[Bug] Setting `pad_token_id` to `eos_token_id`:151645 for open-end generation.

[Hoantrbl]

Checklist

• [x] 1. I have searched related issues but cannot get the expected help.
• [x] 2. The bug has not been fixed in the latest version.
• [x] 3. Please note that if the bug-related issue you submitted lacks corresponding environment info and a minimal reproducible demo, it will be challenging for us to reproduce and resolve the issue, reducing the likelihood of receiving feedback.

Describe the bug

Thanks for your nice work on Internvl3. I found that transformers give the following warning. How can I solve that?

Reproduction

I directly utilize the lmm-evals and its model file ``Internvl2'' to run your latest Internvl3-8B checkpoint.

Environment

torch                         2.6.0
torchvision                   0.21.0
tqdm                          4.67.1
tqdm-multiprocess             0.0.11
traitlets                     5.14.3
transformers                  4.52.0.dev0

Error traceback

[lll2343]

Hi, this shouldn't affect normal generation. If the model repeats or <im_end>, try adding:

"eos_token_id": [
   151645,
   151643
 ]

to your generation_config.json

[Hoantrbl]

Hi, this shouldn't affect normal generation. If the model repeats or <im_end>, try adding:

"eos_token_id": [
   151645,
   151643
 ]

to your generation_config.json

Thanks a lot

[cherishleon]

you can alse use this one

generation_config = dict( max_new_tokens=1024, do_sample=True, eos_token_id=151645, pad_token_id=151645, )