorch_weights_in_tf2_model\r\n tfo = tf_model(tf_inputs, training=False) # Make sure model is built\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/tensorflow_core/python/keras/engine/base_layer.py\", line 822, in __call__\r\n outputs = self.call(cast_inputs, *args, **kwargs)\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_bert.py\", line 943, in call\r\n seq_length = shape_list(input_ids)[2]\r\nIndexError: list index out of range\r\n>>> \r\n```\r\n\r\n### TFBertForTokenClassification (won't load)\r\n```\r\n>>> model = transformers.TFBertForTokenClassification.from_pretrained(model_dir, from_pt = True, config = config)\r\nTraceback (most recent call last):\r\n File \"<stdin>\", line 1, in <module>\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_utils.py\", line 288, in from_pretrained\r\n return load_pytorch_checkpoint_in_tf2_model(model, resolved_archive_file)\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 82, in load_pytorch_checkpoint_in_tf2_model\r\n return load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs=tf_inputs, allow_missing_keys=allow_missing_keys)\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 145, in load_pytorch_weights_in_tf2_model\r\n assert name in pt_state_dict, \"{} not found in PyTorch model\".format(name)\r\nAssertionError: classifier.weight not found in PyTorch model\r\n>>> \r\n```\r\n\r\n### TFBertForQuestionAnswering (won't load)\r\n```\r\n>>> model = transformers.TFBertForQuestionAnswering.from_pretrained(model_dir, from_pt = True, config = config)\r\nTraceback (most recent call last):\r\n File \"<stdin>\", line 1, in <module>\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_utils.py\", line 288, in from_pretrained\r\n return load_pytorch_checkpoint_in_tf2_model(model, resolved_archive_file)\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 82, in load_pytorch_checkpoint_in_tf2_model\r\n return load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs=tf_inputs, allow_missing_keys=allow_missing_keys)\r\n File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 145, in load_pytorch_weights_in_tf2_model\r\n assert name in pt_state_dict, \"{} not found in PyTorch model\".format(name)\r\nAssertionError: qa_outputs.weight not found in PyTorch model\r\n>>> \r\n```", "The same pattern of working (e.g. _TFBertForMaskedLM_) vs not working (e.g. _TFBertForQuestionAnswering_) appears also with the PyTorch version of these models? e.g. _BertForMaskedLM_\r\n\r\n> Thanks for your answer - unfortunately it didn't work..\r\n> \r\n> As I'm fine-tuning the LM on bert-multilingual, I can't try it out with other models. However I have tried to load all the different BERT huggingface-sub-models using my fine-tuned language model and it seems it is only TFBertModel and TFBertForMaskedLM it will load?\r\n> \r\n> Hope that can lead you in a direction?\r\n> \r\n> ```\r\n> import transformers\r\n> model_dir = 'bertlm_model/'\r\n> config = transformers.BertConfig.from_json_file(model_dir + 'config.json')\r\n> ```\r\n> \r\n> ### TFBertModel (works fine)\r\n> ```\r\n> >>> model = transformers.TFBertModel.from_pretrained(model_dir, from_pt = True, config = config)\r\n> >>> \r\n> ```\r\n> \r\n> ### TFBertForPreTraining (won't load)\r\n> ```\r\n> >>> model = transformers.TFBertForPreTraining.from_pretrained(model_dir, from_pt = True, config = config)\r\n> Traceback (most recent call last):\r\n> File \"<stdin>\", line 1, in <module>\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_utils.py\", line 288, in from_pretrained\r\n> return load_pytorch_checkpoint_in_tf2_model(model, resolved_archive_file)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 82, in load_pytorch_checkpoint_in_tf2_model\r\n> return load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs=tf_inputs, allow_missing_keys=allow_missing_keys)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 145, in load_pytorch_weights_in_tf2_model\r\n> assert name in pt_state_dict, \"{} not found in PyTorch model\".format(name)\r\n> AssertionError: cls.seq_relationship.weight not found in PyTorch model\r\n> >>> \r\n> ```\r\n> \r\n> ### TFBertForMaskedLM (works fine)\r\n> ```\r\n> >>> model = transformers.TFBertForMaskedLM.from_pretrained(model_dir, from_pt = True, config = config)\r\n> >>> \r\n> ```\r\n> \r\n> ### TFBertForNextSentencePrediction (won't load)\r\n> ```\r\n> >>> model = transformers.TFBertForNextSentencePrediction.from_pretrained(model_dir, from_pt = True, config = config)\r\n> Traceback (most recent call last):\r\n> File \"<stdin>\", line 1, in <module>\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_utils.py\", line 288, in from_pretrained\r\n> return load_pytorch_checkpoint_in_tf2_model(model, resolved_archive_file)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 82, in load_pytorch_checkpoint_in_tf2_model\r\n> return load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs=tf_inputs, allow_missing_keys=allow_missing_keys)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 145, in load_pytorch_weights_in_tf2_model\r\n> assert name in pt_state_dict, \"{} not found in PyTorch model\".format(name)\r\n> AssertionError: cls.seq_relationship.weight not found in PyTorch model\r\n> >>> \r\n> ```\r\n> \r\n> ### TFBertForSequenceClassification (won't load)\r\n> ```\r\n> >>> model = transformers.TFBertForSequenceClassification.from_pretrained(model_dir, from_pt = True, config = config)\r\n> Traceback (most recent call last):\r\n> File \"<stdin>\", line 1, in <module>\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_utils.py\", line 288, in from_pretrained\r\n> return load_pytorch_checkpoint_in_tf2_model(model, resolved_archive_file)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 82, in load_pytorch_checkpoint_in_tf2_model\r\n> return load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs=tf_inputs, allow_missing_keys=allow_missing_keys)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 145, in load_pytorch_weights_in_tf2_model\r\n> assert name in pt_state_dict, \"{} not found in PyTorch model\".format(name)\r\n> AssertionError: classifier.weight not found in PyTorch model\r\n> >>> \r\n> ```\r\n> \r\n> ### TFBertForMultipleChoice (won't load)\r\n> ```\r\n> >>> model = transformers.TFBertForMultipleChoice.from_pretrained(model_dir, from_pt = True, config = config)\r\n> Traceback (most recent call last):\r\n> File \"<stdin>\", line 1, in <module>\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_utils.py\", line 288, in from_pretrained\r\n> return load_pytorch_checkpoint_in_tf2_model(model, resolved_archive_file)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 82, in load_pytorch_checkpoint_in_tf2_model\r\n> return load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs=tf_inputs, allow_missing_keys=allow_missing_keys)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 109, in load_pytorch_weights_in_tf2_model\r\n> tfo = tf_model(tf_inputs, training=False) # Make sure model is built\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/tensorflow_core/python/keras/engine/base_layer.py\", line 822, in __call__\r\n> outputs = self.call(cast_inputs, *args, **kwargs)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_bert.py\", line 943, in call\r\n> seq_length = shape_list(input_ids)[2]\r\n> IndexError: list index out of range\r\n> >>> \r\n> ```\r\n> \r\n> ### TFBertForTokenClassification (won't load)\r\n> ```\r\n> >>> model = transformers.TFBertForTokenClassification.from_pretrained(model_dir, from_pt = True, config = config)\r\n> Traceback (most recent call last):\r\n> File \"<stdin>\", line 1, in <module>\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_utils.py\", line 288, in from_pretrained\r\n> return load_pytorch_checkpoint_in_tf2_model(model, resolved_archive_file)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 82, in load_pytorch_checkpoint_in_tf2_model\r\n> return load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs=tf_inputs, allow_missing_keys=allow_missing_keys)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 145, in load_pytorch_weights_in_tf2_model\r\n> assert name in pt_state_dict, \"{} not found in PyTorch model\".format(name)\r\n> AssertionError: classifier.weight not found in PyTorch model\r\n> >>> \r\n> ```\r\n> \r\n> ### TFBertForQuestionAnswering (won't load)\r\n> ```\r\n> >>> model = transformers.TFBertForQuestionAnswering.from_pretrained(model_dir, from_pt = True, config = config)\r\n> Traceback (most recent call last):\r\n> File \"<stdin>\", line 1, in <module>\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_utils.py\", line 288, in from_pretrained\r\n> return load_pytorch_checkpoint_in_tf2_model(model, resolved_archive_file)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 82, in load_pytorch_checkpoint_in_tf2_model\r\n> return load_pytorch_weights_in_tf2_model(tf_model, pt_state_dict, tf_inputs=tf_inputs, allow_missing_keys=allow_missing_keys)\r\n> File \"/home/myuser/.local/share/virtualenvs/BERTlm-giwpujkO/lib/python3.7/site-packages/transformers/modeling_tf_pytorch_utils.py\", line 145, in load_pytorch_weights_in_tf2_model\r\n> assert name in pt_state_dict, \"{} not found in PyTorch model\".format(name)\r\n> AssertionError: qa_outputs.weight not found in PyTorch model\r\n> >>> \r\n> ```", "All models load fine using the PyTorch version. So it is only some of the TF versions that are not working..\r\n\r\n\r\n```\r\n>>> model = transformers.BertModel.from_pretrained(model_dir, config = config)\r\n>>> model = transformers.BertForPreTraining.from_pretrained(model_dir, config = config)\r\n>>> model = transformers.BertForMaskedLM.from_pretrained(model_dir, config = config)\r\n>>> model = transformers.BertForNextSentencePrediction.from_pretrained(model_dir, config = config)\r\n>>> model = transformers.BertForSequenceClassification.from_pretrained(model_dir, config = config)\r\n>>> model = transformers.BertForMultipleChoice.from_pretrained(model_dir, config = config)\r\n>>> model = transformers.BertForTokenClassification.from_pretrained(model_dir, config = config)\r\n>>> model = transformers.BertForQuestionAnswering.from_pretrained(model_dir, config = config)\r\n>>> \r\n```", "Hello! If I understand correctly, you fine-tuned a BERT model with a language modeling head (`BertForMaskedLM`), which was then saved and now you're trying to load it in TensorFlow.\r\n\r\nYou can load it with `TFBertModel` and `TFBertForMaskedLM` as the weights are there, but can't load it in other architectures as some weights are lacking. In PyTorch you can load them but it randomly initializes the lacking weights.\r\n\r\nI believe we should have the same behavior between our TensorFlow models and our PyTorch models so I'll take a look at it. In the meantime, here's a workaround that will allow you to load the models in TensorFlow, for example from a `BertForMaskedLM` checkpoint to a `TFBertForSequenceClassification`:\r\n\r\n- Save the `BertForMaskedLM` checkpoint\r\n- Load it in `BertForSequenceClassification`\r\n- Save the checkpoint from `BertForSequenceClassification`\r\n- Load this checkpoint in `TFBertForSequenceClassification`\r\n\r\nHere's an example that will allow you to do that, make sure the directories exist :\r\n\r\n```py\r\nfrom transformers import BertForMaskedLM, BertForSequenceClassification, TFBertForSequenceClassification\r\n\r\n# This must have already been done by the script you used\r\nmodel = BertForMaskedLM.from_pretrained(\"bert-base-cased\")\r\nmodel.save_pretrained(\"here\")\r\n\r\n# Load the saved checkpoint in a PyTorch BertForSequenceClassification model and save it\r\nmodel = BertForSequenceClassification.from_pretrained(\"here\")\r\nmodel.save_pretrained(\"here-seq\")\r\n\r\n# Load the PyTorch model in the TF model of the same type\r\nTFBertForSequenceClassification.from_pretrained(\"here-seq\", from_pt=True)\r\n```", "Perfect - the workaround works - thanks a lot 👍 \r\n\r\nAnd yes, that is sort of the procedure I've used. However I did't run the BertForMaskedLM directly but instead used the run_lm_finetuning.py script to generate my fine-tuned LM:\r\n\r\n```\r\npython run_lm_finetuning.py \\\r\n --train_data_file=<pathToTrain.txt>\\\r\n --output_dir=bertlm_model \\\r\n --eval_data_file=<pathToTest.txt>\\\r\n --model_type=bert \\\r\n --model_name_or_path=bert-base-multilingual-cased \\\r\n --mlm \\\r\n --cache_dir=cache \\\r\n --do_train \\\r\n --do_eval \\\r\n --per_gpu_train_batch_size=8\\\r\n --per_gpu_eval_batch_size=8\r\n```\r\n\r\nAnd from there, I then try to load it with:\r\n```\r\nimport transformers\r\nmodel_dir = 'bertlm_model'\r\n\r\nconfig = transformers.BertConfig.from_json_file(model_dir + '/config.json')\r\nmodel = transformers.TFBertForSequenceClassification.from_pretrained(model_dir, from_pt = True, config = config)\r\n```\r\n" ]

[]

[]

[ "the step1 logits ：\r\nlogits tensor([[ 0.8831, -0.0368, -0.2206, -2.3484, -1.3595]], device='cuda:1',\r\n grad_fn=<AddmmBackward>)\r\nthe step1 loss：\r\ntensor(1.5489, device='cuda:1', grad_fn=NllLossBackward>)\r\nbut why can't loss.backward()?" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2107?src=pr&el=h1) Report\n> Merging [#2107](https://codecov.io/gh/huggingface/transformers/pull/2107?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/0cb163865a4c761c226b151283309eedb2b1ca4d?src=pr&el=desc) will **increase** coverage by `<.01%`.\n> The diff coverage is `100%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/2107?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #2107 +/- ##\n==========================================\n+ Coverage 82.67% 82.67% +<.01% \n==========================================\n Files 111 111 \n Lines 16162 16164 +2 \n==========================================\n+ Hits 13362 13364 +2 \n Misses 2800 2800\n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/2107?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [transformers/modeling\\_bert.py](https://codecov.io/gh/huggingface/transformers/pull/2107/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL21vZGVsaW5nX2JlcnQucHk=) | `87.72% <100%> (+0.04%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/2107?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2107?src=pr&el=footer). Last update [0cb1638...3520be7](https://codecov.io/gh/huggingface/transformers/pull/2107?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "👍 " ]

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

[]

[ "Hi, I've done a short study and I confirm the behavior you see.\r\nI've proposed a simple PR attached that gives interesting results and quite important speed improvement in any case.\r\nTo be discussed!", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

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

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2105?src=pr&el=h1) Report\n> Merging [#2105](https://codecov.io/gh/huggingface/transformers/pull/2105?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/0cb163865a4c761c226b151283309eedb2b1ca4d?src=pr&el=desc) will **not change** coverage.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/2105?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #2105 +/- ##\n=======================================\n Coverage 82.67% 82.67% \n=======================================\n Files 111 111 \n Lines 16162 16162 \n=======================================\n Hits 13362 13362 \n Misses 2800 2800\n```\n\n\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/2105?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2105?src=pr&el=footer). Last update [0cb1638...4757840](https://codecov.io/gh/huggingface/transformers/pull/2105?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

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

[]

[ "What GPU(s) and hyperparameters are you using?\r\n\r\nSpecifically:\r\n--learning_rate ?\r\n--per_gpu_train_batch_size ?\r\n--gradient_accumulation_steps ?\r\n--warmup_steps ?\r\n\r\nI'm on my third xxlarge-v1 fine-tune, ~23 hours each epoch plus eval on 2x NVIDIA 1080Ti. Results are relatively good, best of all the models I've fine-tuned on SQuAD 2.0 so far:\r\n```\r\nalbert_xxlargev1_squad2_512_bs32:\r\n{\r\n \"exact\": 83.67725090541565,\r\n \"f1\": 87.51235434089064,\r\n \"total\": 11873,\r\n \"HasAns_exact\": 81.86572199730094,\r\n \"HasAns_f1\": 89.54692697189559,\r\n \"HasAns_total\": 5928,\r\n \"NoAns_exact\": 85.48359966358284,\r\n \"NoAns_f1\": 85.48359966358284,\r\n \"NoAns_total\": 5945\r\n}\r\n```\r\n\r\n\r\n", "I use 6xP40 for xlarge-v2 and 4xP40 for large-v2 with a same total batch size of 48 (8x6 & 12x4), lr is set to 3e-5 for all the runs. Other options remain default.\r\n\r\nI also launched several runs with same setting, sometimes the problem happened but sometimes didn't, this is weird because I didn't even change the random seed. ", "I meant to include this link in my post above, which details the Google-Research (GR) `run_squad_sp.py` hyperparameters: #https://github.com/huggingface/transformers/issues/1974\r\n\r\nAs demonstrated and referenced in my link, GR's bs=32 was a very slight improvement for me over my initial bs=48 fine-tune as you also chose. Peak learning_rate=5e-5 after a 10% linear lr warm-up proportion and linear lr decay after that.\r\n\r\nHope this helps, please post your results for comparison.", "From tensorboard, the best-performed one is albert-xxlarge-v2 with 88.49 F1 and 84.83 EM at step 25k. I didn't run any experiment on v1 models", "> From tensorboard, the best-performed one is albert-xxlarge-v2 with 88.49 F1 and 84.83 EM at step 25k. I didn't run any experiment on v1 models\r\n\r\nNice results, 6 epochs?\r\n\r\nAccording to GR at the time of V2 release, the xxlarge-V1 model outperforms the xxlarge-V2 model.", "Not sure if this is related, but I found that ALBERT is very unstable. When running in non-deterministic mode, it will sometimes get stuck in a very strange spot and never recover. This becomes very clear when you use a secondary score as a sanity check (e.g. Pearson correlation for regression, f1 for classification). So for the exact same parameters (but each time presumably another random seed), I would sometimes get e.g. `r=0.02` and other times `r=0.77`. \r\n\r\nI'd have to test more to get conclusive results, but it's something that I haven't experienced before with other models.", "The best I can get with xxlarge-v2 is\r\n`\r\nResults: {'exact': 84.86481933799377, 'f1': 88.43795242530017, 'total': 11873, 'HasAns_exact': 82.05128205128206, 'HasAns_f1': 89.20779506504576, 'HasAns_total': 5928, 'NoAns_exact': 87. 67031118587047, 'NoAns_f1': 87.67031118587047, 'NoAns_total': 5945, 'best_exact': 84.86481933799377, 'best_exact_thresh': 0.0, 'best_f1': 88.4379524253, 'best_f1_thresh': 0.0}\r\n`\r\nwith 2e-5 lr, 4xV100, 2 samples per GPU, no gradient accumulation, and ran for 3 epochs.\r\nThe current results are pretty about the same with Roberta large, but I expect better performance from ALBERT.\r\nStill tuning. Any idea on how to improve it? ", "Same issue with `albert-large-v2` but don't know why. Any result?", "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" ]

[]

[]

[ "Did you try replacing all meaningful whitespaces by a special token `<space>` and just add this new token to the tokenizer and train your model with it?", "> Did you try replacing all meaningful whitespaces by a special token `<space>` and just add this new token to the tokenizer and train your model with it?\r\n\r\nI guess it will be a bit tricky to define the \"meaningful\" whitespaces, but I will give it a shot. Thanks :)" ]

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

[]

[ "Any ideas on whether this will be included sooner or later?", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2101?src=pr&el=h1) Report\n> Merging [#2101](https://codecov.io/gh/huggingface/transformers/pull/2101?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/0cb163865a4c761c226b151283309eedb2b1ca4d?src=pr&el=desc) will **decrease** coverage by `2.58%`.\n> The diff coverage is `19.23%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/2101?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #2101 +/- ##\n==========================================\n- Coverage 82.67% 80.08% -2.59% \n==========================================\n Files 111 112 +1 \n Lines 16162 16874 +712 \n==========================================\n+ Hits 13362 13514 +152 \n- Misses 2800 3360 +560\n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/2101?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [transformers/tests/tokenization\\_bert\\_test.py](https://codecov.io/gh/huggingface/transformers/pull/2101/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL3Rva2VuaXphdGlvbl9iZXJ0X3Rlc3QucHk=) | `89.47% <ø> (ø)` | :arrow_up: |\n| [transformers/modeling\\_auto.py](https://codecov.io/gh/huggingface/transformers/pull/2101/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL21vZGVsaW5nX2F1dG8ucHk=) | `29.89% <ø> (-0.72%)` | :arrow_down: |\n| [transformers/tests/tokenization\\_gpt2\\_test.py](https://codecov.io/gh/huggingface/transformers/pull/2101/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL3Rva2VuaXphdGlvbl9ncHQyX3Rlc3QucHk=) | `97.43% <ø> (ø)` | :arrow_up: |\n| [transformers/data/metrics/squad\\_metrics.py](https://codecov.io/gh/huggingface/transformers/pull/2101/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL2RhdGEvbWV0cmljcy9zcXVhZF9tZXRyaWNzLnB5) | `0% <0%> (ø)` | |\n| [transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/2101/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL21vZGVsaW5nX3V0aWxzLnB5) | `90.86% <100%> (ø)` | :arrow_up: |\n| [transformers/tokenization\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/2101/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rva2VuaXphdGlvbl91dGlscy5weQ==) | `91.02% <100%> (+0.55%)` | :arrow_up: |\n| [transformers/tests/tokenization\\_tests\\_commons.py](https://codecov.io/gh/huggingface/transformers/pull/2101/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL3Rva2VuaXphdGlvbl90ZXN0c19jb21tb25zLnB5) | `100% <100%> (ø)` | :arrow_up: |\n| [transformers/tokenization\\_xlm.py](https://codecov.io/gh/huggingface/transformers/pull/2101/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rva2VuaXphdGlvbl94bG0ucHk=) | `83.46% <100%> (+0.13%)` | :arrow_up: |\n| [transformers/data/\\_\\_init\\_\\_.py](https://codecov.io/gh/huggingface/transformers/pull/2101/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL2RhdGEvX19pbml0X18ucHk=) | `100% <100%> (ø)` | :arrow_up: |\n| [transformers/tokenization\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/2101/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rva2VuaXphdGlvbl94bG5ldC5weQ==) | `90.4% <100%> (+0.15%)` | :arrow_up: |\n| ... and [7 more](https://codecov.io/gh/huggingface/transformers/pull/2101/diff?src=pr&el=tree-more) | |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/2101?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2101?src=pr&el=footer). Last update [0cb1638...35737ea](https://codecov.io/gh/huggingface/transformers/pull/2101?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Hi @mandubian, thanks for opening a PR to fix this. I think this raises an issue when there are two new tokens added which are right after each other, as spaces get added before and after the tokens. Here's an example of the issue, with your change and based on #2096:\r\n\r\n```py\r\nfrom transformers import BertTokenizer\r\nbert_tokenizer = BertTokenizer.from_pretrained('bert-base-cased')\r\n\r\nbert_tokenizer.add_tokens(['[ENT]', '[TEN]'])\r\nprint(len(bert_tokenizer))\r\n\r\nx = bert_tokenizer.encode(\"you are the [ENT] [TEN] with [ENT] and [ENT]\")\r\n\r\nprint(bert_tokenizer.decode(x))\r\n# outputs: [CLS] you are the [ENT] [TEN] with [ENT] and [ENT] [SEP]\r\n# with two spaces ----------------^^\r\n```", "You're right, I hadn't thought about the case of 2 consecutive tokens. Let's try to make it better.", "@LysandreJik I've pushed a new version for discussion.\r\n\r\nAdded tokens aren't prepended with space anymore but subtexts are. I've considered different solutions but none is perfect and a compromise has to be made. \r\n\r\nIn Bert tokenizer, `convert_tokens_to_string` joins with space between sub-strings (not in GPT2 tokenizer) but then those sub-strings and added tokens need to be separated also by spaces. So I choose to remove the space before added tokens and add a space in subtexts join so that there are always spaces.\r\n\r\nBut it can add spaces where there weren't. With Bert Tokenizer, if you have `[ABC] toto tata [DEF] [GHI]`, `decode.encode` returns the same string (except lower case). But when you have less spaces `[ABC]toto tata [DEF][GHI]`, `decode.encode` returns `[ABC] toto tata [DEF] [GHI]` with more spaces.\r\nFor GPT2, it's the same, it doesn't respect all spaces from input.\r\n\r\nI've added a test in `tokenizer_bert_test` and `tokenizer_gpt2_test` but it's not so good as it must be implemented for all tokenizers.\r\n\r\nDon't hesitate to give more ideas, it's just a proposition on this quite stupid issue quite far from models 😄 ", "This seems to work, thanks @mandubian! I pushed another commit on your fork to test every tokenizer + rebase on master.", "@LysandreJik (I've deleted my previous message from tonight, I had misread your message on mobile :D) Just to know, don't you squash commits in general?" ]

[]

[]

[ "DistilBERT as any BERT is a Transformer encoder so it encodes a sequence of tokens into a vector in the embedding space. It doesn't return a sequence of tokens.\r\n\r\nThe output of the model is `return output # last-layer hidden-state, (all hidden_states), (all attentions)` https://github.com/huggingface/transformers/blob/master/transformers/modeling_distilbert.py#L484.\r\n\r\nIf you check the size of this hidden-state, it is `torch.Size([1, 4, 768])`.\r\n`768` being the size of the hidden-state ie the size of the embedding vector.\r\n`4` is the number of token in input sequence (`[CLS]` and `[SEP]` tokens are added by tokenizer)\r\n`encode` is meant to return a sequence of token from a sequence of words.\r\n`decode` is meant to return a sequence of words from a sequence of tokens.\r\n\r\nSo if you do:\r\n\r\n```python\r\nencoded = tokenizer.encode('Some text')\r\n# encoded: [101, 2070, 3793, 102]\r\ndecoded = tokenizer.decode(encoded))\r\n# decoded: [CLS] some text [SEP]\r\n```\r\n\r\nBut, you can't use `decode` on the output of the model as it's not a sequence of tokens but an embedding vector.\r\n\r\nDon't hesitate to ask question if my explanation isn't clear.", "Hmh, I'll try to re-phrase my question because your answer did not clear up any of my confusion:\r\n\r\nGiven the output of any hugging face model (e.g. the ones with a language modeling head, take for example `GPT2LMHeadModel`), how does one actually go from the model's output to words ?\r\n", "Hi, you can read the [quickstart](https://huggingface.co/transformers/quickstart.html#openai-gpt-2) of the documentation to see how to use `GPT2LMHeadModel` with the decoding method.", "@George3d6 you can decode if the output of your model has the size of the vocabulary. So you need an output head that convert the hidden-size of the encoder or decoder into the vocabulary size. For `GPT2LMHeadModel`, follow what Lysandre said.", "thank you all for your explanations here. I can do that with the GPT2 models with no issues, but my issue is that now I want to do the same but with the smaller simpler DistilmBERT model which is also multilingual in 104 languages, so I want to generate text in for example Spanish and English and with this lighter model. So I do this:\r\n\r\ntokenizer = DistilBertTokenizer.from_pretrained('distilbert-base-multilingual-cased')\r\nmodel = DistilBertForMaskedLM.from_pretrained('distilbert-base-multilingual-cased')\r\ninput_ids = torch.tensor(tokenizer.encode(\"Hello, my dog is cute\", add_special_tokens=True)).unsqueeze(0) # Batch size 1\r\noutputs = model(input_ids, masked_lm_labels=input_ids)\r\nloss, prediction_scores = outputs[:2]\r\n\r\nbut now, how do I get the continuation of the phrase at that point? I tried to apply tokenizer.decode with no luck there, thank you", "This issue is quite old. But for those who still looking for the answer, you should load the model with LM Head instead.\r\nfor BERT is BertForMaskedLM.from_pretrained()\r\nfor DistilBERT is DistilBertForMaskedLM.from_pretrained()\r\n\r\nThe tokenizer is the same.", "> Hi, you can read the [quickstart](https://huggingface.co/transformers/quickstart.html#openai-gpt-2) of the documentation to see how to use `GPT2LMHeadModel` with the decoding method.\r\n\r\n@LysandreJik This reply is outdated... The link, and its redirection link send you to a 404 error.", "> This issue is quite old. But for those who still looking for the answer, you should load the model with LM Head instead. for BERT is BertForMaskedLM.from_pretrained() for DistilBERT is DistilBertForMaskedLM.from_pretrained()\r\n> \r\n> The tokenizer is the same.\r\n\r\nI've tried `model_MaskedLM = AutoModelForMaskedLM.from_pretrained(checkpoint)` but it doesn't work, since the model still returns logits (or without the head, embeddings) which are not lists of token ids...\r\nHence, we can't simply use decode on the output of that `ForMaskedLM` model.\r\nIt seems that the checkpoint must be compatible with the model's `generate` method, which will output a list of token ids lists.\r\n\r\n", "I want decode the soft prompt embedding to sequence of words. I have the output logit from BertMaskedLm head. the output logits shape is (batch_Size, num_tokens,3,vacab_size). How do I do that? " ]

[]

[]

[ "Can you give some more information? It's not clear what you mean by \"score\". The special classification token for RoBERTa is `<s>`.", "Thanks!", "@tzhxs If that's everything you need, please close this topic.", "ok" ]

[]

[]

[ "There are different kinds of models.\r\nBut as you talk about MLM, you might be talking about BERT-like models.\r\nBERT is based on a transformer encoder so by definition of transformer, it takes a sequence of tokens (a token is just an encoding of each word into a vocabulary of known size) and returns a sequence of vectors in an embedding space: each token of the input sequence has a representation in embedding space (768 or 1024 are the common sizes called hidden-sizes in general).\r\n\r\nNaturally, this embedding is specialized on a given task by adding one or several heads after the encoder. the MLMHead is one of those heads. NextSentencePredictionHead is another one.\r\n\r\nIf you look at BERT code in MLM Mode, it uses the following head:\r\n\r\n```python\r\nclass BertLMPredictionHead(nn.Module):\r\n def __init__(self, config):\r\n super(BertLMPredictionHead, self).__init__()\r\n self.transform = BertPredictionHeadTransform(config)\r\n\r\n # The output weights are the same as the input embeddings, but there is\r\n # an output-only bias for each token.\r\n self.decoder = nn.Linear(config.hidden_size,\r\n config.vocab_size,\r\n bias=False)\r\n\r\n self.bias = nn.Parameter(torch.zeros(config.vocab_size))\r\n\r\n def forward(self, hidden_states):\r\n hidden_states = self.transform(hidden_states)\r\n hidden_states = self.decoder(hidden_states) + self.bias\r\n return hidden_states\r\n```\r\n\r\nYou see here that the output of the model is passed through a simple `nn.Linear(config.hidden_size, config.vocab_size)` converting the embedding vector of size `hidden_size` into a vector of `vocab_size` (vocabulary size). Then you can softmax that into a vector of probability on the whole vocabulary and use argmax to get the most probable token.\r\n\r\nSo for other models, it really depends on the head used. If it is the NextSentencePrediction head, it just classifies the embedding vector into binary true/false so you lose the probabilities on the vocabulary.\r\n\r\nDoes it answer to your questions?", "@mandubian Aha, that last Linear layer was what I was missing. I didn't quite understand how one could get there from simply the output of the BertModel itself (i.e. the encoder). \r\n\r\nI do have one more question, though. The parameters for BertLMPredictionHead are not pretrained, right? Would one still need to finetune that head on a huge dataset? (In particular I'm interested in RoBERTa, but I assume that it works similar to BERT.) More concretely, if I wanted to do inference and just get a token's probability (e.g. 'cookies' in the sentence 'I like cookies'), how could I do that?\r\n\r\nThanks for your time and input!", "BERT is pretrained on MLM and NSP heads and provided by transformers as is. ROBERTA is pretrained on MLM only.\r\nCheck there https://huggingface.co/transformers/model_doc/bert.html#bertforpretraining, you should find what you need in outputs ;)\r\n\r\n", "How did I not see this?! I'm blind. Well, perhaps I should read the documentation rather than the source code sometimes. To be honest, I didn't even know that that documentation existed!\r\n\r\nSo I see\r\n\r\n> **prediction_scores**: torch.FloatTensor of shape (batch_size, sequence_length, config.vocab_size)\r\n> Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).\r\n\r\nSo what this actually returns (if you'd apply SoftMax) is the probability for the whole vocabulary. In other words, a probability across the vocabulary which should then sum to one. So what you'd need to do, then is to find the ID of the input token in the vocabulary, and then get that ID from the output of prediction_scores. But how then can you deal with subword units? If an input token is split, how can you then recover the original token and its output?\r\n\r\nI'm sorry for the flood of questions, but it seems like a snowball effect; every answer results in new questions.\r\n\r\nSeems like the example here is useful but I'll have to dig deeper to understand it completely.\r\n\r\nhttps://github.com/huggingface/transformers/blob/0cb163865a4c761c226b151283309eedb2b1ca4d/examples/utils_squad.py#L803-L808", "No worry.\r\nYes wordpiece tokenizer used in Bert (and BPE in GPT2) can cut a word into several tokens. With WordPiece tokenizer, it prefixes `##` to sub-word tokens. Check that code https://github.com/huggingface/transformers/blob/master/transformers/tokenization_bert.py#L419-L421\r\n\r\nWhen decoding, it's very basic, it just convert back to pieces of strings and re-concatenate them by removing ` ##` in this function for example https://github.com/huggingface/transformers/blob/master/transformers/tokenization_bert.py#L191-L194.\r\n\r\nThe rest can be found in papers, doc or code.", "I understand the encoding/decoding process, but I don't quite understand how you can keep track of the subword units in the model.\r\n\r\nLet's take an example \"I like granola.\", which can be encoded as \"[CLS] i like gran ##ola . [SEP]\" in e.g. BERT. This means we'll get seven output tokens. Here, if we want to get the output values for the input word 'granola' we need index 3 and 4. My question then is, is there a way to keep track of this dynamically/automatically? In other words, a mapping from the input words to the tokenized words, so you can go back and see that granola was split into tokens at indices position 3 and 4.", "In transformers models, AFAIK, there is no tracking of token <-> indices in original text (I can be wrong). In the cases I know, it's just using `##xyz` to mean it's a sub-word token belonging to word. In decoding, final erasure of ` ##` and concatenation rebuilds words.\r\n\r\nYet, in examples/run_squad.py, there might be what you need as it seems to keep track of mapping between tokens and index original doc. Have a look at it, you'll see it's not trivial ;)" ]

[]

[]

[ "Please close this. It's a duplicated of your other question." ]

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

[]

[ "I think you can keep the issue open, this is a bug that should be fixed.", "This could be related, I'm on commit d46147294852694d1dc701c72b9053ff2e726265\r\n\r\nIt's strange that the id for \"student\" changed after adding the special token\r\n", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

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

[]

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

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

[]

[ "Hello! If you try to load your `pytorch_model.bin` directly in `BertForSequenceClassification`, you'll indeed get an error as the model won't know that it is supposed to have three classes. That's what the configuration is for!\r\n\r\nI guess you're doing something similar to this:\r\n\r\n```py\r\nfrom transformers import BertForSequenceClassification\r\n\r\nmodel = BertForSequenceClassification.from_pretrained(\"bert-base-cased\")\r\nmodel.load_state_dict(torch.load(\"SAVED_SST_MODEL_DIR/pytorch_model.bin\"))\r\n# Crashes here\r\n```\r\n\r\nInstead, if you saved using the `save_pretrained` method, then the directory already should have a `config.json` specifying the shape of the model, so you can simply load it using:\r\n\r\n```py\r\nfrom transformers import BertForSequenceClassification\r\n\r\nmodel = BertForSequenceClassification.from_pretrained(\"SAVED_SST_MODEL_DIR\")\r\n```\r\n\r\nIf you didn't save it using `save_pretrained`, but using `torch.save` or another, resulting in a `pytorch_model.bin` file containing your model state dict, you can initialize a configuration from your initial configuration (in this case I guess it's `bert-base-cased`) and assign three classes to it. You can then load your model by specifying which configuration to use:\r\n\r\n```py\r\nfrom transformers import BertForSequenceClassification, BertConfig\r\n\r\nconfig = BertConfig.from_pretrained(\"bert-base-cased\", num_labels=3)\r\nmodel = BertForSequenceClassification.from_pretrained(\"bert-base-cased\", config=config)\r\nmodel.load_state_dict(torch.load(\"SAVED_SST_MODEL_DIR/pytorch_model.bin\"))\r\n```\r\n\r\nLet me know how it works out for you.", "Yes!!!\r\nSetting the num_labels is useful!\r\nAnd I found that if i delete the classifier.weights and classifier.bias before i use torch.save(model_to_save.state_dict(), output_model_file), the pytorch_model.bin will be loaded well when further fine-tuning. And this model can be also used for QA or MultipleChoice.\r\n\r\n\r\n> Hello! If you try to load your `pytorch_model.bin` directly in `BertForSequenceClassification`, you'll indeed get an error as the model won't know that it is supposed to have three classes. That's what the configuration is for!\r\n> \r\n> I guess you're doing something similar to this:\r\n> \r\n> ```python\r\n> from transformers import BertForSequenceClassification\r\n> \r\n> model = BertForSequenceClassification.from_pretrained(\"bert-base-cased\")\r\n> model.load_state_dict(torch.load(\"SAVED_SST_MODEL_DIR/pytorch_model.bin\"))\r\n> # Crashes here\r\n> ```\r\n> \r\n> Instead, if you saved using the `save_pretrained` method, then the directory already should have a `config.json` specifying the shape of the model, so you can simply load it using:\r\n> \r\n> ```python\r\n> from transformers import BertForSequenceClassification\r\n> \r\n> model = BertForSequenceClassification.from_pretrained(\"SAVED_SST_MODEL_DIR\")\r\n> ```\r\n> \r\n> If you didn't save it using `save_pretrained`, but using `torch.save` or another, resulting in a `pytorch_model.bin` file containing your model state dict, you can initialize a configuration from your initial configuration (in this case I guess it's `bert-base-cased`) and assign three classes to it. You can then load your model by specifying which configuration to use:\r\n> \r\n> ```python\r\n> from transformers import BertForSequenceClassification, BertConfig\r\n> \r\n> config = BertConfig.from_pretrained(\"bert-base-cased\", num_labels=3)\r\n> model = BertForSequenceClassification.from_pretrained(\"bert-base-cased\", config=config)\r\n> model.load_state_dict(torch.load(\"SAVED_SST_MODEL_DIR/pytorch_model.bin\"))\r\n> ```\r\n> \r\n> Let me know how it works out for you.\r\n\r\nYes!!!\r\nSetting the num_labels is useful!\r\nAnd I found that if i delete the classifier.weights and classifier.bias before i use torch.save(model_to_save.state_dict(), output_model_file), the pytorch_model.bin will be loaded well when further fine-tuning. And this model can be also used for QA or MultipleChoice.\r\n", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2093?src=pr&el=h1) Report\n> Merging [#2093](https://codecov.io/gh/huggingface/transformers/pull/2093?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/2670b0d682746e1fe94ab9c7b4d2fd7f4af03193?src=pr&el=desc) will **decrease** coverage by `0.01%`.\n> The diff coverage is `100%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/2093?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #2093 +/- ##\n==========================================\n- Coverage 82.67% 82.65% -0.02% \n==========================================\n Files 111 111 \n Lines 16162 16162 \n==========================================\n- Hits 13362 13359 -3 \n- Misses 2800 2803 +3\n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/2093?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [transformers/tests/optimization\\_tf\\_test.py](https://codecov.io/gh/huggingface/transformers/pull/2093/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL29wdGltaXphdGlvbl90Zl90ZXN0LnB5) | `86.76% <100%> (ø)` | :arrow_up: |\n| [transformers/tests/modeling\\_tf\\_common\\_test.py](https://codecov.io/gh/huggingface/transformers/pull/2093/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL21vZGVsaW5nX3RmX2NvbW1vbl90ZXN0LnB5) | `95.38% <0%> (-1.54%)` | :arrow_down: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/2093?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2093?src=pr&el=footer). Last update [2670b0d...010489c](https://codecov.io/gh/huggingface/transformers/pull/2093?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "@julien-c Could you merge this please? :-)" ]

[]

[]

[ "I encountered this issue when using apex mixed precision, and I put `amp.initialize` after wrapping the model in `DistributedDataParallel`, and I believe reversing the order to first call `amp.initialize` fixed it", "I did not use mixed precision.", "+1, also having this issue for *-v1 and *-v2 models. I'm not using mixed precision either.", "I encountered the same problem. This problem can be fixed by removing line 289 and line 331 in modeling_albert.py (Those lines are not existed in this repo):\r\n```\r\nclass AlbertLayerGroup(nn.Module):\r\n def __init__(self, config):\r\n super(AlbertLayerGroup, self).__init__()\r\n\r\n self.output_attentions = config.output_attentions\r\n self.output_hidden_states = config.output_hidden_states\r\n self.albert_layers = nn.ModuleList([AlbertLayer(config) for _ in range(config.inner_group_num)])\r\n\r\n def forward(self, hidden_states, attention_mask=None, head_mask=None):\r\n layer_hidden_states = ()\r\n layer_attentions = ()\r\n\r\n for layer_index, albert_layer in enumerate(self.albert_layers):\r\n layer_output = albert_layer(hidden_states, attention_mask, head_mask[layer_index])\r\n hidden_states = layer_output[0]\r\n\r\n print(\"Layer index\", layer_index)\r\n\r\n if self.output_attentions:\r\n ...\r\n```\r\n```\r\nclass AlbertTransformer(nn.Module):\r\n def __init__(self, config):\r\n super(AlbertTransformer, self).__init__()\r\n\r\n self.config = config\r\n self.output_attentions = config.output_attentions\r\n self.output_hidden_states = config.output_hidden_states\r\n self.embedding_hidden_mapping_in = nn.Linear(config.embedding_size, config.hidden_size)\r\n self.albert_layer_groups = nn.ModuleList([AlbertLayerGroup(config) for _ in range(config.num_hidden_groups)])\r\n\r\n def forward(self, hidden_states, attention_mask=None, head_mask=None):\r\n hidden_states = self.embedding_hidden_mapping_in(hidden_states)\r\n\r\n all_attentions = ()\r\n\r\n if self.output_hidden_states:\r\n all_hidden_states = (hidden_states,)\r\n\r\n for i in range(self.config.num_hidden_layers):\r\n # Number of layers in a hidden group\r\n layers_per_group = int(self.config.num_hidden_layers / self.config.num_hidden_groups)\r\n\r\n # Index of the hidden group\r\n group_idx = int(i / (self.config.num_hidden_layers / self.config.num_hidden_groups))\r\n\r\n # Index of the layer inside the group\r\n layer_idx = int(i - group_idx * layers_per_group)\r\n\r\n print(group_idx, layer_idx)\r\n\r\n ...\r\n\r\n```", "> print(group_idx, layer_idx)\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 } ]

[]

[ "Hi! How did you obtain the train-v2.0 and dev-v2.0 files? Did you put the `--version_2_with_negative` flag to specify you're using SQuAD V2?", "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" ]

[]

[]

[ "Dpulicated to #2052 and closed it .\r\n\r\n" ]

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

[]

[ "Hello, the script would need to be adapted to run on TPU to take full advantage of the chips. We're actively working with the Cloud TPU team on scripts for fine-tuning on TPUs, which should be available in the coming weeks.", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[ "Hi @thomwolf - any suggestion would be greatly appreciated. \r\n\r\nI am looking forward to hosting one of the fine-tuned model (pytorch) using bert-as-a-service library. However, TF conversion seems to be the way to go, and I'm stuck as the script throws above errors that I am unable to understand. \r\n\r\n", "Hello! Indeed there seems to be a bug with the conversion script. In the meantime, here's how you can load your PyTorch checkpoint in a TF model:\r\n\r\n```py\r\nfrom transformers import BertForMaskedLM, TFBertForMaskedLM\r\n\r\n# The script should have already done that\r\nmodel = BertForMaskedLM.from_pretrained(\"bert-base-cased\")\r\nmodel.save_pretrained(\"here\")\r\n\r\n# Load the PyTorch model in TensorFlow\r\ntf_model = TFBertForMaskedLM.from_pretrained(\"here\", from_pt=True)\r\n\r\n# Save the TensorFlow model\r\ntf_model.save_pretrained(\"tf_test\")\r\n```\r\n\r\nYou can then convert the generated `.h5` model in a ckpt, like is described in [this issue](https://github.com/keras-team/keras/issues/9040) or this [stackoverflow issue](https://stackoverflow.com/questions/52650842/how-to-convert-hdf5-to-tensorflow-checkpoint)", "Thanks for the suggestion @LysandreJik \r\nI just tried this approach.\r\n\r\nIn my case, I fine tuned a model on MLM using run_lm_finetuning.py \r\n\r\n```\r\nfrom transformers import BertConfig, BertTokenizer, BertModel, BertForMaskedLM\r\nimport os\r\ntokenizer = BertTokenizer.from_pretrained(ft_cbert)\r\nmodel = BertModel.from_pretrained(ft_cbert)\r\nmodel.save_pretrained(str(os.path.join(ft_cbert, \"pt_bertmodel\")))\r\nmodel = BertForMaskedLM.from_pretrained(str(os.path.join(ft_cbert, \"pt_bertmodel\")))\r\nmodel.save_pretrained(str(os.path.join(ft_cbert, \"pt_maskedlm_bertmodel\")))\r\nmodel = TFBertModel.from_pretrained(os.path.join(ft_cbert, \"pt_maskedlm_bertmodel\"), from_pt=True)\r\nmodel.save_pretrained(os.path.join(ft_cbert, \"tf_maskedlm_bertmodel\"))\r\n```\r\nNow, when loading the pytorch model, TF doesn't seem to find weights and initializing all of the layers to 0 (correct me if I am interpreting incorrectly); I see a list of weights not loaded from pytorch model at the end of the log. \r\n\r\n> I1212 16:31:52.322784 139685136627520 modeling_utils.py:334] loading weights file /home/imagen/skc/bert/data/gold-regions/gold-finetune/cb-finetune-with-eval/pt_bertmodel/pytorch_model.bin\r\n> I1212 16:31:55.378468 139685136627520 configuration_utils.py:71] Configuration saved in /home/imagen/skc/bert/data/gold-regions/gold-finetune/cb-finetune-with-eval/pt_maskedlm_bertmodel/config.json\r\n> I1212 16:31:57.219412 139685136627520 modeling_utils.py:205] Model weights saved in /home/imagen/skc/bert/data/gold-regions/gold-finetune/cb-finetune-with-eval/pt_maskedlm_bertmodel/pytorch_model.bin\r\n> I1212 16:31:57.220998 139685136627520 configuration_utils.py:148] loading configuration file /home/imagen/skc/bert/data/gold-regions/gold-finetune/cb-finetune-with-eval/pt_maskedlm_bertmodel/config.json\r\n> I1212 16:31:57.222085 139685136627520 configuration_utils.py:168] Model config {\r\n> \"attention_probs_dropout_prob\": 0.1,\r\n> \"finetuning_task\": null,\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> \"is_decoder\": false,\r\n> \"layer_norm_eps\": 1e-12,\r\n> \"max_position_embeddings\": 512,\r\n> \"num_attention_heads\": 12,\r\n> \"num_hidden_layers\": 12,\r\n> \"num_labels\": 2,\r\n> \"output_attentions\": false,\r\n> \"output_hidden_states\": false,\r\n> \"output_past\": true,\r\n> \"pruned_heads\": {},\r\n> \"torchscript\": false,\r\n> \"type_vocab_size\": 2,\r\n> \"use_bfloat16\": false,\r\n> \"vocab_size\": 28996\r\n> }\r\n> \r\n> I1212 16:31:57.222966 139685136627520 modeling_tf_utils.py:255] loading weights file /home/imagen/skc/bert/data/gold-regions/gold-finetune/cb-finetune-with-eval/pt_maskedlm_bertmodel/pytorch_model.bin\r\n> I1212 16:31:57.293533 139685136627520 modeling_tf_pytorch_utils.py:78] Loading PyTorch weights from /home/imagen/skc/bert/data/gold-regions/gold-finetune/cb-finetune-with-eval/pt_maskedlm_bertmodel/pytorch_model.bin\r\n> I1212 16:31:58.017100 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/embeddings/word_embeddings/weight:0\r\n> I1212 16:31:58.018263 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/embeddings/position_embeddings/embeddings:0\r\n> I1212 16:31:58.019075 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/embeddings/token_type_embeddings/embeddings:0\r\n> I1212 16:31:58.019884 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/embeddings/LayerNorm/gamma:0\r\n> I1212 16:31:58.020372 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/embeddings/LayerNorm/beta:0\r\n> I1212 16:31:58.020853 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/attention/self/query/kernel:0\r\n> I1212 16:31:58.021338 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/attention/self/query/bias:0\r\n> I1212 16:31:58.021814 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/attention/self/key/kernel:0\r\n> I1212 16:31:58.022383 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/attention/self/key/bias:0\r\n> I1212 16:31:58.022871 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/attention/self/value/kernel:0\r\n> I1212 16:31:58.023389 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/attention/self/value/bias:0\r\n> I1212 16:31:58.023855 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/attention/output/dense/kernel:0\r\n> I1212 16:31:58.024335 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/attention/output/dense/bias:0\r\n> I1212 16:31:58.024829 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/attention/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.025296 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/attention/output/LayerNorm/beta:0\r\n> I1212 16:31:58.025762 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/intermediate/dense/kernel:0\r\n> I1212 16:31:58.026222 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/intermediate/dense/bias:0\r\n> I1212 16:31:58.026710 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/output/dense/kernel:0\r\n> I1212 16:31:58.027182 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/output/dense/bias:0\r\n> I1212 16:31:58.027667 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.028124 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._0/output/LayerNorm/beta:0\r\n> I1212 16:31:58.028624 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/attention/self/query/kernel:0\r\n> I1212 16:31:58.029091 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/attention/self/query/bias:0\r\n> I1212 16:31:58.029582 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/attention/self/key/kernel:0\r\n> I1212 16:31:58.030059 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/attention/self/key/bias:0\r\n> I1212 16:31:58.030566 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/attention/self/value/kernel:0\r\n> I1212 16:31:58.031049 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/attention/self/value/bias:0\r\n> I1212 16:31:58.031528 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/attention/output/dense/kernel:0\r\n> I1212 16:31:58.032037 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/attention/output/dense/bias:0\r\n> I1212 16:31:58.032562 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/attention/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.033143 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/attention/output/LayerNorm/beta:0\r\n> I1212 16:31:58.033643 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/intermediate/dense/kernel:0\r\n> I1212 16:31:58.034140 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/intermediate/dense/bias:0\r\n> I1212 16:31:58.034643 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/output/dense/kernel:0\r\n> I1212 16:31:58.035099 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/output/dense/bias:0\r\n> I1212 16:31:58.035623 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.036166 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._1/output/LayerNorm/beta:0\r\n> I1212 16:31:58.036743 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/attention/self/query/kernel:0\r\n> I1212 16:31:58.037309 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/attention/self/query/bias:0\r\n> I1212 16:31:58.037782 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/attention/self/key/kernel:0\r\n> I1212 16:31:58.038266 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/attention/self/key/bias:0\r\n> I1212 16:31:58.038728 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/attention/self/value/kernel:0\r\n> I1212 16:31:58.039192 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/attention/self/value/bias:0\r\n> I1212 16:31:58.039664 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/attention/output/dense/kernel:0\r\n> I1212 16:31:58.040130 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/attention/output/dense/bias:0\r\n> I1212 16:31:58.040640 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/attention/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.041108 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/attention/output/LayerNorm/beta:0\r\n> I1212 16:31:58.041579 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/intermediate/dense/kernel:0\r\n> I1212 16:31:58.042079 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/intermediate/dense/bias:0\r\n> I1212 16:31:58.042617 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/output/dense/kernel:0\r\n> I1212 16:31:58.043088 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/output/dense/bias:0\r\n> I1212 16:31:58.043587 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.044040 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._2/output/LayerNorm/beta:0\r\n> I1212 16:31:58.044509 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/attention/self/query/kernel:0\r\n> I1212 16:31:58.045005 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/attention/self/query/bias:0\r\n> I1212 16:31:58.050858 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/attention/self/key/kernel:0\r\n> I1212 16:31:58.051367 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/attention/self/key/bias:0\r\n> I1212 16:31:58.051822 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/attention/self/value/kernel:0\r\n> I1212 16:31:58.052374 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/attention/self/value/bias:0\r\n> I1212 16:31:58.052869 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/attention/output/dense/kernel:0\r\n> I1212 16:31:58.053370 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/attention/output/dense/bias:0\r\n> I1212 16:31:58.053862 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/attention/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.054336 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/attention/output/LayerNorm/beta:0\r\n> I1212 16:31:58.054825 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/intermediate/dense/kernel:0\r\n> I1212 16:31:58.055315 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/intermediate/dense/bias:0\r\n> I1212 16:31:58.055775 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/output/dense/kernel:0\r\n> I1212 16:31:58.056253 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/output/dense/bias:0\r\n> I1212 16:31:58.056724 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.057177 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._3/output/LayerNorm/beta:0\r\n> I1212 16:31:58.057679 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/attention/self/query/kernel:0\r\n> I1212 16:31:58.058135 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/attention/self/query/bias:0\r\n> I1212 16:31:58.058606 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/attention/self/key/kernel:0\r\n> I1212 16:31:58.059053 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/attention/self/key/bias:0\r\n> I1212 16:31:58.059546 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/attention/self/value/kernel:0\r\n> I1212 16:31:58.060031 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/attention/self/value/bias:0\r\n> I1212 16:31:58.060508 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/attention/output/dense/kernel:0\r\n> I1212 16:31:58.060971 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/attention/output/dense/bias:0\r\n> I1212 16:31:58.061455 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/attention/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.061920 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/attention/output/LayerNorm/beta:0\r\n> I1212 16:31:58.062463 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/intermediate/dense/kernel:0\r\n> I1212 16:31:58.062933 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/intermediate/dense/bias:0\r\n> I1212 16:31:58.063439 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/output/dense/kernel:0\r\n> I1212 16:31:58.063920 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/output/dense/bias:0\r\n> I1212 16:31:58.064412 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.064872 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._4/output/LayerNorm/beta:0\r\n> I1212 16:31:58.066597 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/attention/self/query/kernel:0\r\n> I1212 16:31:58.068921 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/attention/self/query/bias:0\r\n> I1212 16:31:58.069412 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/attention/self/key/kernel:0\r\n> I1212 16:31:58.069909 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/attention/self/key/bias:0\r\n> I1212 16:31:58.070411 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/attention/self/value/kernel:0\r\n> I1212 16:31:58.070859 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/attention/self/value/bias:0\r\n> I1212 16:31:58.071335 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/attention/output/dense/kernel:0\r\n> I1212 16:31:58.071808 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/attention/output/dense/bias:0\r\n> I1212 16:31:58.072312 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/attention/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.072788 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/attention/output/LayerNorm/beta:0\r\n> I1212 16:31:58.073315 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/intermediate/dense/kernel:0\r\n> I1212 16:31:58.073767 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/intermediate/dense/bias:0\r\n> I1212 16:31:58.074249 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/output/dense/kernel:0\r\n> I1212 16:31:58.074745 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/output/dense/bias:0\r\n> I1212 16:31:58.075211 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.075714 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._5/output/LayerNorm/beta:0\r\n> I1212 16:31:58.076181 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/attention/self/query/kernel:0\r\n> I1212 16:31:58.076673 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/attention/self/query/bias:0\r\n> I1212 16:31:58.077143 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/attention/self/key/kernel:0\r\n> I1212 16:31:58.077627 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/attention/self/key/bias:0\r\n> I1212 16:31:58.078094 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/attention/self/value/kernel:0\r\n> I1212 16:31:58.078586 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/attention/self/value/bias:0\r\n> I1212 16:31:58.079055 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/attention/output/dense/kernel:0\r\n> I1212 16:31:58.079540 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/attention/output/dense/bias:0\r\n> I1212 16:31:58.080033 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/attention/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.080506 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/attention/output/LayerNorm/beta:0\r\n> I1212 16:31:58.080977 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/intermediate/dense/kernel:0\r\n> I1212 16:31:58.081467 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/intermediate/dense/bias:0\r\n> I1212 16:31:58.081947 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/output/dense/kernel:0\r\n> I1212 16:31:58.082474 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/output/dense/bias:0\r\n> I1212 16:31:58.082974 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.083476 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._6/output/LayerNorm/beta:0\r\n> I1212 16:31:58.083951 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/attention/self/query/kernel:0\r\n> I1212 16:31:58.084461 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/attention/self/query/bias:0\r\n> I1212 16:31:58.084934 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/attention/self/key/kernel:0\r\n> I1212 16:31:58.085417 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/attention/self/key/bias:0\r\n> I1212 16:31:58.085875 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/attention/self/value/kernel:0\r\n> I1212 16:31:58.086349 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/attention/self/value/bias:0\r\n> I1212 16:31:58.086802 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/attention/output/dense/kernel:0\r\n> I1212 16:31:58.087476 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/attention/output/dense/bias:0\r\n> I1212 16:31:58.087949 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/attention/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.088423 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/attention/output/LayerNorm/beta:0\r\n> I1212 16:31:58.089007 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/intermediate/dense/kernel:0\r\n> I1212 16:31:58.089831 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/intermediate/dense/bias:0\r\n> I1212 16:31:58.090376 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/output/dense/kernel:0\r\n> I1212 16:31:58.090837 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/output/dense/bias:0\r\n> I1212 16:31:58.091311 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.091777 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._7/output/LayerNorm/beta:0\r\n> I1212 16:31:58.092295 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/attention/self/query/kernel:0\r\n> I1212 16:31:58.092808 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/attention/self/query/bias:0\r\n> I1212 16:31:58.093313 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/attention/self/key/kernel:0\r\n> I1212 16:31:58.093771 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/attention/self/key/bias:0\r\n> I1212 16:31:58.094259 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/attention/self/value/kernel:0\r\n> I1212 16:31:58.099888 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/attention/self/value/bias:0\r\n> I1212 16:31:58.100401 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/attention/output/dense/kernel:0\r\n> I1212 16:31:58.100865 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/attention/output/dense/bias:0\r\n> I1212 16:31:58.101369 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/attention/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.101860 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/attention/output/LayerNorm/beta:0\r\n> I1212 16:31:58.102412 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/intermediate/dense/kernel:0\r\n> I1212 16:31:58.103574 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/intermediate/dense/bias:0\r\n> I1212 16:31:58.104034 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/output/dense/kernel:0\r\n> I1212 16:31:58.104549 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/output/dense/bias:0\r\n> I1212 16:31:58.105008 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.105483 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._8/output/LayerNorm/beta:0\r\n> I1212 16:31:58.105949 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/attention/self/query/kernel:0\r\n> I1212 16:31:58.106442 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/attention/self/query/bias:0\r\n> I1212 16:31:58.106897 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/attention/self/key/kernel:0\r\n> I1212 16:31:58.107369 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/attention/self/key/bias:0\r\n> I1212 16:31:58.107837 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/attention/self/value/kernel:0\r\n> I1212 16:31:58.108303 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/attention/self/value/bias:0\r\n> I1212 16:31:58.108789 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/attention/output/dense/kernel:0\r\n> I1212 16:31:58.109263 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/attention/output/dense/bias:0\r\n> I1212 16:31:58.109742 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/attention/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.110190 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/attention/output/LayerNorm/beta:0\r\n> I1212 16:31:58.110669 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/intermediate/dense/kernel:0\r\n> I1212 16:31:58.111116 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/intermediate/dense/bias:0\r\n> I1212 16:31:58.111589 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/output/dense/kernel:0\r\n> I1212 16:31:58.112125 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/output/dense/bias:0\r\n> I1212 16:31:58.112630 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.113107 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._9/output/LayerNorm/beta:0\r\n> I1212 16:31:58.113591 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/attention/self/query/kernel:0\r\n> I1212 16:31:58.114055 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/attention/self/query/bias:0\r\n> I1212 16:31:58.114537 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/attention/self/key/kernel:0\r\n> I1212 16:31:58.115001 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/attention/self/key/bias:0\r\n> I1212 16:31:58.115493 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/attention/self/value/kernel:0\r\n> I1212 16:31:58.115964 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/attention/self/value/bias:0\r\n> I1212 16:31:58.116458 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/attention/output/dense/kernel:0\r\n> I1212 16:31:58.116904 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/attention/output/dense/bias:0\r\n> I1212 16:31:58.117376 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/attention/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.117864 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/attention/output/LayerNorm/beta:0\r\n> I1212 16:31:58.118321 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/intermediate/dense/kernel:0\r\n> I1212 16:31:58.118805 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/intermediate/dense/bias:0\r\n> I1212 16:31:58.119260 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/output/dense/kernel:0\r\n> I1212 16:31:58.119747 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/output/dense/bias:0\r\n> I1212 16:31:58.120195 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.120673 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._10/output/LayerNorm/beta:0\r\n> I1212 16:31:58.121122 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/attention/self/query/kernel:0\r\n> I1212 16:31:58.121608 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/attention/self/query/bias:0\r\n> I1212 16:31:58.122125 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/attention/self/key/kernel:0\r\n> I1212 16:31:58.122639 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/attention/self/key/bias:0\r\n> I1212 16:31:58.123139 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/attention/self/value/kernel:0\r\n> I1212 16:31:58.127967 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/attention/self/value/bias:0\r\n> I1212 16:31:58.128448 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/attention/output/dense/kernel:0\r\n> I1212 16:31:58.128974 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/attention/output/dense/bias:0\r\n> I1212 16:31:58.129623 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/attention/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.130099 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/attention/output/LayerNorm/beta:0\r\n> I1212 16:31:58.130589 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/intermediate/dense/kernel:0\r\n> I1212 16:31:58.131052 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/intermediate/dense/bias:0\r\n> I1212 16:31:58.131555 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/output/dense/kernel:0\r\n> I1212 16:31:58.132040 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/output/dense/bias:0\r\n> I1212 16:31:58.132566 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/output/LayerNorm/gamma:0\r\n> I1212 16:31:58.133050 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/encoder/layer_._11/output/LayerNorm/beta:0\r\n> I1212 16:31:58.133538 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/pooler/dense/kernel:0\r\n> I1212 16:31:58.133999 139685136627520 modeling_tf_pytorch_utils.py:159] Initialize TF weight tf_bert_model_3/bert/pooler/dense/bias:0\r\n> I1212 16:31:58.654147 139685136627520 modeling_tf_pytorch_utils.py:169] Weights or buffers not loaded from PyTorch model: {'cls.predictions.transform.dense.bias', 'cls.predictions.transform.LayerNorm.bias', 'cls.predictions.transform.dense.weight', 'cls.predictions.decoder.weight', 'cls.predictions.bias', 'cls.predictions.transform.LayerNorm.weight'}", "Hmm it says it's initializing all the weights from the PyTorch model, so they're not initialized to zero. \r\n\r\nIt's indeed not loading some weights from the PyTorch models, which are not needed for the TF model you're initializing (you're loading a BertForMaskedLM in a TFBertModel, so some weights are not used).", "Thanks for the clarification @LysandreJik \r\n\r\nThis way I am able to save to the model as .h5 version. However, since this step only saves model weights, converting .h5 to .ckpt is not straightforward as it requires the suitable architecture defined (when I am loading it in non-hface libs like tf.keras). It seems the model is not saved using model.save() instead with save_weights(). One needs to define the architecture to load weights and save as .ckpt. It would be great if there is an option to save the model including the necessary architecture to be loaded in TF. Let me know if I am missing something here. \r\n\r\nThank you.", "> Thanks for the clarification @LysandreJik\r\n> \r\n> This way I am able to save to the model as .h5 version. However, since this step only saves model weights, converting .h5 to .ckpt is not straightforward as it requires the suitable architecture defined (when I am loading it in non-hface libs like tf.keras). It seems the model is not saved using model.save() instead with save_weights(). One needs to define the architecture to load weights and save as .ckpt. It would be great if there is an option to save the model including the necessary architecture to be loaded in TF. Let me know if I am missing something here.\r\n> \r\n> Thank you.\r\n\r\nsame question", "\r\n\r\n> Thanks for the clarification @LysandreJik\r\n> \r\n> This way I am able to save to the model as .h5 version. However, since this step only saves model weights, converting .h5 to .ckpt is not straightforward as it requires the suitable architecture defined (when I am loading it in non-hface libs like tf.keras). It seems the model is not saved using model.save() instead with save_weights(). One needs to define the architecture to load weights and save as .ckpt. It would be great if there is an option to save the model including the necessary architecture to be loaded in TF. Let me know if I am missing something here.\r\n> \r\n> Thank you.\r\n\r\nHi, I have the same question, been stuck with this, have you solved the issue? \r\n\r\nThanks you." ]

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

[]

[ "Not sure to understand what you mean by `Using a simple similarity algorithm will always return the most similar even if it is not really correct`. What kind of simple similarity algo are you evoking here? What do you mean those simple algorithms aren't precise enough for your usecase?\r\n\r\nConsidering sentence similarity algorithms, I know:\r\n- Statistical approach using bag-of-words TF-IDF-based methods like BM25 (better on longer docs than sentences).\r\n- bag-of-word sentences and pooling (average-like) on word embeddings (word2vec-like) weighted by TF-IDF on a corpus for example.\r\n- Full Sentence Embedding learnt directly on a similarity training set (maybe finetuned on your domain). It builds sentence embedding in a vector-space in which you can compute distance between sentences. Those models often uses siamese approach based on pre-trained language models such as BERT.\r\n\r\nSentence embedding technique requires more work and domain knowledge but is the one reaching the highest metrics in SOTA. Other techniques can be enough depending on your needs and domain.\r\n\r\n", "Excuse me for not being clear enough, \r\nI wanted to say that if I have an FAQ data set and I want to get the most similar question for the user's question then, cosine similarity or TF-IDF will always give back a question even if it is not related. \r\nThis is about classifying either that question does have a similar one or not like does the snippet of code I posted earlier from deeppavlov. ", "No need to excuse ;)\r\nExcept an approach based on a dataset classifying \"similar and non-similar\" on your domain, any approach based on a score will require that you set a threshold to discriminate similar and non-similar. It can be fuzzy but sentence similarity is a very relative concept in any case.\r\n\r\n\r\n", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

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

[]

[ "I have wondered about this comment as well. I have implemented multi-GPU evaluation and it works perfectly fine. By evaluation I mean that the the work of evaluating is distributed and all results are then gathered to the main GPU (e.g. 0) or CPU which then calculates loss and secondary metrics (f1/pearson). I haven't experienced any issues with it but perhaps there is a reason that I don't know about. ", "Thanks, @BramVanroy. BTW, since you mentioned CPU, did you succeed in distributing LM fine-tuning on multiple CPUs? I tried that using torch.distributed and the 'gloo' backend and it seemed to be working fine, except that the total speed hardly improved.", "Oh no, what I meant was doing some calculations such as correlations in the CPU. I've never done fine-tuning on CPU. I can imagine that it takes a long time. ", "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", "@BramVanroy I was thinking of doing the same. Can you point me to your training script which uses multi-gpu eval the way you have described? I am hoping that if I see that first, I can avoid common mistakes. \r\nThanks!", "@dhruvdcoder Unfortunately, that code is in no state to be made public, in part because it is too complex and not incredibly well written. If I find the time, I plan to improve it and to add a PR here to update the example scripts to make use of multi-GPU evaluation." ]

[]

[]

[ "nevermind, it suddenly started working." ]

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

[]

[ "bert large is bigger than bert base. You're using a batch size of 24 (which is big, especially with 12 gradient accumulation steps). \r\n\r\nReduce your batch size in order for your model + your tensors to fit on the GPU and you won't experience the same error!", "Right @LysandreJik , reducing the batch size did fix the error but it looks like the generated model we receive is not same as provided by huggingface. \r\n\r\nIn our demo of closed domain QnA, https://demos.pragnakalp.com/bert-chatbot-demo, the answers are pretty good where we are using the model provided by huggingface (bert-large-uncased-whole-word-masking-finetuned-squad). But when we finetune on our own and even though we get 93.XX f1 score the accuracy of the model is not same as demo.\r\n\r\nWhat other parameters were set by huggingface to generate \"bert-large-uncased-whole-word-masking-finetuned-squad\" model? ", "If the only difference between the command you used and the command available [here](https://huggingface.co/transformers/examples.html#id1) is the batch size, you could try and adjust the gradient accumulation so that the resulting batch size is unchanged. For example if you put batch size equal to 6 (1/4 of the specified batch size, 24), you can multiply by 4 the gradient accumulation steps (-> 48) so that you keep the same batch size.\r\n\r\nWhat `exact_match` result did you obtain alongside the 93.xx F1 score?", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

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

[]

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

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

[]

[ "It is in the [optimization.py](https://github.com/huggingface/transformers/blob/df99f8c5a1c54d64fb013b43107011390c3be0d5/transformers/optimization.py), at line 45. It creates a schedule with a learning rate that decreases linearly after linearly increasing during a warmup period. In order to import it, you have to do the following:\r\n```\r\n> from transformers import get_linear_schedule_with_warmup\r\n> ...\r\n```\r\n\r\nI've tested this statement with **Python 3.6.9**, **Transformers 2.2.1** (installed with `pip install transformers`), **PyTorch 1.3.1** and **TensorFlow 2.0**.\r\n\r\n> $ pip show transformers\r\n> Name: transformers\r\n> Version: 2.2.1\r\n> Summary: State-of-the-art Natural Language Processing for TensorFlow 2.0 and PyTorch\r\n> Home-page: https://github.com/huggingface/transformers\r\n> Author: Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Google AI Language Team Authors, Open AI team Authors, Facebook AI Authors, Carnegie Mellon University Authors\r\n> Author-email: [[email protected]](mailto:[email protected])\r\n> License: Apache\r\n> Location: /home/ubuntu/anaconda3/lib/python3.6/site-packages\r\n> Requires: numpy, requests, regex, sacremoses, tqdm, sentencepiece, boto3\r\n> Required-by:", "I have the same error", "Do you see my comment above? Did you try out?\r\n\r\n> I have the same error", "> Do you see my comment above? Did you try out?\r\n> \r\n> > I have the same error\r\n\r\nI try to install from the git, fix the problem", "I tried both pip and git, still having the issue", "Still having this issue on 2.3.0 too", "Use get_linear_schedule_with_warmup() instead of WarmupLinearSchedule. I think they have the same function.", "> Use get_linear_schedule_with_warmup() instead of WarmupLinearSchedule. I think they have the same function.\r\n\r\nThe API is not quite the same, but it's similar enough that it should be easy enough to convert.\r\n\r\nFor example:\r\n\r\n```python\r\nscheduler = WarmupLinearSchedule(optimizer, warmup_steps=WARMUP_STEPS, t_total = -1)\r\n```\r\n\r\nbecomes...\r\n\r\n```python\r\nscheduler = get_linear_schedule_with_warmup(optimizer, num_warmup_steps=WARMUP_STEPS, num_training_steps = -1)\r\n```\r\n\r\nFor future visitors, see [docs](https://huggingface.co/transformers/main_classes/optimizer_schedules.html?highlight=get_linear_schedule_with_warmup#transformers.get_linear_schedule_with_warmup)", "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" ]

[]

[]

[ "Does this fix the non-deterministic behavior mentioned in #2027 ?", "Yes, this should return `[]` for every string that only contains whitespace characters. ", "Ok, great, merging then, thanks!" ]

[]

[]

[ "I got the same issue for version 2.2.1. ", "I also meet this issue and you may check out the possible root cause from #2052. \r\n\r\nMy workaround is backoff to 2.1.1 version.\r\n", "Should have been fixed with https://github.com/huggingface/transformers/pull/2051" ]

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

[]

[ "You may use the word as the input and make the sentence embedding as the word embedding.\r\nfor example, input is \r\n\"puppeteer\"\r\ntokens as\r\n'[CLS]', 'puppet', '##eer', '[SEP]'\r\nand then get embedding of this tokens list output.", "I have similar usage as well, I did a simple experiment, and observe that the subword embedding [subword1, subword2, subword3...] when input a whole sentence,\r\n the cosine similarity of [subword1,subword2],[subword1,subword3]... tends to above 90%.\r\n So that sum and average subwords' embedding doesn't change much.\r\nBtw, I tested this with Roberta models, and I observe quite different result for Bert models.", "Take a look at how bert-sense does it :)\r\nhttps://github.com/uhh-lt/bert-sense/blob/bfecb3c0e677d36ccfab4e2131ef9183995efaef/BERT_Model.py#L342", "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", "Excuse me did someone solve it ?", "@mathshangw I turn to use the `TransformerWordEmbeddings` of `flair` library to handle this. Here is an [example](https://github.com/flairNLP/flair/blob/master/examples/ner/run_ner.py#L119).", "I found a way to obtain the subtoken mask.\r\n\r\nThere is an argument called `return_offsets_mapping`. When you pass the tokenized sequence to the tokenizer, the returned offsets mapping records the start position of each token instead of the entire sentence, for example.\r\n\r\n```python\r\ntokens: list[str] = 'this is a niceing work'.split()\r\n# NOT THIS => tokens: str = 'this is a niceing work'\r\n\r\ntokenizer.tokenize(tokens, add_special_tokens=True)\r\n# ['▁this', '▁is', '▁a', '▁nice', 'ing', '▁work', '</s>', 'en_XX']\r\n\r\ntokens = tokenizer(tokens, add_special_tokens=True, is_split_into_words=True,\r\n return_offsets_mapping=True, return_tensors='pt')\r\n# {\r\n# 'input_ids': tensor([[903, 83, 10, 26267, 214, 4488, 2, 250004]]),\r\n# 'attention_mask': tensor([[1, 1, 1, 1, 1, 1, 1, 1]]),\r\n# 'offset_mapping': tensor([[[0, 4],\r\n# [0, 2],\r\n# [0, 1],\r\n# [0, 4],\r\n# [4, 7],\r\n# [0, 4],\r\n# [0, 0],\r\n# [0, 0]]])}\r\n\r\nsubtoken_mask = tokens['offset_mapping'][..., 0] != 0\r\n# tensor([[False, False, False, False, True, False, False, False]])\r\n```\r\n\r\nForget about the weird word *niceing*, I just want to get some sub-tokens. Now, by simply checking if this token starts from the beginning of the given word, we can tell if it is a sub-token.\r\n\r\nHope this is helpful to you guys." ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2078?src=pr&el=h1) Report\n> Merging [#2078](https://codecov.io/gh/huggingface/transformers/pull/2078?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/35ff345fc9df9e777b27903f11fa213e4052595b?src=pr&el=desc) will **increase** coverage by `0.01%`.\n> The diff coverage is `100%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/2078?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #2078 +/- ##\n==========================================\n+ Coverage 83.16% 83.18% +0.01% \n==========================================\n Files 109 109 \n Lines 15858 15874 +16 \n==========================================\n+ Hits 13188 13204 +16 \n Misses 2670 2670\n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/2078?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [transformers/hf\\_api.py](https://codecov.io/gh/huggingface/transformers/pull/2078/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL2hmX2FwaS5weQ==) | `97.5% <100%> (+0.62%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/2078?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2078?src=pr&el=footer). Last update [35ff345...5543617](https://codecov.io/gh/huggingface/transformers/pull/2078?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2077?src=pr&el=h1) Report\n> Merging [#2077](https://codecov.io/gh/huggingface/transformers/pull/2077?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/35ff345fc9df9e777b27903f11fa213e4052595b?src=pr&el=desc) will **not change** coverage.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/2077?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #2077 +/- ##\n=======================================\n Coverage 83.16% 83.16% \n=======================================\n Files 109 109 \n Lines 15858 15858 \n=======================================\n Hits 13188 13188 \n Misses 2670 2670\n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/2077?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [transformers/modeling\\_ctrl.py](https://codecov.io/gh/huggingface/transformers/pull/2077/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL21vZGVsaW5nX2N0cmwucHk=) | `96.47% <ø> (ø)` | :arrow_up: |\n| [transformers/modeling\\_tf\\_ctrl.py](https://codecov.io/gh/huggingface/transformers/pull/2077/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL21vZGVsaW5nX3RmX2N0cmwucHk=) | `97.86% <ø> (ø)` | :arrow_up: |\n| [transformers/modeling\\_gpt2.py](https://codecov.io/gh/huggingface/transformers/pull/2077/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL21vZGVsaW5nX2dwdDIucHk=) | `84.44% <ø> (ø)` | :arrow_up: |\n| [transformers/modeling\\_tf\\_gpt2.py](https://codecov.io/gh/huggingface/transformers/pull/2077/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL21vZGVsaW5nX3RmX2dwdDIucHk=) | `94.75% <ø> (ø)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/2077?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2077?src=pr&el=footer). Last update [35ff345...d0383e4](https://codecov.io/gh/huggingface/transformers/pull/2077?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "LGTM, merging!" ]

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

[]

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

[]

[]

[ "It works so well the CI failed because of a broken link :)", "Ok great!\r\n\r\nMaybe in the future, we would like to ensure model files can also be loaded without problems but this will suffice for now (and be fast)!\r\n\r\nmerging (when I've converted and added the missing model)", "Yes it would be great too! The only limit is the RAM available and the bandwidth on Circle CI's side. Assuming they're big enough we can download and load all files at the same time, it is easy to do. Maybe next time a related issue pops up?", "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2075?src=pr&el=h1) Report\n> Merging [#2075](https://codecov.io/gh/huggingface/transformers/pull/2075?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/9c58b236ef5fbbe5d0cbde4932eb342a73eaa0dc?src=pr&el=desc) will **increase** coverage by `0.01%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/2075?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #2075 +/- ##\n==========================================\n+ Coverage 80.35% 80.36% +0.01% \n==========================================\n Files 114 114 \n Lines 17091 17091 \n==========================================\n+ Hits 13733 13736 +3 \n+ Misses 3358 3355 -3\n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/2075?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [transformers/tests/modeling\\_common\\_test.py](https://codecov.io/gh/huggingface/transformers/pull/2075/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL21vZGVsaW5nX2NvbW1vbl90ZXN0LnB5) | `74.53% <0%> (+0.55%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/2075?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2075?src=pr&el=footer). Last update [9c58b23...d5712f7](https://codecov.io/gh/huggingface/transformers/pull/2075?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Ok great, merging, thanks @rlouf" ]

[]

[]

[ "It works so well that the CI failed because of a broken link :)" ]

[]

[]

[]

[]

[]

[ "Hi, there are several ways to check out the embeddings.\r\n\r\n1 - The easy way is to get the `embeddings` and use it as a `torch.nn.Module` (which it inherits from):\r\n\r\nFor example, this is the output of the embedding layer of the sentence \"Alright, let's do this\", of dimension (batch_size, sequence_length, hidden_size):\r\n\r\n```py\r\nfrom transformers import RobertaTokenizer, RobertaModel\r\nimport torch\r\n\r\ntok = RobertaTokenizer.from_pretrained(\"roberta-base\")\r\nmodel = RobertaModel.from_pretrained(\"roberta-base\")\r\n\r\nsentence = torch.tensor([tok.encode(\"Alright, let's do this\")])\r\nembedding_output = model.embeddings(sentence)\r\n```\r\n\r\n2 - A different way you can access them is by accessing the hidden states. You have to create a configuration object in order to specify that you would like the model to output its hidden states. You can then initialize the model from that configuration.\r\n\r\nUsing the example described above:\r\n```py\r\nfrom transformers import RobertaTokenizer, RobertaModel, RobertaConfig\r\nimport torch\r\n\r\nconfig = RobertaConfig.from_pretrained(\"roberta-base\")\r\nconfig.output_hidden_states = True\r\n\r\ntok = RobertaTokenizer.from_pretrained(\"roberta-base\")\r\nmodel = RobertaModel.from_pretrained(\"roberta-base\", config=config)\r\n\r\nsentence = torch.tensor([tok.encode(\"Alright, let's do this\")])\r\n\r\noutput = model(sentence) # returns a tuple(sequence_output, pooled_output, hidden_states)\r\nhidden_states = output[-1]\r\n\r\nembedding_output = hidden_states[0]\r\n```\r\n\r\nThose are the embeddings using only the embeddings layer, which do not change much when fine-tuned. If you want to access the sentence representations of the two models, you can simply use the `sequence_outputs`:\r\n\r\n```py\r\noutput = model(input)\r\nsequence_output = output[0]\r\n\r\nfinetuned_output = finetuned_model(input)\r\nfinetuned_sequence_output = finetuned_output[0]\r\n```\r\n\r\nYou can then compare those however you see fit!", "which model from roberta i can use for RU lang?\r\nor better using `xlm-mlm-17-1280` or `bert-base-multilingual-cased`?", "@LysandreJik thank you very much for your response! \r\n\r\nTo check my understanding, I can access the output of the embedding layer of roberta using the procedures you described (1&2). I can also access the embeddings learned at the last layer of roberta (the final layer) doing the following:\r\n```python\r\nfrom transformers import RobertaTokenizer, RobertaModel, RobertaConfig\r\nimport torch\r\n\r\nconfig = RobertaConfig.from_pretrained(\"roberta-base\")\r\nconfig.output_hidden_states = True\r\n\r\ntok = RobertaTokenizer.from_pretrained(\"roberta-base\")\r\nmodel = RobertaModel.from_pretrained(\"roberta-base\", config=config)\r\n\r\nsentence = torch.tensor([tok.encode(\"Alright, let's do this\")])\r\n\r\noutput = model(sentence)\r\nfinal_embeddings = output[0]\r\n```\r\nIs my understanding correct or have I missed something?", "@aclifton314 i think u should take `output[-1]` not a `output[0]`\r\nbtw, the 1st & 2nd example return similar vectors\r\nalso, the 2nd example has been working longer than 1st", "@vtrokhymenko Do you know what the difference is between `output[-1]` and `output[0]`?", "@aclifton314 the answer u can find here:\r\n>You have to create a configuration object in order to specify that you would like the model to output its hidden states", "@aclifton314 Referring to the output of the last layer as embeddings may be a bit ambiguous here, but yes, your `final_embeddings` variable holds the representation of your sequence at the uppermost layer (having gone through every model layer).\r\n\r\n`output[-1]` returns the hidden states while `output[0]` returns the sequence output.", "@LysandreJik @vtrokhymenko , thank you both for your replies! Closing this issue." ]

[]

[]

[ "Hi! Could you specify the command you used to launch `run_generation` as well as the versions in your environment? Pyton, pytorch, transformers? Thanks.", "`python scripts_htx/run_generation.py --model_type ctrl --model_name ctrl --repetition 1.2`\r\npython=3.7.3\r\ntorch=1.3.0\r\ntransformers=2.2.1\r\n\r\nBut I guess this issue is not related to the versions....", "This is actually the same as #1920 \r\nNow fixed on master (will be in the next release)." ]

[ { "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" ]

[]

[]

[ "Cool!", "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2069?src=pr&el=h1) Report\n> Merging [#2069](https://codecov.io/gh/huggingface/transformers/pull/2069?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/ee53de7aac8312140e87d452718e15e3d42e27dd?src=pr&el=desc) will **not change** coverage.\n> The diff coverage is `100%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/2069?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #2069 +/- ##\n=======================================\n Coverage 83.16% 83.16% \n=======================================\n Files 109 109 \n Lines 15858 15858 \n=======================================\n Hits 13188 13188 \n Misses 2670 2670\n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/2069?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/2069/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL21vZGVsaW5nX3V0aWxzLnB5) | `90.86% <100%> (ø)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/2069?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2069?src=pr&el=footer). Last update [ee53de7...1d87b37](https://codecov.io/gh/huggingface/transformers/pull/2069?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "@thomwolf and @VictorSanh - could you please look into #2088 (PT to TF)? tagging you guys in this thread as it seems relevant, please let me know otherwise. \r\n\r\nThanks", "This thread is not relevant but I'll give a look at your issue soon.", "thanks!" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2068?src=pr&el=h1) Report\n> Merging [#2068](https://codecov.io/gh/huggingface/transformers/pull/2068?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/2d5d86e03779b4b316698438caff0f675ee54abd?src=pr&el=desc) will **increase** coverage by `0.01%`.\n> The diff coverage is `100%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/2068?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #2068 +/- ##\n==========================================\n+ Coverage 83.15% 83.17% +0.01% \n==========================================\n Files 109 109 \n Lines 15869 15869 \n==========================================\n+ Hits 13196 13199 +3 \n+ Misses 2673 2670 -3\n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/2068?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [transformers/modeling\\_bert.py](https://codecov.io/gh/huggingface/transformers/pull/2068/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL21vZGVsaW5nX2JlcnQucHk=) | `87.68% <100%> (ø)` | :arrow_up: |\n| [transformers/tests/modeling\\_common\\_test.py](https://codecov.io/gh/huggingface/transformers/pull/2068/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL21vZGVsaW5nX2NvbW1vbl90ZXN0LnB5) | `74.23% <0%> (+0.57%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/2068?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2068?src=pr&el=footer). Last update [2d5d86e...18fb935](https://codecov.io/gh/huggingface/transformers/pull/2068?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "lgtm", "should this fix be added here as well? @thomwolf https://github.com/huggingface/transformers/blob/6cc06d17394f5715cdf2d13a1ef7680bedaee9e2/src/transformers/modeling_utils.py#L700" ]

[]

[]

[ "Did you get the solution to this? @elixium ", "Hi any update on this? I would like to deploy huggingface Transformers model with Tensorflow Serving too" ]

[ { "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" ]

[]

[]

[ "Merging now to fix the xlnet test issue on master at the same time.", "Also cc'ing @louismartin on this.", "Thanks for fixing that.\r\nThis comes from a problem in fairseq where special tokens are added twice when using SentencePiece.\r\nCross-referencing the fairseq issue: [https://github.com/pytorch/fairseq/issues/1309](https://github.com/pytorch/fairseq/issues/1309)" ]

[]

[]

[ "I don't think it is a problem. \r\nYour model will learn the embeddings of the words in your own dictionary.\r\n\r\nActually Nothing will be unchanged if you changed dictionary position as well as you keeped the embedding weight just the same order with your dictionary. ", "Thanks @karajan1001. \r\nI am not sure if I understood well the last sentence. \r\nYou meant that I must not change the dictionary after I train the LM based on that? \r\nAfter that I train the SentencePiece tokenizer and the relative vocabulary is given, than the vocab is an input of Language Model training and it cannot be changed anymore.", "> After that I train the SentencePiece tokenizer and the relative vocabulary is given, than the vocab is an input of Language Model training and it cannot be changed anymore.\r\n\r\nI think so. The vocabulary tells the model which array to get with the input tokens.\r\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2063?src=pr&el=h1) Report\n> Merging [#2063](https://codecov.io/gh/huggingface/transformers/pull/2063?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/fb0d2f1da102d699c6457fd98be35f89852d08b9?src=pr&el=desc) will **not change** coverage.\n> The diff coverage is `100%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/2063?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #2063 +/- ##\n=======================================\n Coverage 83.58% 83.58% \n=======================================\n Files 105 105 \n Lines 15568 15568 \n=======================================\n Hits 13012 13012 \n Misses 2556 2556\n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/2063?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [transformers/tokenization\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/2063/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rva2VuaXphdGlvbl91dGlscy5weQ==) | `91.87% <100%> (ø)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/2063?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2063?src=pr&el=footer). Last update [fb0d2f1...7f998b1](https://codecov.io/gh/huggingface/transformers/pull/2063?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Thanks for catching it!\r\n\r\n@LysandreJik do we want to have `special_tokens_mask` returned as a tensor when `encode` is called with `return_tensors='pt' or 'tf'`. I would say no.", "Thanks for that @guillaume-be, we can merge.\r\n\r\n@thomwolf I don't really see a use-case where having it as a tensor would be useful. I believe its main use is be for pre-processing, maybe it would be useful to have it as a tensor then but I'm not convinced.", "Ok, great! merging" ]

[]

[]

[ "Solved it by typecasting posixpath to string" ]

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

[]

[ "Can you show us the full error message?", "Can it be related to #2109 in some way?\r\n\r\n> ## Bug\r\n> Model I am using (Bert, XLNet....):\r\n> \r\n> Language I am using the model on (English, Chinese....):\r\n> \r\n> The problem arise when using:\r\n> \r\n> * [x] the official example scripts: (give details)\r\n> * [ ] my own modified scripts: (give details)\r\n> \r\n> The tasks I am working on is:\r\n> \r\n> * [ ] an official GLUE/SQUaD task: (give the name)\r\n> * [x] my own task or dataset: (give details)\r\n> \r\n> I used this script to load the bert model i had finetuned for a classification task following google_research_bert.\r\n> I want to convert those TF checkpoints to pytorch.\r\n> \r\n> ```python\r\n> config = BertConfig.from_pretrained('bert-base-uncased')\r\n> config.num_labels=4\r\n> # You will need to load a BertForSequenceClassification model\r\n> model = BertForSequenceClassification(config)\r\n> \r\n> tf_checkpoint_path = init_checkpoint\r\n> # Load weights from tf checkpoint\r\n> load_tf_weights_in_bert(model, config, tf_checkpoint_path)\r\n> \r\n> pytorch_dump_path = \"./pytorch_bert_output\"\r\n> # Save pytorch-model\r\n> print(\"Save PyTorch model to {}\".format(pytorch_dump_path))\r\n> torch.save(model.state_dict(), pytorch_dump_path)\r\n> ```\r\n> \r\n> When i execute this, i get the following error, BertForSequenceClassification' object has no attribute 'bias. Any leads would be helpful.\r\n> 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" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2060?src=pr&el=h1) Report\n> Merging [#2060](https://codecov.io/gh/huggingface/transformers/pull/2060?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/5bfcd0485ece086ebcbed2d008813037968a9e58?src=pr&el=desc) will **not change** coverage.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/2060?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #2060 +/- ##\n=======================================\n Coverage 83.58% 83.58% \n=======================================\n Files 105 105 \n Lines 15568 15568 \n=======================================\n Hits 13012 13012 \n Misses 2556 2556\n```\n\n\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/2060?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2060?src=pr&el=footer). Last update [5bfcd04...12d18d4](https://codecov.io/gh/huggingface/transformers/pull/2060?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[]

[ "Did you solve it? I have the same problem as you." ]

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

[]

[ "What is the batch size you used?", "The same bug occurs with Python 3.6.9, Transformers 2.2.1 (installed with `pip install transformers`), PyTorch 1.3.1 and TensorFlow 2.0.\r\nStack trace:\r\n```\r\nPython 3.6.9 |Anaconda, Inc.| (default, Jul 30 2019, 19:07:31) \r\n[GCC 7.3.0] on linux\r\nType \"help\", \"copyright\", \"credits\" or \"license\" for more information.\r\n>>> from transformers import TFAlbertModel\r\n/home/<user>/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:541: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\r\n _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\r\n/home/<user>/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:542: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\r\n _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\r\n/home/<user>/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:543: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\r\n _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\r\n/home/<user>/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:544: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\r\n _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\r\n/home/<user>/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:545: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\r\n _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\r\n/home/<user>/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:550: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\r\n np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\r\n2019-12-05 09:55:19.006308: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA\r\n2019-12-05 09:55:19.027197: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 3408000000 Hz\r\n2019-12-05 09:55:19.027888: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x5582d4e6e4c0 executing computations on platform Host. Devices:\r\n2019-12-05 09:55:19.027909: I tensorflow/compiler/xla/service/service.cc:175] StreamExecutor device (0): Host, Default Version\r\n>>> model = TFAlbertModel.from_pretrained('albert-base-v1')\r\n100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 484/484 [00:00<00:00, 185172.23B/s]\r\n299B [00:00, 131456.70B/s]\r\n2019-12-05 09:55:28.628697: W tensorflow/python/util/util.cc:299] Sets are not currently considered sequences, but this may change in the future, so consider avoiding using them.\r\nTraceback (most recent call last):\r\n File \"<stdin>\", line 1, in <module>\r\n File \"/home/<user>/anaconda3/envs/huggingface/lib/python3.6/site-packages/transformers/modeling_tf_utils.py\", line 289, in from_pretrained\r\n model.load_weights(resolved_archive_file, by_name=True)\r\n File \"/home/<user>/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorflow_core/python/keras/engine/training.py\", line 181, in load_weights\r\n return super(Model, self).load_weights(filepath, by_name)\r\n File \"/home/<user>/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorflow_core/python/keras/engine/network.py\", line 1171, in load_weights\r\n with h5py.File(filepath, 'r') as f:\r\n File \"/home/<user>/anaconda3/envs/huggingface/lib/python3.6/site-packages/h5py/_hl/files.py\", line 408, in __init__\r\n swmr=swmr)\r\n File \"/home/<user>/anaconda3/envs/huggingface/lib/python3.6/site-packages/h5py/_hl/files.py\", line 173, in make_fid\r\n fid = h5f.open(name, flags, fapl=fapl)\r\n File \"h5py/_objects.pyx\", line 54, in h5py._objects.with_phil.wrapper\r\n File \"h5py/_objects.pyx\", line 55, in h5py._objects.with_phil.wrapper\r\n File \"h5py/h5f.pyx\", line 88, in h5py.h5f.open\r\nOSError: Unable to open file (file signature not found)\r\n```\r\nIf I try to use the PyTorch version of Albert with _albert-base-v1_, it works as expected!\r\nStack trace:\r\n```\r\n>>> model = AlbertModel.from_pretrained('albert-base-v1')\r\n100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 47376396/47376396 [00:04<00:00, 10033199.48B/s]\r\n```\r\n\r\nSaid this, I suspect that the TensorFlow version of Albert is not developed correctly (it misses the config). Is it possible? Now I'm investigating on..\r\n\r\nUPDATE 1: I've gone to the Transformers' source code in the [modeling_tf_albert.py](https://github.com/huggingface/transformers/blob/e85855f2c408f65a4aaf5d15baab6ca90fd26050/transformers/) and I've downloaded the .h5 model **correctly** (from [this link](https://s3.amazonaws.com/models.huggingface.co/bert/albert-base-v1-tf_model.h5)). So I suspect there is an internal error that is independent from the download of the .h5 file.\r\n\r\n> ## Bug\r\n> Model I am using :ALBERT\r\n> \r\n> Language I am using the model on (English, Chinese....):English\r\n> \r\n> > from transformers import TFAlbertModel\r\n> \r\n> > model2=TFAlbertModel.from_pretrained('albert-base-v1')\r\n> \r\n> Then:\r\n> \r\n> > \r\n> \r\n> OSError Traceback (most recent call last)\r\n> in \r\n> ----> 1 model2=TFAlbertModel.from_pretrained('albert-base-v1')\r\n> \r\n> ~/anaconda3/lib/python3.7/site-packages/transformers/modeling_tf_utils.py in from_pretrained(cls, pretrained_model_name_or_path, *model_args, **kwargs)\r\n> 287 # 'by_name' allow us to do transfer learning by skipping/adding layers\r\n> 288 # see https://github.com/tensorflow/tensorflow/blob/00fad90125b18b80fe054de1055770cfb8fe4ba3/tensorflow/python/keras/engine/network.py#L1339-L1357\r\n> --> 289 model.load_weights(resolved_archive_file, by_name=True)\r\n> 290\r\n> 291 ret = model(model.dummy_inputs, training=False) # Make sure restore ops are run\r\n> \r\n> ~/anaconda3/lib/python3.7/site-packages/tensorflow_core/python/keras/engine/training.py in load_weights(self, filepath, by_name)\r\n> 179 raise ValueError('Load weights is not yet supported with TPUStrategy '\r\n> 180 'with steps_per_run greater than 1.')\r\n> --> 181 return super(Model, self).load_weights(filepath, by_name)\r\n> 182\r\n> 183 @trackable.no_automatic_dependency_tracking\r\n> \r\n> ~/anaconda3/lib/python3.7/site-packages/tensorflow_core/python/keras/engine/network.py in load_weights(self, filepath, by_name)\r\n> 1169 'first, then load the weights.')\r\n> 1170 self._assert_weights_created()\r\n> -> 1171 with h5py.File(filepath, 'r') as f:\r\n> 1172 if 'layer_names' not in f.attrs and 'model_weights' in f:\r\n> 1173 f = f['model_weights']\r\n> \r\n> ~/anaconda3/lib/python3.7/site-packages/h5py/_hl/files.py in **init**(self, name, mode, driver, libver, userblock_size, swmr, rdcc_nslots, rdcc_nbytes, rdcc_w0, track_order, **kwds)\r\n> 392 fid = make_fid(name, mode, userblock_size,\r\n> 393 fapl, fcpl=make_fcpl(track_order=track_order),\r\n> --> 394 swmr=swmr)\r\n> 395\r\n> 396 if swmr_support:\r\n> \r\n> ~/anaconda3/lib/python3.7/site-packages/h5py/_hl/files.py in make_fid(name, mode, userblock_size, fapl, fcpl, swmr)\r\n> 168 if swmr and swmr_support:\r\n> 169 flags |= h5f.ACC_SWMR_READ\r\n> --> 170 fid = h5f.open(name, flags, fapl=fapl)\r\n> 171 elif mode == 'r+':\r\n> 172 fid = h5f.open(name, h5f.ACC_RDWR, fapl=fapl)\r\n> \r\n> h5py/_objects.pyx in h5py._objects.with_phil.wrapper()\r\n> \r\n> h5py/_objects.pyx in h5py._objects.with_phil.wrapper()\r\n> \r\n> h5py/h5f.pyx in h5py.h5f.open()\r\n> \r\n> OSError: Unable to open file (file signature not found)\r\n> \r\n> I tried to trace the GPU state, the memory usage is 15513MiB / 16130MiB, it is obvious that\r\n> when I create a model, it automatically allocates memory in GPU, but when I tried this in colab and use the same TF version, it works well, after creating model, there still are much free memory.\r\n> *OS: Linux version 4.9.0-11-amd64\r\n> \r\n> * Python version:3.7\r\n> * TF version:TF2.0\r\n> * Transformers version (or branch):2.2\r\n> * Using GPU ?GPU\r\n> \r\n> ## Additional context", "> What is the batch size you used?\r\n\r\nI haven't tried to train, I just run one line code to create the model, then problem happened. \r\n\r\n> model2=TFAlbertModel.from_pretrained('albert-base-v1')\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" ]

[]

[]

[ "It is located under `configuration_roberta.py`, see it [here](https://github.com/huggingface/transformers/blob/1c542df7e554a2014051dd09becf60f157fed524/transformers/configuration_roberta.py#L31) :)", "Thanks @stefan-it ! Missed the readme part of calling `distilroberta-base` with `RobertaModel` instead of `DistilBertModel`. Closing. " ]

[]

[]

[ "This could be related to this issue here: https://github.com/huggingface/transformers/issues/1837 :)", "I copied the&nbsp;get_linear_schedule_with_warmup function code add to my project in the&nbsp;transformers/optimization.py\r\nand then&nbsp;it worked&nbsp;\r\nThank you for developing such an brilliant library.\r\n\r\n\r\n\r\n\r\n------------------&nbsp;原始邮件&nbsp;------------------\r\n发件人:&nbsp;\"Stefan Schweter\"<[email protected]&gt;;\r\n发送时间:&nbsp;2019年12月5日(星期四) 上午7:19\r\n收件人:&nbsp;\"huggingface/transformers\"<[email protected]&gt;;\r\n抄送:&nbsp;\"FOXaaFOX\"<[email protected]&gt;;\"Author\"<[email protected]&gt;;\r\n主题:&nbsp;Re: [huggingface/transformers] cannot import name 'get_linear_schedule_with_warmup' from 'transformers.optimization' (#2056)\r\n\r\n\r\n\r\n\r\nThis could be related to this issue here: #1837 :)\r\n \r\n—\r\nYou are receiving this because you authored the thread.\r\nReply to this email directly, view it on GitHub, or unsubscribe." ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2055?src=pr&el=h1) Report\n> Merging [#2055](https://codecov.io/gh/huggingface/transformers/pull/2055?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/35ff345fc9df9e777b27903f11fa213e4052595b?src=pr&el=desc) will **decrease** coverage by `0.45%`.\n> The diff coverage is `95.45%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/2055?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #2055 +/- ##\n=========================================\n- Coverage 83.16% 82.7% -0.46% \n=========================================\n Files 109 109 \n Lines 15858 15943 +85 \n=========================================\n- Hits 13188 13186 -2 \n- Misses 2670 2757 +87\n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/2055?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [transformers/modeling\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/2055/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL21vZGVsaW5nX29wZW5haS5weQ==) | `81.45% <0%> (-0.55%)` | :arrow_down: |\n| [transformers/tokenization\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/2055/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rva2VuaXphdGlvbl94bG5ldC5weQ==) | `90.24% <100%> (ø)` | :arrow_up: |\n| [transformers/tests/modeling\\_tf\\_auto\\_test.py](https://codecov.io/gh/huggingface/transformers/pull/2055/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL21vZGVsaW5nX3RmX2F1dG9fdGVzdC5weQ==) | `36.36% <100%> (ø)` | :arrow_up: |\n| [transformers/tokenization\\_albert.py](https://codecov.io/gh/huggingface/transformers/pull/2055/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rva2VuaXphdGlvbl9hbGJlcnQucHk=) | `89.74% <100%> (ø)` | :arrow_up: |\n| [transformers/tests/modeling\\_tf\\_openai\\_gpt\\_test.py](https://codecov.io/gh/huggingface/transformers/pull/2055/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL21vZGVsaW5nX3RmX29wZW5haV9ncHRfdGVzdC5weQ==) | `94.73% <100%> (ø)` | :arrow_up: |\n| [transformers/tests/modeling\\_tf\\_bert\\_test.py](https://codecov.io/gh/huggingface/transformers/pull/2055/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL21vZGVsaW5nX3RmX2JlcnRfdGVzdC5weQ==) | `96.22% <100%> (ø)` | :arrow_up: |\n| [transformers/tests/tokenization\\_auto\\_test.py](https://codecov.io/gh/huggingface/transformers/pull/2055/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL3Rva2VuaXphdGlvbl9hdXRvX3Rlc3QucHk=) | `50% <100%> (ø)` | :arrow_up: |\n| [transformers/tests/tokenization\\_distilbert\\_test.py](https://codecov.io/gh/huggingface/transformers/pull/2055/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL3Rva2VuaXphdGlvbl9kaXN0aWxiZXJ0X3Rlc3QucHk=) | `63.63% <100%> (ø)` | :arrow_up: |\n| [transformers/tests/modeling\\_tf\\_roberta\\_test.py](https://codecov.io/gh/huggingface/transformers/pull/2055/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL21vZGVsaW5nX3RmX3JvYmVydGFfdGVzdC5weQ==) | `75.2% <100%> (ø)` | :arrow_up: |\n| [transformers/tests/tokenization\\_transfo\\_xl\\_test.py](https://codecov.io/gh/huggingface/transformers/pull/2055/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL3Rva2VuaXphdGlvbl90cmFuc2ZvX3hsX3Rlc3QucHk=) | `97.43% <100%> (ø)` | :arrow_up: |\n| ... and [39 more](https://codecov.io/gh/huggingface/transformers/pull/2055/diff?src=pr&el=tree-more) | |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/2055?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2055?src=pr&el=footer). Last update [35ff345...61978c1](https://codecov.io/gh/huggingface/transformers/pull/2055?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Given that the PR touches the whole test suite and that tests pass, if there's no opposition,I'd like to merge it before master diverges.\r\n\r\nWe can figure out running tests on the GPU on CircleCI separately.", "Agreed. Squashed and merged." ]

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

[]

[ "I found [this](https://huggingface.co/transformers/_modules/transformers/modeling_bert.html) code which has transpose_for_scores but I am not sure how this can be used with the above code.", "Yes, the `attentions` outputs of the model are the softmax values.", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

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

[]

[ "I had the same problem with different tasks . I've tried all the XLM pre-training models and got random results. Please let us know if you have solved this problem. I'm trying to figure it out. @DanKing1903 ", "I was able to reproduce the results of XLM on XNLI.\r\nIt was highly sensitive to hyper parameters.\r\nI would suggest that you tune your learning_rate ~ 1.5e-6 and batch_size ~ 4.", "> I was able to reproduce the results of XLM on XNLI.\r\n> It was highly sensitive to hyper parameters.\r\n> I would suggest that you tune your learning_rate ~ 1.5e-6 and batch_size ~ 4.\r\n\r\nThank you for your answer. I modified the parameters as you suggested. xlm-mlm-17-1280 and xlm-mlm-100-1280 batch_size only go up to 2, and others modified the parameters as before. I wonder if it has anything to do with the task. Looking forward to your reply", "I got the same problem.\r\nI suggest you to go with RMSprop (require less memory than Adam so you can have a bigger batch size) with learning rate 3e-5 (very important to use a small learning rate otherwise it diverge) and clipnorm of 1.0. Personally, I use a global batch size of 20 where each GPU has a batch size of 10. I haven't tested with accumulated gradient since tf2.0 does not have a wrapper for it at the moment, but I think it will help.\r\nIt might also help adding momentum to RMSprop or a scheduling learning rate, but haven't test it yet. If you have some hint or previous experience on it please let me know\r\n", "> > I was able to reproduce the results of XLM on XNLI.\r\n> > It was highly sensitive to hyper parameters.\r\n> > I would suggest that you tune your learning_rate ~ 1.5e-6 and batch_size ~ 4.\r\n> \r\n> Thank you for your answer. I modified the parameters as you suggested. xlm-mlm-17-1280 and xlm-mlm-100-1280 batch_size only go up to 2, and others modified the parameters as before. I wonder if it has anything to do with the task. Looking forward to your reply\r\n\r\nIn my limited experience, XLM is highly sensitive to HPs (it seems to be also the case with RoBERTa on GLUE, to a lesser extent). However, it is not something I observed with mBERT (and Distil-mBERT). So I don't think it has to do with XNLI since there is no consistent pattern across different models.", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "Has anyone come up with a good set of hyper-parameters to train XLM models very well? Thanks for sharing the experience!" ]

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

[]

[ "With Transformers **2.2.0**, it works as expected!\r\n```\r\nPython 3.6.9 |Anaconda, Inc.| (default, Jul 30 2019, 19:07:31) \r\n[GCC 7.3.0] on linux\r\nType \"help\", \"copyright\", \"credits\" or \"license\" for more information.\r\n>>> import torch\r\n>>> from transformers import BertTokenizer\r\n/home/vidiemme/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:541: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\r\n _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\r\n/home/vidiemme/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:542: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\r\n _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\r\n/home/vidiemme/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:543: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\r\n _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\r\n/home/vidiemme/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:544: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\r\n _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\r\n/home/vidiemme/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:545: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\r\n _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\r\n/home/vidiemme/anaconda3/envs/huggingface/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:550: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\r\n np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\r\n2019-12-05 10:19:00.776555: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA\r\n2019-12-05 10:19:00.799189: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 3408000000 Hz\r\n2019-12-05 10:19:00.799911: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x55971fe06640 executing computations on platform Host. Devices:\r\n2019-12-05 10:19:00.799929: I tensorflow/compiler/xla/service/service.cc:175] StreamExecutor device (0): Host, Default Version\r\n>>> tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')\r\n>>> tokenizer.tokenize('The quick brown [MASK] jumps over the lazy dog.')\r\n['the', 'quick', 'brown', '[MASK]', 'jumps', 'over', 'the', 'lazy', 'dog', '.']\r\n```\r\nWith Transformers **2.2.1**, the bug you've highlighted occurs to me too!\r\n\r\n> ## Bug\r\n> Model I am using (Bert, XLNet....): 'bert-base-uncased'\r\n> \r\n> Language I am using the model on (English, Chinese....): English\r\n> \r\n> After upgrading to 2.2.1 version, the BERT tokenizer cannot tokenize special word while it works in 2.1.1 version.\r\n> \r\n> According to [here](https://github.com/huggingface/transformers/blob/master/transformers/tokenization_bert.py#L71), 'bert-base-uncased' should perform lower case operation. Inputs follow this config to [perform lower case operation](https://github.com/huggingface/transformers/blob/master/transformers/tokenization_utils.py#L615), while no corresponding action for [special tokens](https://github.com/huggingface/transformers/blob/master/transformers/tokenization_utils.py#L1100). Eventually, it tokenizes '[MASK]' to 3 subwords (e.g. [, mask and ]) rather than skip the tokenization operation in [here](https://github.com/huggingface/transformers/blob/master/transformers/tokenization_utils.py#L658).\r\n> \r\n> Error occurs after this [commit](https://github.com/huggingface/transformers/commit/7246d3c2f93c4461f3ec8ada7a26a002d8f196ea).\r\n> \r\n> ## To Reproduce\r\n> Steps to reproduce the behavior:\r\n> \r\n> ```\r\n> import torch\r\n> from transformers import BertTokenizer\r\n> tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')\r\n> tokenizer.tokenize('The quick brown [MASK] jumps over the lazy dog.')\r\n> ```\r\n> \r\n> ## Expected behavior\r\n> Expected output is\r\n> ['the', 'quick', 'brown', **'[mask]'**, 'jumps', 'over', 'the', 'lazy', 'dog', '.']\r\n> while actual output is\r\n> ['the', 'quick', 'brown', **'[', 'mask', ']'**, 'jumps', 'over', 'the', 'lazy', 'dog', '.']", "I've tracked it yesterday evening and I confirm all of that too. `PreTrainedTokenizer.add_tokens` forces added tokens to lower-case but tokens coming from BertTokenizer constructor aren't lower-cased.\r\n\r\nYet considering a viable patch, I tend to think there is an issue linked to current design of Tokenizers with respect to flags such as `do_lower_case`.\r\n\r\nFor example, current BertTokenizer is:\r\n\r\n```python\r\nclass BertTokenizer(PreTrainedTokenizer):\r\n...\r\n\r\n def __init__(self, vocab_file, do_lower_case=True, do_basic_tokenize=True, never_split=None,\r\n unk_token=\"[UNK]\", sep_token=\"[SEP]\", pad_token=\"[PAD]\", cls_token=\"[CLS]\",\r\n mask_token=\"[MASK]\", tokenize_chinese_chars=True, **kwargs):\r\n\r\n super(BertTokenizer, self).__init__(unk_token=unk_token, sep_token=sep_token,\r\n pad_token=pad_token, cls_token=cls_token,\r\n mask_token=mask_token, **kwargs)\r\n```\r\n\r\nSo `BertTokenizer` knows about `do_lower_case` but not the super class `PreTrainedTokenizer`. Moreover, by default `do_lower_case` is True but all tokens are defined in upper_case.\r\n\r\nThen, in `PreTrainedTokenizer`, there are some `if self.init_kwargs.get('do_lower_case', False):` in different places of the code to force text or added_tokens to lower_case before tokenization. But this means you inject a knowledge of `lower_case` in a class that doesn't know it by construction. It works but as we see in the case of token case, it's error-prone and not so robust. Moreover, if there were several flags, it would become even harder to track.\r\n\r\nA solution could be to provide a simple callback system in `PreTrainedTokenizer` with callbacks `prepare_tokens` and `prepare_text` provided by the implementing Tokenizer class which takes into account its own flags.\r\n\r\nYet it requires a bigger modification of code and a bit more reflection (I can propose a PR on this if we agree on something).\r\nFor now, an immediate solution to current issue would be to force BertTokenizer to lower_case its tokens by construction: \r\n```python\r\nclass BertTokenizer(PreTrainedTokenizer):\r\n...\r\n\r\n def __init__(self, vocab_file, do_lower_case=True, do_basic_tokenize=True, never_split=None,\r\n unk_token=\"[UNK]\", sep_token=\"[SEP]\", pad_token=\"[PAD]\", cls_token=\"[CLS]\",\r\n mask_token=\"[MASK]\", tokenize_chinese_chars=True, **kwargs):\r\n\r\n if do_lower_case:\r\n unk_token, sep_token, pad_token, cls_token, mask_token = unk_token.lower(), sep_token.lower(), pad_token.lower(), cls_token.lower(), mask_token.lower()\r\n\r\n super(BertTokenizer, self).__init__(unk_token=unk_token, sep_token=sep_token,\r\n pad_token=pad_token, cls_token=cls_token,\r\n mask_token=mask_token, **kwargs)\r\n```\r\n\r\nWDYT?", "Should have been fixed with #2051", "I confirm it should solve the issue! It introduces a bit more external logic about `do_lower_case` in `PreTrainedTokenizer` as I explained. It's not critical but keep in mind for the future, there are solutions to improve that in the code ;)", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2051?src=pr&el=h1) Report\n> Merging [#2051](https://codecov.io/gh/huggingface/transformers/pull/2051?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/5bfcd0485ece086ebcbed2d008813037968a9e58?src=pr&el=desc) will **increase** coverage by `<.01%`.\n> The diff coverage is `100%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/2051?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #2051 +/- ##\n==========================================\n+ Coverage 83.58% 83.58% +<.01% \n==========================================\n Files 105 105 \n Lines 15568 15574 +6 \n==========================================\n+ Hits 13012 13018 +6 \n Misses 2556 2556\n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/2051?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [transformers/tests/tokenization\\_tests\\_commons.py](https://codecov.io/gh/huggingface/transformers/pull/2051/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rlc3RzL3Rva2VuaXphdGlvbl90ZXN0c19jb21tb25zLnB5) | `100% <100%> (ø)` | :arrow_up: |\n| [transformers/tokenization\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/2051/diff?src=pr&el=tree#diff-dHJhbnNmb3JtZXJzL3Rva2VuaXphdGlvbl91dGlscy5weQ==) | `91.96% <100%> (+0.09%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/2051?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/2051?src=pr&el=footer). Last update [5bfcd04...0025a20](https://codecov.io/gh/huggingface/transformers/pull/2051?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "@LysandreJik @thomwolf my last PR seems to have introduced a nasty bug, which got into the most recent release. Could one of you (or someone else at 🤗) review this PR, which should fix it? Sorry for the regression and inconvenience :disappointed: ", "Indeed, this is an issue! No worries, bugs happen.\r\n\r\nUsing a regex may be a bit slow but we'll merge this as to fix the bug, and think of optimization afterward.", ">Using a regex may be a bit slow but we'll merge this as to fix the bug, and think of optimization afterward.\r\n\r\n@LysandreJik true! It just seemed like the fastest way to get to some fix for now. I can help improve it later if necessary :) " ]

[]

[]

[ "ping author @louismartin :)", "Hi @loretoparisi, \r\nWe sampled 10**7 lines randomly from the pretraining corpus.\r\nThe size of the vocabulary was chosen to somewhat match the original BERT paper which used a 30k wordpiece vocabulary, so yes it's mostly arbitrary. ", "@louismartin thanks a lot for the details. I was wondering about the 32k size if this could have been biased by the language...w", "Yes maybe there is a more adapted vocabulary size, we did not investigate that :)\r\n\r\nCan I close the issue now?", "Adding a reference to https://github.com/google/sentencepiece/issues/415", "@louismartin: \r\n> We sampled 10**7 lines randomly from the pretraining corpus.\r\n\r\nMay I ask how did you come up with that number? I'm trying to figure out how many lines I should select to train a model. Assuming I have access to 1 billion rows of ngrams with mean length of 7 words; I'm not sure how many random lines/ngrams would be enough to train a tokenizer with fixed vocab of size 50k?", "I think this is more of a resource allocation question. How much time or compute do you want to allocate to training your tokenizer? Alternate phrasing: why wouldn't you train your tokenizer on the full corpus?", "@julien-c Well sure, that's a valid point. In theory I can train a tokenizer on full corpus by setting the fixed size for the vocabulary. It will just take more and more time (& possibly more compute resources) with increased size of dataset. I was wondering if there's any correlation between quality of fixed threshold vocabulary generated from increasing size of training dataset. I can see that this may be task dependent and requires iterative experiments. Is there any paper that I can look into regarding this? Thanks! ", "PS: did you check out [`tokenizers`](https://github.com/huggingface/tokenizers)? It is pretty fast 😄 \r\n\r\nI've trained a byte-level BPE on 10 GB of text in ~15 minutes.", "@julien-c are you suggesting that, thanks to 🤗 amazing library `tokenizer` we could potentially train the sentence piece tokenizer without setting up a boundary?\r\nThis means that, potentially, current models could improve a lot: from 32K subwords tokens to let's say 1M, what will happen?", "Those are two different things: size of vocab, and size of corpus that you train your tokenizer on.", "@julien-c that's true, and it also seems to have no clear relation in terms of final overall accuracy. Let's say we take as metrics the PPL, and we consider a fixed corpus size and vary the vocab size in batches of 8K, like: 8K, 16K, 32K, 64K, 128K, until we are closer to the whole size of the non unique tokens vocabulary. What will be the related PPL for each training?\r\n\r\n(PPL or BLEU, or other...)" ]

[]

[]

[ "with referring to that file https://github.com/huggingface/transformers/blob/master/examples/distillation/requirements.txt\r\n\r\nrun:\r\n`pip install -r requirements.txt`", "> with referring to that file https://github.com/huggingface/transformers/blob/master/examples/distillation/requirements.txt\r\n> \r\n> run:\r\n> `pip install -r requirements.txt`\r\n\r\nthx, got it~" ]

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

[]

[ "Hi,\r\nThe first ones are loaded and there is currently no simple way to control this.", "**Is there any evidence than the first layers is the best choice when reducing the number of layers ?**\r\n\r\nFor example in your article about Distil-Bert, you chose to initialize the student by taking the even layers. Why so ?", "> For example in your article about Distil-Bert, you chose to initialize the student by taking the even layers. Why so ?\r\n\r\nIt empirically produces stronger performance.\r\nThere are some other empirical evidences in [this paper](https://arxiv.org/abs/1909.11556) from Angela Fan, Edouard Grave and Armand Joulin.", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "I also want to ask this " ]

[]

[]

[ "Oops, that appears to be my fault. Should be a quick fix though, so I'll try to make a PR on it right away. Sorry about that! :grimacing: ", "Thanks man! :) yeah no worries, thought it may be a good idea to report haha" ]