Skip to content

support q3.6-27b#178

Merged
tastelikefeet merged 3 commits intomodelscope:mainfrom
tastelikefeet:feat/q36-27b
Apr 22, 2026
Merged

support q3.6-27b#178
tastelikefeet merged 3 commits intomodelscope:mainfrom
tastelikefeet:feat/q36-27b

Conversation

@tastelikefeet
Copy link
Copy Markdown
Collaborator

PR type

  • Bug Fix
  • New Feature
  • Document Updates
  • More Models or Datasets Support

PR information

Write the detail information belongs to this PR.

Experiment results

Paste your experiment result here(if needed).

Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist Bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code Review

This pull request updates the default base model from the MoE-based Qwen3.6-35B-A3B to the dense Qwen3.6-27B across documentation, configuration files, examples, and notebooks. Key changes include updating model IDs, adjusting server configurations for the new model's context length and memory requirements, and replacing Megatron-specific export steps in notebooks with a more general PEFT-based LoRA merging approach. Review feedback highlights several instances where model descriptions or link texts were not fully updated to reflect the new model architecture, as well as environment-specific metadata changes in a notebook that should be reverted for better portability.

Comment thread README.md Outdated
Comment thread README_ZH.md Outdated
Comment thread cookbook/rl/short_math_grpo_multi_lora.py Outdated
Comment thread notebook/self_cognition.ipynb
Comment thread notebook/self_cognition.ipynb
@tastelikefeet tastelikefeet merged commit 8461131 into modelscope:main Apr 22, 2026
1 of 3 checks passed
tastelikefeet added a commit that referenced this pull request Apr 22, 2026
(cherry picked from commit 8461131)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants