Skip to content

Conversation

@hmellor
Copy link
Member

@hmellor hmellor commented Dec 1, 2025

  • Qwen2_5_VLForConditionalGeneration, Qwen2VLForConditionalGeneration, JinaVLForRanking - Uses Qwen 2/2.5 VL which recently had the hacky __getattribute__ removed in Let's break Qwen-VL 🚨  huggingface/transformers#42420. This means that you cannot just get config fields from the root config and expect them to get fetched from the text config
  • ModernBertModel, ModernBertForTokenClassification, ModernBertForSequenceClassification - Instantiated its RoPE layer in a weird way. It now uses the get_rope method and uses the layer_types field which is now present in Transformers v5

@mergify mergify bot added the qwen Related to Qwen models label Dec 1, 2025
@hmellor hmellor marked this pull request as ready for review December 1, 2025 18:36
@hmellor hmellor requested a review from sighingnow as a code owner December 1, 2025 18:36
@chatgpt-codex-connector
Copy link

Codex usage limits have been reached for code reviews. Please check with the admins of this repo to increase the limits by adding credits.

@Isotr0py Isotr0py enabled auto-merge (squash) December 2, 2025 02:47
@github-actions github-actions bot added the ready ONLY add when PR is ready to merge/full CI is needed label Dec 2, 2025
@Isotr0py Isotr0py merged commit f5b0846 into vllm-project:main Dec 2, 2025
52 checks passed
@hmellor hmellor deleted the fix-some-tests branch December 2, 2025 10:05
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

qwen Related to Qwen models ready ONLY add when PR is ready to merge/full CI is needed

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants