num_assistant_tokens not configured in assistant model

[geoffrey-young]

System Info

num_assistant_tokens and other parameters intended for the assistant model are not properly passed down, leaving the assistant model with default values.

outputs = model.generate(**inputs, assistant_model=assistant_model, num_assistant_tokens=5)
print(f'num_assistant_tokens: {assistant_model.generation_config.num_assistant_tokens}')

This behavior is present both in the official 4.56.0 as well as main (currently 4.57.0.dev0/bb45d36)

Fix in PR #40740

• transformers version: 4.57.0.dev0
• Platform: macOS-15.6.1-arm64-arm-64bit
• Python version: 3.12.11
• Huggingface_hub version: 0.34.4
• Safetensors version: 0.6.2
• Accelerate version: 1.10.1
• Accelerate config: not found
• DeepSpeed version: not installed
• PyTorch version (accelerator?): 2.8.0 (NA)
• Tensorflow version (GPU?): not installed (NA)
• Flax version (CPU?/GPU?/TPU?): not installed (NA)
• Jax version: not installed
• JaxLib version: not installed
• Using distributed or parallel set-up in script?:?

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

outputs = model.generate(**inputs, assistant_model=assistant_model, num_assistant_tokens=5)
print(f'num_assistant_tokens: {assistant_model.generation_config.num_assistant_tokens}')

see also this Collab: https://colab.research.google.com/drive/1BIY6yklrsarrPmXWrciV1chAqV5I9-wo?usp=sharing

Expected behavior

num_assistant_tokens should be 5 not the default 20