Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Allow overriding eos_token_id #1654

Merged

Conversation

Wovchena
Copy link
Collaborator

Phi3_V eos_token_id has different values for GenerationConfig and Tokenizer. It's required to allow overriding the token_id to align with the sample from model cards.

My patches to the original models are ignored:

  1. https://huggingface.co/microsoft/Phi-3-vision-128k-instruct/discussions/68
  2. https://huggingface.co/microsoft/Phi-3.5-vision-instruct/discussions/35

Phi3_V eos_token_id has different values for GenerationConfig and Tokenizer. It's required to allow overriding the token_id to align with the sample from model cards.

My patches to the original models are ignored:
1. https://huggingface.co/microsoft/Phi-3-vision-128k-instruct/discussions/68
2. https://huggingface.co/microsoft/Phi-3.5-vision-instruct/discussions/35
@Wovchena Wovchena requested a review from yatarkan January 30, 2025 11:43
@github-actions github-actions bot added category: visual language Visual language pipeline category: sampling Sampling / Decoding algorithms labels Jan 30, 2025
@ilya-lavrenov ilya-lavrenov added this to the 2025.1 milestone Jan 30, 2025
@ilya-lavrenov ilya-lavrenov self-assigned this Jan 30, 2025
@andrei-kochin andrei-kochin added this pull request to the merge queue Jan 30, 2025
@github-merge-queue github-merge-queue bot removed this pull request from the merge queue due to failed status checks Jan 30, 2025
@Wovchena Wovchena added this pull request to the merge queue Jan 30, 2025
github-merge-queue bot pushed a commit that referenced this pull request Jan 30, 2025
Phi3_V eos_token_id has different values for GenerationConfig and
Tokenizer. It's required to allow overriding the token_id to align with
the sample from model cards.

My patches to the original models are ignored:
1.
https://huggingface.co/microsoft/Phi-3-vision-128k-instruct/discussions/68
2.
https://huggingface.co/microsoft/Phi-3.5-vision-instruct/discussions/35
@github-merge-queue github-merge-queue bot removed this pull request from the merge queue due to failed status checks Jan 30, 2025
@ilya-lavrenov ilya-lavrenov added this pull request to the merge queue Jan 31, 2025
@github-merge-queue github-merge-queue bot removed this pull request from the merge queue due to failed status checks Jan 31, 2025
@Wovchena Wovchena added this pull request to the merge queue Jan 31, 2025
@github-merge-queue github-merge-queue bot removed this pull request from the merge queue due to failed status checks Jan 31, 2025
@Wovchena Wovchena added this pull request to the merge queue Jan 31, 2025
Merged via the queue into openvinotoolkit:master with commit 3b13a40 Jan 31, 2025
70 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
category: sampling Sampling / Decoding algorithms category: visual language Visual language pipeline
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants