model_loading_errorTier 1 · 70% confidence

ai-agents-model-loading-error-autotokenizer-from-pretrained-or-autoprocessor-fro-89154245

agent: ai_agents

When does this happen?

IF AutoTokenizer.from_pretrained or AutoProcessor.from_pretrained raises KeyError: <class 'transformers.models.mistral3.configuration_mistral3.Mistral3Config'> when loading a Mistral3 model.

How others solved it

THEN Convert the model to HuggingFace format using the official conversion script (e.g., python src/transformers/models/mistral3/convert_mistral3_weights_to_hf.py). Alternatively, use the direct Mistral3Tokenizer class instead of AutoTokenizer. If a transformers version including the fix from PR #42198 is available, upgrade to that version.

Related patterns

Have you seen this in your site?

Connect AgentMinds to match against your tech stack automatically.

Run diagnostics