From 814a3f4e017020d67ce69048e2647e7bf5d0784f Mon Sep 17 00:00:00 2001 From: Lysandre Date: Wed, 21 Aug 2019 14:11:14 -0400 Subject: [PATCH] Removed `attention_mask` from GPT-2 and GPT documentation. Corrected `multiple_choice_labels` to actual name `mc_labels` --- pytorch_transformers/modeling_gpt2.py | 10 +--------- pytorch_transformers/modeling_openai.py | 12 ++---------- 2 files changed, 3 insertions(+), 19 deletions(-) diff --git a/pytorch_transformers/modeling_gpt2.py b/pytorch_transformers/modeling_gpt2.py index 9022048d6d..2980cf269a 100644 --- a/pytorch_transformers/modeling_gpt2.py +++ b/pytorch_transformers/modeling_gpt2.py @@ -408,10 +408,6 @@ GPT2_INPUTS_DOCSTRING = r""" Inputs: list of ``torch.FloatTensor`` (one for each layer): that contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see `past` output below). Can be used to speed up sequential decoding. - **attention_mask**: (`optional`) ``torch.FloatTensor`` of shape ``(batch_size, sequence_length)``: - Mask to avoid performing attention on padding token indices. - Mask values selected in ``[0, 1]``: - ``1`` for tokens that are NOT MASKED, ``0`` for MASKED tokens. **head_mask**: (`optional`) ``torch.FloatTensor`` of shape ``(num_heads,)`` or ``(num_layers, num_heads)``: Mask to nullify selected heads of the self-attention modules. Mask values selected in ``[0, 1]``: @@ -642,10 +638,6 @@ class GPT2DoubleHeadsModel(GPT2PreTrainedModel): list of ``torch.FloatTensor`` (one for each layer): that contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see `past` output below). Can be used to speed up sequential decoding. - **attention_mask**: (`optional`) ``torch.FloatTensor`` of shape ``(batch_size, num_choices, sequence_length)``: - Mask to avoid performing attention on padding token indices. - Mask values selected in ``[0, 1]``: - ``1`` for tokens that are NOT MASKED, ``0`` for MASKED tokens. **head_mask**: (`optional`) ``torch.FloatTensor`` of shape ``(num_heads,)`` or ``(num_layers, num_heads)``: Mask to nullify selected heads of the self-attention modules. Mask values selected in ``[0, 1]``: @@ -656,7 +648,7 @@ class GPT2DoubleHeadsModel(GPT2PreTrainedModel): Indices are selected in ``[-1, 0, ..., config.vocab_size]`` All labels set to ``-1`` are ignored (masked), the loss is only computed for labels in ``[0, ..., config.vocab_size]`` - **multiple_choice_labels**: (`optional`) ``torch.LongTensor`` of shape ``(batch_size)``: + **mc_labels**: (`optional`) ``torch.LongTensor`` of shape ``(batch_size)``: Labels for computing the multiple choice classification loss. Indices should be in ``[0, ..., num_choices]`` where `num_choices` is the size of the second dimension of the input tensors. (see `input_ids` above) diff --git a/pytorch_transformers/modeling_openai.py b/pytorch_transformers/modeling_openai.py index cfea768736..690aa7812b 100644 --- a/pytorch_transformers/modeling_openai.py +++ b/pytorch_transformers/modeling_openai.py @@ -415,11 +415,7 @@ OPENAI_GPT_INPUTS_DOCSTRING = r""" Inputs: **token_type_ids**: (`optional`) ``torch.LongTensor`` of shape ``(batch_size, sequence_length)``: A parallel sequence of tokens (can be used to indicate various portions of the inputs). The embeddings from these tokens will be summed with the respective token embeddings. - Indices are selected in the vocabulary (unlike BERT which has a specific vocabulary for segment indices). - **attention_mask**: (`optional`) ``torch.FloatTensor`` of shape ``(batch_size, sequence_length)``: - Mask to avoid performing attention on padding token indices. - Mask values selected in ``[0, 1]``: - ``1`` for tokens that are NOT MASKED, ``0`` for MASKED tokens. + Indices are selected in the vocabulary (unlike BERT which has a specific vocabulary for segment indices) **head_mask**: (`optional`) ``torch.FloatTensor`` of shape ``(num_heads,)`` or ``(num_layers, num_heads)``: Mask to nullify selected heads of the self-attention modules. Mask values selected in ``[0, 1]``: @@ -626,10 +622,6 @@ class OpenAIGPTDoubleHeadsModel(OpenAIGPTPreTrainedModel): A parallel sequence of tokens (can be used to indicate various portions of the inputs). The embeddings from these tokens will be summed with the respective token embeddings. Indices are selected in the vocabulary (unlike BERT which has a specific vocabulary for segment indices). - **attention_mask**: (`optional`) ``torch.FloatTensor`` of shape ``(batch_size, num_choices, sequence_length)``: - Mask to avoid performing attention on padding token indices. - Mask values selected in ``[0, 1]``: - ``1`` for tokens that are NOT MASKED, ``0`` for MASKED tokens. **head_mask**: (`optional`) ``torch.FloatTensor`` of shape ``(num_heads,)`` or ``(num_layers, num_heads)``: Mask to nullify selected heads of the self-attention modules. Mask values selected in ``[0, 1]``: @@ -640,7 +632,7 @@ class OpenAIGPTDoubleHeadsModel(OpenAIGPTPreTrainedModel): Indices are selected in ``[-1, 0, ..., config.vocab_size]`` All labels set to ``-1`` are ignored (masked), the loss is only computed for labels in ``[0, ..., config.vocab_size]`` - **multiple_choice_labels**: (`optional`) ``torch.LongTensor`` of shape ``(batch_size)``: + **mc_labels**: (`optional`) ``torch.LongTensor`` of shape ``(batch_size)``: Labels for computing the multiple choice classification loss. Indices should be in ``[0, ..., num_choices]`` where `num_choices` is the size of the second dimension of the input tensors. (see `input_ids` above)