Skip to content

issues Search Results · repo:huggingface/transformers language:Python

Filter by

16k results
 (73 ms)

16k results

inhuggingface/transformers (press backspace or delete to remove)

huggingface/transformers
a logic error in _preprocess function of Qwen2VLImageProcessor Class

System Info in the _preprocess function of Qwen2VLImageProcessor Class(https://github.com/huggingface/transformers/blob/348f3285c5114159d2ff4933b4b8ae36866d01a7/src/transformers/models/qwen2_vl/image_processing_qwen2_vl.py#L278), ...
bug
  • InsaneGe
  • Opened 
    1 hour ago
  • #37064

huggingface/transformers
Incorrect calculation of strides leading to loss of param data upon tensor parallel use while sliced model loading

System Info - transformers version: 4.50.2 - Python version: 3.12.0 - Huggingface_hub version: 0.29.3 - Safetensors version: 0.5.3 - Accelerate version: 1.5.2 - Accelerate config: not found ...
bug
  • kmehant
  • Opened 
    11 hours ago
  • #37051

huggingface/transformers
AutoTrain Unsloth support

System Info Any system (incl. Google Colab) Certain versions (incl. 4.50.0 and up) produce this behavior The version that Colab has installed by default works fine Reproduction #@title 🤗 AutoTrain ...
bug
  • urroxyz
  • Opened 
    12 hours ago
  • #37050

huggingface/transformers
Persistent generation issues with MT5 models (base and fine-tuned) across environments

I m experiencing consistent text generation failures with both pretrained google/mt5-base and custom fine-tuned MT5 models across multiple environments (local machines, Google Colab). The models produce ...
  • Elpharran
  • Opened 
    13 hours ago
  • #37048

huggingface/transformers
Optionality of `attention_mask` argument in Attention classes/functions.

In the current stable version of transformers attention_mask argument is annotated as Optional[torch.Tensor] (see for example modeling_llama.py). However, in fact it is a required argument. At the same ...
  • Godofnothing
  • Opened 
    14 hours ago
  • #37046

huggingface/transformers
Latest TorchAO config breaks serialization

PR #36526 breaks https://github.com/huggingface/transformers/blob/49b5ab6a27511de5168c72e83318164f1b4adc43/tests/quantization/torchao_integration/test_torchao.py#L345 Originally posted by @airMeng in ...
  • airMeng
  • 4
  • Opened 
    22 hours ago
  • #37035

huggingface/transformers
TypeError: llama_flash_attn_forward() got an unexpected keyword argument 'cache_position'

System Info transformers==4.50.0 torch==2.4.0+cu121 flash_attn==2.7.4.post1 Who can help? @ArthurZucker @muellerzr @SunMarc Information - [ ] The official example scripts - [x] My own modified ...
bug
  • buaaxiejun
  • Opened 
    22 hours ago
  • #37030

huggingface/transformers
add MiniCPM-o

Model description As discussed in #31836, I would like to add the miniCPM-o model. I believe the miniCPM family has had a significant impact on the LMM, LLM field. Currently, the miniCPM-o code is uploaded ...
New model
  • jp1924
  • 4
  • Opened 
    23 hours ago
  • #37029

huggingface/transformers
run_mim.py script from image-pretraining example is not working

System Info - transformers version: 4.51.0.dev0 - Platform: Linux-5.15.0-73-generic-x86_64-with-glibc2.35 - Python version: 3.12.9 - Huggingface_hub version: 0.29.3 - Safetensors version: 0.5.3 ...
bug
  • jafraustro
  • Opened 
    yesterday
  • #37020

huggingface/transformers
SwitchTransformer: Initialization of tensor to collect expert results is incorrect for dropped tokens (from ML POV)

System Info This is a about a logical bug from ML point of view. It will not result in crashes but influence model behavior significantly. In the transformers code of SwitchTransfomer, we initialize ...
bug
  • mario-aws
  • 1
  • Opened 
    yesterday
  • #37017
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Press the
/
key to activate the search input again and adjust your query.
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Press the
/
key to activate the search input again and adjust your query.
Issue search results · GitHub