Search results

python模式使用vllm引擎推理，如何使用多卡推理

Status: Open.

#6472 In modelscope/ms-swift;

· huhk-sysu opened on Nov 6, 2025

megatron lora续训Qwen3-Next-80B-A3B-Instruct 报错

Status: Open.

#6471 In modelscope/ms-swift;

· xiaotianns opened on Nov 6, 2025

How to use llm as a reward func？

Status: Open.

#6470 In modelscope/ms-swift;

· happened opened on Nov 6, 2025

请问如何训练Qwen3VL-30B-A2B模型的function call能力

Status: Open.

#6468 In modelscope/ms-swift;

· zzk6626 opened on Nov 6, 2025

agent训练数据function的参数混乱

bugSomething isn't workingSomething isn't working

Status: Open.

#6467 In modelscope/ms-swift;

· buer103 opened on Nov 6, 2025

how to save the tokenized data using save_to_disk

Status: Open.

#6466 In modelscope/ms-swift;

· BaiMeiyingxue opened on Nov 6, 2025

NaN in local grad norm after verion bumping to use mcore bridge

Status: Open.

#6465 In modelscope/ms-swift;

· zhoutong-hai opened on Nov 6, 2025

DPO训练报错

Status: Open.

#6464 In modelscope/ms-swift;

· BeerTai opened on Nov 6, 2025

用gptq量化lora微调合并后的qwen3-next-80b-A3b模型，报错KeyError: 'self_attn.q_proj'

bugSomething isn't workingSomething isn't working

Status: Open.

#6459 In modelscope/ms-swift;

· xiaotianns opened on Nov 6, 2025

jina-reranker-v3

Status: Open.

#6456 In modelscope/ms-swift;

· KunDoer opened on Nov 6, 2025

在NPU lora微调Qwen3-coder-30B Agent能力超级慢

npu

Status: Open.

#6455 In modelscope/ms-swift;

· buer103 opened on Nov 6, 2025

Qwen3 Omni训练合并lora权重失败

Status: Open.

#6454 In modelscope/ms-swift;

· qq31415926 opened on Nov 6, 2025