-
Notifications
You must be signed in to change notification settings - Fork 27.3k
Issues: huggingface/transformers
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Author
Label
Projects
Milestones
Assignee
Sort
Issues list
StopStringCriteria relies on
len(tokenizer)==model.config.vocab_size, leading to index errors
bug
#35244
opened Dec 12, 2024 by
Kripner
2 of 4 tasks
Shape mismatch in RoPE embeddings gpt_neox model when rotary_ndims is odd
bug
#35233
opened Dec 12, 2024 by
mseeger
2 of 4 tasks
if split_special_tokens==True，fast_tokenizer is slower than slow_tokenizer
bug
#35227
opened Dec 12, 2024 by
gongel
4 tasks
logged loss is not correct with gradient accumulation
bug
#35204
opened Dec 11, 2024 by
Jack47
2 of 4 tasks
gradient calculation is not correct with gradient accumulation in Pretrain
bug
#35203
opened Dec 11, 2024 by
Jack47
2 of 4 tasks
PaliGemma2 Processor returns wrong labels array when <image> token is present in
text
bug
#35200
opened Dec 11, 2024 by
probicheaux
2 of 4 tasks
QuantizedCache first token processing is counterintuitive / worse than in papers
bug
Generation
Quantization
#35185
opened Dec 10, 2024 by
goodevening13
2 of 4 tasks
Incorrect file structure in convert_mask2former_original_pytorch_checkpoint_to_pytorch.py?
bug
#35178
opened Dec 10, 2024 by
yjwnb6
2 of 4 tasks
Detokenization discrepancy with Llama3.1
bug
#35175
opened Dec 9, 2024 by
AbrahamSanders
2 of 4 tasks
LlavaForConditionalGeneration._merge_input_ids_with_image_features throws error
bug
Multimodal
WIP
Label your PR/Issue with WIP for some long outstanding Issues/PRs that are work in progress
#35169
opened Dec 9, 2024 by
NicolasDrapier
4 tasks done
DynamicCache does not support variable lengths, except for FA2
bug
#35168
opened Dec 9, 2024 by
SimJeg
2 of 4 tasks
Mimi model gives different outputs when using batch encode vs single encode
bug
#35166
opened Dec 9, 2024 by
avishaiElmakies
2 of 4 tasks
Calling Trainer.create_model_card() with an empty dataset list causes an IndexError
bug
#35163
opened Dec 9, 2024 by
FelixSchneiderZoom
4 tasks
Impossible to change attention implementation
bug
#35153
opened Dec 8, 2024 by
stoical07
2 of 4 tasks
how to load the weight of decoder.embed_tokens.weight seperately from the shared weight?
bug
#35152
opened Dec 8, 2024 by
SoSongzhi
1 of 4 tasks
RuntimeError: shape '[1, 3098, 6, 5, 128]' is invalid for input of size 12689408
bug
#35146
opened Dec 7, 2024 by
LuchenZhou
4 tasks
resizing token embeddings causes output embedding to be reinitialized in
post_init when tie_word_embedding is False
bug
#35141
opened Dec 7, 2024 by
avishaiElmakies
2 of 4 tasks
Special token ids are not longer typed properly in 4.47.0
bug
#35126
opened Dec 6, 2024 by
chanind
1 of 4 tasks
(sort of) a bug with token offsets: some special tokens have (0, 0) offsets regardless of their position in the document
bug
#35125
opened Dec 6, 2024 by
ViktorooReps
4 tasks
Previous Next
ProTip!
Updated in the last three days: updated:>2024-12-10.