Search results

Mcore bridge unable to merge LoRA weights and save as a Hugging Face model.

Status: Open.

#6447 In modelscope/ms-swift;

· Minami-su opened on Nov 5, 2025

On-Policy Distillation train

bugSomething isn't workingSomething isn't working

Status: Open.

#6446 In modelscope/ms-swift;

· RonanDou opened on Nov 5, 2025

grpo异步方式训练时，显存单调上涨，直至OOM

Status: Open.

#6445 In modelscope/ms-swift;

· hlchen23 opened on Nov 5, 2025

是否可以支持KTransformers，优化Moe的微调成本？

Status: Open.

#6444 In modelscope/ms-swift;

· xsg5886779 opened on Nov 5, 2025

[RFC] Support Megatron GKD Training

enhancementNew feature or requestNew feature or request

Status: Open.

#6443 In modelscope/ms-swift;

· chapter544 opened on Nov 5, 2025

微调Qwen2.5-Omni模型的Talker部分并将音频作为label

Status: Open.

#6442 In modelscope/ms-swift;

· ooochen-30 opened on Nov 5, 2025

raise ValueError(f"max_new_tokens must be greater than 0, but is {self.max_new_tokens}.") ValueError: max_new_tokens must be greater than 0, but is -5923.

Status: Open.

#6441 In modelscope/ms-swift;

· SLKAlgs opened on Nov 5, 2025

qwen3-8B:ms-swift/examples/train/think_model/qwen3_demo1.sh测试时报错

Status: Open.

#6439 In modelscope/ms-swift;

· williamyang4978 opened on Nov 5, 2025

qwen3_omni训练时无法单机多卡并行，同一时间只有一个卡工作

Status: Open.

#6438 In modelscope/ms-swift;

· pyssionately opened on Nov 5, 2025

Missing the MTP module when exporting the Qwen3-Next HF model to an MScore model.

enhancementNew feature or requestNew feature or request

Status: Open.

#6436 In modelscope/ms-swift;

· 155394551lzk opened on Nov 5, 2025

How to import callback function into the framework

Status: Open.

#6433 In modelscope/ms-swift;

· Iven-Wu opened on Nov 5, 2025

Support Qwen3-VL and Qwen2.5-VL in the same environment

enhancementNew feature or requestNew feature or request

Status: Open.

#6432 In modelscope/ms-swift;

· thisiskofi opened on Nov 4, 2025