SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

fix docs of past_key_values (#12049)

Browse Source
This commit is contained in:
Suraj Patil
2021-06-07 15:24:03 +05:30
committed by GitHub
parent 3857f2b4e3
commit 185122ef22
11 changed files with 230 additions and 62 deletions
Download Patch File Download Diff File Expand all files Collapse all files

35
templates/adding_a_new_model/cookiecutter-template-{{cookiecutter.modelname}}/modeling_{{cookiecutter.lowercase_modelname}}.py
View File

@@ -1063,11 +1063,20 @@ class {{cookiecutter.camelcase_modelname}}ForCausalLM({{cookiecutter.camelcase_m
- 1 for tokens that are **not masked**,
- 0 for tokens that are **masked**.
past_key_values (:obj:`tuple(tuple(torch.FloatTensor))` of length :obj:`config.n_layers` with each tuple having 4 tensors of shape :obj:`(batch_size, num_heads, sequence_length - 1, embed_size_per_head)`):
Contains precomputed key and value hidden states of the attention blocks. Can be used to speed up decoding.
If :obj:`past_key_values` are used, the user can optionally input only the last :obj:`decoder_input_ids`
past_key_values (:obj:`tuple(tuple(torch.FloatTensor))`, `optional`, returned when ``use_cache=True`` is passed or when ``config.use_cache=True``):
Tuple of :obj:`tuple(torch.FloatTensor)` of length :obj:`config.n_layers`, with each tuple having 2
tensors of shape :obj:`(batch_size, num_heads, sequence_length, embed_size_per_head)`) and 2 additional
tensors of shape :obj:`(batch_size, num_heads, encoder_sequence_length, embed_size_per_head)`. The two
additional tensors are only required when the model is used as a decoder in a Sequence to Sequence
model.
Contains pre-computed hidden-states (key and values in the self-attention blocks and in the
cross-attention blocks) that can be used (see :obj:`past_key_values` input) to speed up sequential
decoding.
If :obj:`past_key_values` are used, the user can optionally input only the last ``decoder_input_ids``
(those that don't have their past key value states given to this model) of shape :obj:`(batch_size, 1)`
instead of all :obj:`decoder_input_ids` of shape :obj:`(batch_size, sequence_length)`.
instead of all ``decoder_input_ids`` of shape :obj:`(batch_size, sequence_length)`.
labels (:obj:`torch.LongTensor` of shape :obj:`(batch_size, sequence_length)`, `optional`):
Labels for computing the left-to-right language modeling loss (next word prediction). Indices should be in
``[-100, 0, ..., config.vocab_size]`` (see ``input_ids`` docstring) Tokens with indices set to ``-100`` are
@@ -2089,8 +2098,13 @@ class {{cookiecutter.camelcase_modelname}}PreTrainedModel(PreTrainedModel):
:obj:`attentions`) :obj:`last_hidden_state` of shape :obj:`(batch_size, sequence_length, hidden_size)`,
`optional`) is a sequence of hidden-states at the output of the last layer of the encoder. Used in the
cross-attention of the decoder.
past_key_values (:obj:`Tuple[Tuple[torch.Tensor]]` of length :obj:`config.n_layers` with each tuple having 2 tuples each of which has 2 tensors of shape :obj:`(batch_size, num_heads, sequence_length - 1, embed_size_per_head)`):
Contains precomputed key and value hidden-states of the attention blocks. Can be used to speed up decoding.
past_key_values (:obj:`tuple(tuple(torch.FloatTensor))`, `optional`, returned when ``use_cache=True`` is passed or when ``config.use_cache=True``):
Tuple of :obj:`tuple(torch.FloatTensor)` of length :obj:`config.n_layers`, with each tuple having 2 tensors
of shape :obj:`(batch_size, num_heads, sequence_length, embed_size_per_head)`) and 2 additional tensors of
shape :obj:`(batch_size, num_heads, encoder_sequence_length, embed_size_per_head)`.
Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention
blocks) that can be used (see :obj:`past_key_values` input) to speed up sequential decoding.
If :obj:`past_key_values` are used, the user can optionally input only the last :obj:`decoder_input_ids`
(those that don't have their past key value states given to this model) of shape :obj:`(batch_size, 1)`
@@ -2429,8 +2443,13 @@ class {{cookiecutter.camelcase_modelname}}Decoder({{cookiecutter.camelcase_model
- 1 indicates the head is **not masked**,
- 0 indicates the head is **masked**.
past_key_values (:obj:`Tuple[Tuple[torch.Tensor]]` of length :obj:`config.n_layers` with each tuple having 2 tuples each of which has 2 tensors of shape :obj:`(batch_size, num_heads, sequence_length - 1, embed_size_per_head)`):
Contains precomputed key and value hidden-states of the attention blocks. Can be used to speed up
past_key_values (:obj:`tuple(tuple(torch.FloatTensor))`, `optional`, returned when ``use_cache=True`` is passed or when ``config.use_cache=True``):
Tuple of :obj:`tuple(torch.FloatTensor)` of length :obj:`config.n_layers`, with each tuple having 2
tensors of shape :obj:`(batch_size, num_heads, sequence_length, embed_size_per_head)`) and 2 additional
tensors of shape :obj:`(batch_size, num_heads, encoder_sequence_length, embed_size_per_head)`.
Contains pre-computed hidden-states (key and values in the self-attention blocks and in the
cross-attention blocks) that can be used (see :obj:`past_key_values` input) to speed up sequential
decoding.
If :obj:`past_key_values` are used, the user can optionally input only the last