LLama 3.2 vision - unable to convert

[pdufour]

System Info

python3 --version

Python 3.10.15

requirements.txt

huggingface_hub
streamlit
transformers[torch]==4.46.0
onnxruntime==1.19.2
optimum==1.23.2
onnx==1.17.0
onnxconverter-common==1.14.0
tqdm==4.66.5
onnxslim==0.1.35
--extra-index-url https://pypi.ngc.nvidia.com
onnx_graphsurgeon==0.5.2

Who can help?

No response

Information

• The official example scripts
• My own modified scripts

Tasks

• An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
• My own task or dataset (give details below)

Reproduction (minimal, reproducible, runnable)

1. Clone https://github.com/huggingface/transformers.js (Note: I updated the transformers version as seen in above requirements.txt)
2. Run python3 -m scripts.convert --quantize --model_id meta-llama/Llama-3.2-11B-Vision

Gives error:

Traceback (most recent call last):
  File "/opt/homebrew/Cellar/[email protected]/3.10.15/Frameworks/Python.framework/Versions/3.10/lib/python3.10/runpy.py", line 196, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/opt/homebrew/Cellar/[email protected]/3.10.15/Frameworks/Python.framework/Versions/3.10/lib/python3.10/runpy.py", line 86, in _run_code
    exec(code, run_globals)
  File "transformers.js/scripts/convert.py", line 462, in <module>
    main()
  File "transformers.js/scripts/convert.py", line 349, in main
    main_export(**export_kwargs)
  File ".venv/lib/python3.10/site-packages/optimum/exporters/onnx/__main__.py", line 303, in main_export
    model = TasksManager.get_model_from_task(
  File ".venv/lib/python3.10/site-packages/optimum/exporters/tasks.py", line 2071, in get_model_from_task
    model_class = TasksManager.get_model_class_for_task(
  File ".venv/lib/python3.10/site-packages/optimum/exporters/tasks.py", line 1394, in get_model_class_for_task
    raise KeyError(
KeyError: "Unknown task: image-text-to-text. Possible values are: `audio-classification` for AutoModelForAudioClassification, `audio-frame-classification` for AutoModelForAudioFrameClassification, `audio-xvector` for AutoModelForAudioXVector, `automatic-speech-recognition` for ('AutoModelForSpeechSeq2Seq', 'AutoModelForCTC'), `depth-estimation` for AutoModelForDepthEstimation, `feature-extraction` for AutoModel, `fill-mask` for AutoModelForMaskedLM, `image-classification` for AutoModelForImageClassification, `image-segmentation` for ('AutoModelForImageSegmentation', 'AutoModelForSemanticSegmentation'), `image-to-image` for AutoModelForImageToImage, `image-to-text` for AutoModelForVision2Seq, `mask-generation` for AutoModel, `masked-im` for AutoModelForMaskedImageModeling, `multiple-choice` for AutoModelForMultipleChoice, `object-detection` for AutoModelForObjectDetection, `question-answering` for AutoModelForQuestionAnswering, `semantic-segmentation` for AutoModelForSemanticSegmentation, `text-to-audio` for ('AutoModelForTextToSpectrogram', 'AutoModelForTextToWaveform'), `text-generation` for AutoModelForCausalLM, `text2text-generation` for AutoModelForSeq2SeqLM, `text-classification` for AutoModelForSequenceClassification, `token-classification` for AutoModelForTokenClassification, `zero-shot-image-classification` for AutoModelForZeroShotImageClassification, `zero-shot-object-detection` for AutoModelForZeroShotObjectDetection"

Expected behavior

It should be able to handle this type of conversion