- 
          
 - 
                Notifications
    
You must be signed in to change notification settings  - Fork 11k
 
          [Misc] Upgrade transformers to 4.49
          #13602
        
          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
Conversation
Signed-off-by: Roger Wang <ywang@roblox.com>
| 
           👋 Hi! Thank you for contributing to the vLLM project. 💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels. Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run  Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging. To run CI, PR reviewers can either: Add  🚀  | 
    
| 
           
  | 
    
| 
           We need to update   | 
    
| 
           IIRC, multiple people reported that the fast processor does not match the original one. Did we consider this issue?  | 
    
          
 @WoosukKwon Yea I tried to see if the numerical differences would lead to actual differences in the result, but it looks like that's not the case since the tests all passed (with different image size factors). I'm also okay with us not including   | 
    
| 
           @ywang96 Got it. I'd personally prefer to not use it until we are certain about the correctness.  | 
    
          
 Sounds good - I'll revert the change  | 
    
| 
           Since #13617 we no longer need to manually update   | 
    
| 
           In a separate PR which also updates transformers I'm seeing some  It'd be good to check if this happens here too.  | 
    
          
 @hmellor Thanks for the heads up!  | 
    
| 
           Rebuilding  | 
    
This PR upgrades the version of transformers to 4.49 for out-of-box support of Qwen2.5-VL