Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / InternLM/xtuner issues and pull requests

#956 - torch编译错误

Issue - State: open - Opened by tcxia 15 days ago

#954 - 对Minicpm3进行了支持

Pull Request - State: closed - Opened by LDLINGLINGLING 18 days ago

#953 - 怎么设置random seed控制train dpo随机性

Issue - State: open - Opened by pingbowen23 19 days ago

#952 - xtuner 微调internLM2.5出错

Issue - State: open - Opened by sakura073 21 days ago - 6 comments

#951 - fix qwen2 tokenizer name

Pull Request - State: open - Opened by amulil 23 days ago

#949 - 请问如何在微调训练时查看数据

Issue - State: open - Opened by liguoyu666 28 days ago

#948 - CheckpointHook开启by_epoch时未保存模型

Issue - State: open - Opened by aizhweiwei 28 days ago - 1 comment

#945 - xtuner如何根据checkpoints继续预训练

Issue - State: open - Opened by aizhweiwei about 1 month ago - 1 comment

#943 - Does xtuner support DPO for InternVL?

Issue - State: open - Opened by fabriceyhc about 1 month ago - 1 comment

#942 - llava-llama3-8b 微调过程中 loss nan

Issue - State: open - Opened by liboaccn about 1 month ago - 1 comment

#940 - 是否支持微调Flux.1 dev

Issue - State: open - Opened by TongrongHuang about 1 month ago

#939 - support for training lamma 3.2 - vision

Issue - State: open - Opened by JAVerma about 1 month ago

#937 - docker利用xtuner微调时，出错，不知道哪的问题？

Issue - State: open - Opened by 159357hou about 1 month ago - 2 comments

#936 - 请问目前支持qwen2吗？

Issue - State: open - Opened by Zheng-Jay about 2 months ago - 3 comments

#934 - 选择四卡训练卡住

Issue - State: open - Opened by AlittlePIE about 2 months ago - 1 comment

#933 - intern2.5-20B微调 后词表长度不一致

Issue - State: open - Opened by topology1 about 2 months ago

#932 - 使用lengthgroupedsampler代替原本的sampler后卡死

Issue - State: open - Opened by xcy9614 about 2 months ago

#931 - [Fix] Fix OOM when qlora converting

Pull Request - State: open - Opened by fanqiNO1 about 2 months ago

#930 - [Bugs] fix qlora convert bugs

Pull Request - State: closed - Opened by HIT-cwh about 2 months ago

#929 - 如何进行val和test？

Issue - State: open - Opened by Diyigelieren about 2 months ago

#928 - version `GLIBCXX_3.4.29' not found

Issue - State: open - Opened by amannier about 2 months ago

#927 - Failed to inference single image using xtuner chat with llava-llama3-8b model

Issue - State: closed - Opened by J0eky about 2 months ago - 1 comment

#926 - 奖励模型问题

Issue - State: open - Opened by Eren139 about 2 months ago - 1 comment

#925 - transformers == 4.44.2 xtuner == 0.1.23 训练 qwen2 时报错

Issue - State: open - Opened by thomZ1 2 months ago - 2 comments

#920 - QwenVL支持

Issue - State: open - Opened by liyan1997 2 months ago

#918 - 一些关于步数统计的疑问

Issue - State: open - Opened by young-chao 2 months ago

#917 - add rescale sp loss

Pull Request - State: open - Opened by HIT-cwh 2 months ago

#916 - reward model训练完如何预测？

Issue - State: open - Opened by tcxia 2 months ago - 1 comment

#915 - qlora微调的模型是不支持中断后继续训练吗？

Issue - State: open - Opened by deep-practice 2 months ago - 3 comments

#914 - sharegpt4v数据集map错误

Issue - State: closed - Opened by bjzhb666 2 months ago - 1 comment

#911 - 如何修改 master port

Issue - State: open - Opened by AislantVentus 3 months ago - 1 comment

#910 - train time decrease from 13 hours to 9

Issue - State: open - Opened by mylesgoose 3 months ago

#908 - 微调数据集策略（dataset make confuse）

Issue - State: open - Opened by EasonQYS 3 months ago

#906 - [Feature] Support balanced dataset to speed-up VL training

Pull Request - State: open - Opened by yqyao 3 months ago - 2 comments

#903 - fine-tuning codegeex4

Issue - State: open - Opened by sgjohnson1981 3 months ago

#901 - Load failure with the converted finetune InternVL2-2B model

Issue - State: closed - Opened by leagend 3 months ago - 1 comment

#898 - 如何实现多张卡共同存放单个模型

Issue - State: open - Opened by RyanOvO 3 months ago - 1 comment

#897 - 之前理解错误，麻烦删除

Issue - State: closed - Opened by Hellcat1005 3 months ago

#896 - Packer 好像没有没有分块 attention_mask

Issue - State: closed - Opened by WallE-Chang 3 months ago

#895 - there is no script for gpt fintune

Issue - State: open - Opened by zhenghuawang6 3 months ago

#894 - 长对话的微调训练

Issue - State: open - Opened by RyanOvO 3 months ago

#893 - 怎么自己指定fp16 fp32 bp16？

Issue - State: closed - Opened by bjzhb666 3 months ago - 1 comment

#892 - accumulative_counts起作用了吗？

Issue - State: closed - Opened by YixinSong-e 3 months ago - 3 comments

#891 - 多机训练的数据集问题

Issue - State: closed - Opened by YixinSong-e 3 months ago

#889 - Adjust the order of InternVL dataset log printing

Pull Request - State: open - Opened by KooSung 3 months ago - 1 comment

#888 - fix

Pull Request - State: open - Opened by ArtificialZeng 3 months ago

#887 - please fix llava 70B config

Issue - State: open - Opened by ds-kczerski 3 months ago

#886 - 无法识别修改后的配置文件

Issue - State: open - Opened by happyrenxiaozhao 3 months ago

#885 - [Bug] fix openai_map_fn bugs

Pull Request - State: closed - Opened by HIT-cwh 3 months ago

#884 - [Fix] Fix output_with_loss in openai_map_fn

Pull Request - State: closed - Opened by fanqiNO1 3 months ago

#883 - 训练时卡死在"Checkpoints will be saved to...”

Issue - State: open - Opened by No360201 3 months ago - 1 comment

#882 - qwen1.5-32b config 支持

Issue - State: open - Opened by WallE-Chang 3 months ago

#880 - CheckpointHook开启by_epoch时未保存模型

Issue - State: open - Opened by SingL3 3 months ago - 2 comments

#878 - support transformers >= 4.43

Pull Request - State: closed - Opened by HIT-cwh 3 months ago

#877 - [Feature] Pipeline Parallelization of Different Stages in RLHF

Pull Request - State: open - Opened by llkn-2 3 months ago - 4 comments

#876 - transformers 4.42.4以后已经没有_flash_attention_forward方法

Issue - State: open - Opened by Snowdar 3 months ago - 2 comments

#875 - 有效节省显存的原因

Issue - State: open - Opened by 2020zyc 3 months ago

#873 - [Bug] fix dsv2 attn dispatch (softmax_scale)

Pull Request - State: closed - Opened by HIT-cwh 3 months ago

#872 - Add internlm2 5 cfgs

Pull Request - State: closed - Opened by HIT-cwh 3 months ago

#871 - xtuner可以支持v100上全量训练7b模型吗？

Issue - State: open - Opened by zhou888888 3 months ago - 2 comments

#870 - Json Error in vis_data

Issue - State: open - Opened by kleinzcy 3 months ago

#869 - readme中增加了MiniCPM的支持

Pull Request - State: closed - Opened by LDLINGLINGLING 3 months ago

#867 - llama3.1 support

Issue - State: open - Opened by yinjun622 4 months ago - 3 comments

#865 - custom system/prompt template support?

Issue - State: open - Opened by HeegonJin 4 months ago - 3 comments

#862 - bump version to 0.1.23

Pull Request - State: closed - Opened by HIT-cwh 4 months ago

#857 - 关于dpo训练时chat template的使用问题

Issue - State: open - Opened by pokerc 4 months ago - 3 comments

#839 - Qwen2-72B，16K长文本，convert转换为HF模型OOM

Issue - State: open - Opened by daiyafei2013 4 months ago - 2 comments

#837 - [Feature] Add LLaST(WIP)

Pull Request - State: open - Opened by ChenX17 4 months ago

#836 - 多机多卡 如何启动训练

Issue - State: open - Opened by poisonwine 4 months ago - 3 comments

#815 - 关于训练的一些疑惑

Issue - State: open - Opened by Zheng-Jay 4 months ago - 3 comments