OpenMOSS/MOSS

Issues

• streamlit run moss_web_demo_streamlit.py --server.port 8888

  #378 opened a month ago by MR-HOLO
  0

• 请问moss-003-pm-data偏好数据什么时间能开源？

  #368 opened a year ago by liumingzhu6060
  2

• moss-moon-003-sft Jupyter测试失败

  #376 opened 10 months ago by YoloZyk
  2

• 这是小时候的550W吗

  #364 opened a year ago by mason334
  1

• moss为什么不更新了

  #377 opened 8 months ago by endNone
  0

• 想问下moss的多轮sft数据是怎么制作出来的

  #373 opened a year ago by xealml
  1

• 你好，我想知道moss能否采用别的gpu调试

  #372 opened a year ago by user10246
  1

• 多轮对话数据构造的时候是否会有上下文不一致的问题

  #375 opened 10 months ago by muziyongshixin
  0

• 你好，我有一个很有意思的想法。能不能提供帮助

  #374 opened 10 months ago by user10246
  3

• 求一個moss邀請碼

  #337 opened a year ago by Hansonwong47
  3

• MOSS next release?

  #371 opened a year ago by yhyu13
  3

• do_sample=True

  #370 opened a year ago by zhaoyang-coder
  0

• RuntimeError: indices should be either on cpu or on the same device as the indexed tensor (cuda:1)

  #363 opened a year ago by lizhixi212
  1

• how to run fnlp/moss-base-7b in multi gpus

  #369 opened a year ago by FakerYFX
  0

• export CUDA_LAUNCH_BLOCKING=1 Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

  #367 opened a year ago by whay181
  0

• how to continue pretrain?

  #366 opened a year ago by h-ccc
  0

• 两张v100部署失败

  #362 opened a year ago by Cocoalate
  2

• langchain+moss

  #365 opened a year ago by secslim
  0

• cuda error

  #353 opened a year ago by wqh17101
  1

• moss-003-sft-plugin-data里面有很多错误

  #361 opened a year ago by guozhiyao
  0

• 微信群聊过期了

  #360 opened a year ago by FunEx06
  0

• 叫moss 为什么科幻片不推荐流浪地球

  #354 opened a year ago by hutairan123
  1

• finetune的时候为何没有把<|Human|>的loss给mask掉？

  #359 opened a year ago by jouw
  0

• finetuning loss

  #344 opened a year ago by ZiboZ
  1

• 求邀请码

  #358 opened a year ago by FunEx06
  0

• 多轮对话数据处理

  #349 opened a year ago by 447428054
  1

• moss-003-sft-plugin-data的30w条数据都是人工写的吗？

  #357 opened a year ago by amenota
  0

• Inner Thoughts, Tool Responses 这些数据怎么训练？

  #356 opened a year ago by wjy3326
  0

• 使用fp8 后微调速度特别慢

  #355 opened a year ago by lhtpluto
  0

• 使用jittor加载模型的时候这个文件pytorch_model.bin.index.json是如何生成的？

  #348 opened a year ago by wanglaiqi
  1

• 如何开发自己的插件呢？开发后如何注册到moss里面呢？🙏

  #329 opened a year ago by yihuaxiang
  3

• 模型显存占用问题

  #352 opened a year ago by wqh17101
  2

• How to train a custom tokenizer for Chinese from scratch

  #351 opened a year ago by SparkJiao
  0

• 请问moss-002-sft-data和moss-003-sft-data是否有重复的数据 ？

  #350 opened a year ago by CoinCheung
  0

• 单机单卡80G显存微调报错torch.distributed.elastic.multiprocessing.api:failed

  #341 opened a year ago by qinkongyouyun
  0

• squad数据集

  #347 opened a year ago by WTJ-github-ing
  0

• cuda

  #346 opened a year ago by altria-zewei-wang
  0

• 下游任务也是中文，使用中文instruction or 英文instruction ?

  #343 opened a year ago by feng-1985
  0

• 4张v100可以完成微调吗？

  #335 opened a year ago by wangyou0804
  2

• readme里面的网页demo有开源吗

  #342 opened a year ago by lucasjinreal
  0

• 微调后导出67g的bin文件，之后怎么做呢

  #327 opened a year ago by mayurou
  17

• [Help] <输入token数目有限制吗>

  #340 opened a year ago by LebronJames0423
  0

• 微调报错warning没有具体信息

  #339 opened a year ago by altria-zewei-wang
  0

• TypeError: '<' not supported between instances of 'tuple' and 'float'

  #338 opened a year ago by DexiuCai
  0

• 求 发布像Web MOSS插件版本

  #336 opened a year ago by lhtpluto
  0

• 官方微信群的二维码过期了，能否更新一下

  #330 opened a year ago by bltcn
  6

• 请问有没有训练过程相关的文档？

  #334 opened a year ago by sunway1988
  1

• CUDA_LAUNCH_BLOCKING=1、TORCH_USE_CUDA_DSA

  #328 opened a year ago by lhtpluto
  4

• 关于RLHF的问题

  #333 opened a year ago by dongdongrj
  0

• .

  #331 opened a year ago by Chenhuaqi6
  0