riaz.somc/transformers
1
0
Fork 0
mirror of https://github.com/huggingface/transformers.git synced 2025-08-02 11:11:05 +06:00
Code Issues Actions 3 Packages Projects Releases Wiki Activity

Fix issue of canine forward requiring input_ids anyway (#26290)

Browse Source 
* fix issue of canine forward requires input_ids anyway

The `forward` requires `input_ids` for deriving other variables in all cases. Change this to use the given one between `input_ids` and `inputs_embeds`

* fix canine forward

The current `forward` requires (the shape of) `input_ids` for deriving other variables whenever `input_ids` or `inputs_embeds` is provided. Change this to use the given one instead of `input_ids` all the time.

* fix format

* fix format
This commit is contained in:
marcmk6 2023-10-02 17:06:40 +08:00 committed by GitHub
parent 7d77d7f79c
commit 6d02ca4bb9
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 3 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
src/transformers/models/canine/modeling_canine.py
View File

@ -1169,7 +1169,9 @@ class CanineModel(CaninePreTrainedModel):
# Contextualize character embeddings using shallow Transformer.
# We use a 3D attention mask for the local attention.
# `input_char_encoding`: shape (batch_size, char_seq_len, char_dim)
char_attention_mask = self._create_3d_attention_mask_from_input_mask(input_ids, attention_mask)
char_attention_mask = self._create_3d_attention_mask_from_input_mask(
input_ids if input_ids is not None else inputs_embeds, attention_mask
)
init_chars_encoder_outputs = self.initial_char_encoder(
input_char_embeddings,
attention_mask=char_attention_mask,