modelscope/swift
ms-swift: Use PEFT or Full-parameter to finetune 250+ LLMs or 35+ MLLMs. (Qwen2, GLM4, Internlm2, Yi, Llama3, Llava, Deepseek, Baichuan2...)
PythonApache-2.0
Issues
- 0
如何批量导出MS数据集为swift能加载的格式
#1041 opened by WSC741606 - 0
是否支持idm-vton模型的微调呢?
#1040 opened by awzhgw - 2
- 3
目前支持多模态模型部署了吗
#987 opened by LRHstudy - 4
自定义数据微调MiniCPM-Llama3-V-2_5报错
#1030 opened by zhudongwork - 5
qwen1.5-32b-chat 使用vllm推理很慢
#986 opened by zhangfan-algo - 2
swift微调多模态大模型后,比如Intern VL 1.5,可以使用lmdeploy部署吗?
#1033 opened by wangdong1992 - 2
自定义数据集报错
#980 opened by Vindicator645 - 0
单机多卡微调报错: Expected all tensors to be on the same device, but found at least two devices, cuda:2 and cuda:3! is:closed
#1022 opened by Elissa0723 - 0
使用appui功能无法指定ip和端口,还是会使用默认配置
#1034 opened by zhangfan-algo - 1
NPU微调报错:ValueError: Your assigned backend {original_backend} is not avaliable, please use {backend}
#1020 opened by FlynnShi - 0
Support RLAIF-V
#981 opened by choyakawa - 0
bitsandbytes was compiled without GPU support
#1031 opened by stellarxxu - 1
react模板放在system和user的区别?
#1025 opened by nauyiahc - 1
可以支持一下InternLM2-Math-Plus-Mixtral8x22B的微调吗
#1019 opened by zhangfan-algo - 2
- 6
- 2
infer 怎么指定输出文件夹?
#957 opened by AlexJJJChen - 12
infer 无法跑完所有data
#1006 opened by AlexJJJChen - 0
多模态模型(qwen_vl_chat)量化失败
#1015 opened by Luccadoremi - 1
- 2
可以支持一下qwen1.5系列模型的预训练吗
#1018 opened by zhangfan-algo - 1
模型并行自我认知微调报错
#1014 opened by wssywh - 8
微调qwen后会循环输出
#1009 opened by sherry085 - 1
采用 --dtype fp16 的方式DPO训练后,无法推理。
#1011 opened by stevezhang88 - 1
使用教程训练会爆cuda错误
#1012 opened by yangtianyu92 - 1
多模态提前将图片处理,然后再训练LLM
#1007 opened by choyakawa - 0
微调minicpmv2时cpu占用率超高
#1008 opened by strawhatboy - 3
awq 微调后如何推理
#1002 opened by hehuang139 - 1
Qwen-VL-Chat-Int4 是否可以full finetune?
#1000 opened by Luccadoremi - 1
cogvlm2添加history报错
#993 opened by LRHstudy - 0
微调量化后qwen1half-14b-chat-gptq-int8推理时向量报错RuntimeError: probability tensor contains either `inf`, `nan` or element < 0
#994 opened by AnsongLi - 2
- 0
grok-1载入时间过长
#988 opened by Di-Zayn - 1
支持swift接入数据集的下载和离线训练。
#982 opened by stitchshaw - 0
微调internvl-V1-5一直报warning
#985 opened by sunzx8 - 6
- 3
模型微调后不停止回答<|im_end|>
#960 opened by againcui - 1
是否支持deepseek-v2-chat-lite model ?
#955 opened by awzhgw - 2
peft加载qwen1half_72b_chat的lora模型报错
#975 opened by jhjiang10 - 6
- 2
能否支持多卡DDP部署
#956 opened by WSC741606 - 1
并发
#963 opened by wuguangshuo - 1
qwen1half-14b-chat-int4使用lora微调后合并模型报错
#966 opened by AnsongLi - 1
怎么做 batch infer 来提高显卡利用率呢?
#968 opened by thesby - 1
New version code issue for Internvl "cannot import name 'BOFTConfig' from 'peft'"
#976 opened by MVP-D77 - 0
使用zero3进行多机多卡全量微调,保存的模型权重不完整
#972 opened by ultrazhl98 - 2
- 4
ValueError: model_type: 'yi-1_5-6b' is not registered
#959 opened by jacnmm4 - 0
ValueError: model_type: 'yi-1_5-6b' is not registered.
#958 opened by jacnmm4