Model fails to load using transformers with trust_remote_code=True
#40
by
Pratikk25
- opened
I’m attempting to load the MiniMaxAI/MiniMax-M2 model using the Hugging Face transformers library. However, the model initialization fails with the following error:
OSError: Could not locate ..processing_utils.py inside MiniMaxAI/MiniMax-M2.
This occurs whether I use AutoModelForCausalLM directly.
CUDA version: 12.1.1
Python version: 3.10
Transformers version: 4.46.3
Execution environment: NVIDIA deep learning container (Ubuntu 22.04 base)
Model variant: MiniMaxAI/MiniMax-M2
We have verified this on the latest version of the Transformers module, v4.57.1. Please update the Transformers module and try it out.
Sorry, there is an error in the modeling_minimax_m2.py file. I will fix it as soon as possible.
I've fixed it, please try it!