Search results

liger_kernel with Qwen3 Embedding fine-tuning - KeyError: 'last_hidden_state'

Status: Open.

#5219 In modelscope/ms-swift;

· MLKoz2 opened on Aug 1, 2025

KIMI VL SFT ERROR

Status: Open.

#5218 In modelscope/ms-swift;

· MooMoo-Yang opened on Aug 1, 2025

关于Support deepspeed-AutoTP的疑问

Status: Open.

#5217 In modelscope/ms-swift;

· effort-yq opened on Aug 1, 2025

多轮对话是否支持自定义选择某几轮不计算loss

Status: Open.

#5215 In modelscope/ms-swift;

· zht8506 opened on Aug 1, 2025

多机GRPO训练QwenVL模型时swanlab不能log，colocate模式

Status: Open.

#5213 In modelscope/ms-swift;

· Zillwang opened on Aug 1, 2025

grpo训练的时候训练几个epoch就会卡住

Status: Open.

#5212 In modelscope/ms-swift;

· wuxi-dixi opened on Aug 1, 2025

如何在sft过程中获得ROUGE和BLEU指标？

Status: Open.

#5211 In modelscope/ms-swift;

· HT-NEKO opened on Aug 1, 2025

Support Dr. GRPO

Status: Open.

#5210 In modelscope/ms-swift;

· dipta007 opened on Aug 1, 2025

Does ms-swift support B200 GPU?

Status: Open.

#5209 In modelscope/ms-swift;

· JhaceLam opened on Aug 1, 2025

Internvl3 DPO微调 模型加载再保存后性能下降

Status: Open.

#5206 In modelscope/ms-swift;

· dle666 opened on Aug 1, 2025

qwen reranker训练数据能带think么

Status: Open.

#5205 In modelscope/ms-swift;

· SuiJiGuoChengSuiJiGuo opened on Aug 1, 2025

LoRA Model Separate Export During HF Conversion

Status: Open.

#5204 In modelscope/ms-swift;

· cxz1418 opened on Aug 1, 2025