423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-08-03 20:22:49 +08:00
Code Issues Packages Projects Releases Wiki Activity

tiny fix

Browse Source 
Former-commit-id: 81167cd19dac4926da9bd259f4e3cb064c22825c
This commit is contained in:
hoshi-hiyouga 2023-12-13 17:32:36 +08:00
parent 3f8d33695c
commit fa99ead86b
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
src/llmtuner/model/utils.py
View File

@ -177,12 +177,12 @@ def resize_embedding_layer(model: "PreTrainedModel", tokenizer: "PreTrainedToken
r"""
Resize token embeddings.
"""
if not isinstance(model.get_output_embeddings(), torch.nn.Linear):
logger.warning("Current model does not support resizing token embeddings.")
return
current_embedding_size = model.get_input_embeddings().weight.size(0)
if len(tokenizer) > current_embedding_size:
if not isinstance(model.get_output_embeddings(), torch.nn.Linear):
logger.warning("Current model does not support resizing token embeddings.")
return
model.resize_token_embeddings(len(tokenizer), pad_to_multiple_of=64)
new_embedding_size = model.get_input_embeddings().weight.size(0)
logger.info("Resized token embeddings from {} to {}.".format(current_embedding_size, new_embedding_size))