InternLM/InternLM

Issues

• [Bug] InternLM2 int4 出现重复说话、重复前置内容（system prompt）现象

  #706 opened 3 months ago by sanbuphy
  11

• [QA] 设置do_sample=False(贪婪解码),下的解码问题.

  #738 opened a month ago by fxb392
  0

• [Bug] failing everytime and getting CUDA out of memory

  #724 opened a month ago by HAIXteam
  6

• [Feature] 一个把internLM2 LoRA部分转换成llama风格的脚本

  #737 opened a month ago by ChengYouFancy
  0

• [QA] Question about phase 2 long context pretraining batch size

  #733 opened a month ago by skyshine102
  3

• [Feature] 是否已经支持tensorrt-llm或者计划支持？

  #714 opened 2 months ago by mblank5
  2

• [QA] 请问如何在昇腾910上进行模型微调？

  #736 opened 2 months ago by rourouZ
  3

• [QA] 使用InternLM微调qa机器人，如何让它从训练资料中选择答案，而不是自己生成？

  #728 opened 2 months ago by ILG2021
  2

• [QA] Can internlm2 be supported in fastchat?

  #723 opened 2 months ago by WuLindong1997
  6

• [Bug] safetensors_rust.SafetensorError: Error while deserializing header: HeaderTooLarge

  #731 opened 2 months ago by yuanphoenix
  3

• [Feature] convert2llama.py

  #729 opened 2 months ago by bladeswill
  1

• [Bug] internlm2_chat_1.8b 模型不支持多轮对话

  #726 opened 2 months ago by hello-gary-2022
  2

• [Bug] llama.cpp internlm2 function calling bug

  #696 opened 2 months ago by bdqfork
  6

• [QA] 书生2模型有关chat_template的问题

  #700 opened 3 months ago by DirtyKnightForVi
  5

• [QA] InternLM 2 对文字种类的识别, 生成能力以及微调相关问题

  #704 opened 3 months ago by timousT
  6

• [QA] 请问是否会开源PPO的训练code和reward model？

  #712 opened 3 months ago by LYMDLUT
  1

• 请问是否支持在MindSpore的910A或者910B上部署？

  #698 opened 3 months ago by ChingKwanCheung
  5

• [QA] Number of training tokens for Internlm2 1.8B, 7B, and 20B?

  #711 opened 3 months ago by RicardoDominguez
  3

• 一人血书求讲解怎么来洗数据！

  #718 opened 3 months ago by Jianfeng777
  5

• 万人血书 InternLM2-4B ❗❕❗❕❗

  #719 opened 3 months ago by SaaRaaS-1300
  0

• 请问是否支持200k上文的微调，需要什么样的配置？[QA]

  #702 opened 3 months ago by Labmem009
  2

• [Bug] Special tokens are still mismatched.

  #715 opened 3 months ago by Li-Qingyun
  2

• [Bug] internlm2-chat-20b huggingface下载503报错，模型不存在

  #716 opened 3 months ago by DefTruth
  0

• [QA] Windows11使用bitsandbytes运行InternLM2-chat-7B-4bits量化，大模型精神错乱

  #680 opened 3 months ago by Sawtone
  19

• [Bug] When loading a model by using transformers and using stream chat, it seems no whitespace character in English response.

  #709 opened 3 months ago by zhulinJulia24
  1

• [Bug] 微调eval阶段使用generate的结果会出现</s>

  #701 opened 3 months ago by AZYoung233
  2

• [Bug] `Unrecognized configuration class Error` returned by `AutoTokenizer.from_pretrained` with InternLM-chat-1.8b-sft (transformers==4.36)

  #688 opened 4 months ago by openmmlab-bot
  2

• [QA] 如果想基于internLM使用领域数据微调出一个领域模型，应该使用internLM-20B-chat还是internLM-20B-sft？

  #694 opened 4 months ago by honglianglv
  3

• [QA] 如何提升推理速度？

  #697 opened 4 months ago by luzhongyu
  2

• [QA] 为什么我通过lmdepoly的教程部署后，输出的是一堆乱码呢，具体图片如下，感谢帮助

  #665 opened 4 months ago by sunshicheng1
  6

• [Bug] 运行魔塔示例代码出现 AttributeError: 'ChatStreamer' object has no attribute 'cache'

  #668 opened 4 months ago by tcexeexe
  4

• [Bug] internlm2 emit content with [UNUSED_TOKEN_145] at times

  #654 opened 4 months ago by gaord
  5

• [Bug] AttributeError: module 'tokenizers.decoders' has no attribute 'Replace'

  #655 opened 4 months ago by jiaoyang3
  5

• [Bug] internlm2-chat-20b转换成llama使用vllm后会输出[UNUSED_TOKEN_145]

  #662 opened 4 months ago by dafen12
  4

• 卡住问题和max_position_embeddings[QA]

  #690 opened 4 months ago by xxg98
  2

• [QA] tokenizer对<|im_start|>的特殊编码不对

  #693 opened 4 months ago by shipengai
  0

• InternLM2 下载模型后，用本地路径加载tokenizer和模型报错

  #691 opened 4 months ago by syp1997
  2

• [Bug] 模型推理中止

  #669 opened 4 months ago by Patrick-Ni
  7

• [QA] InternLM2微调时数据的max_token

  #681 opened 4 months ago by MING-ZCH
  0

• internlm2-chat-7b本身支持的token多大？[QA]

  #683 opened 4 months ago by xxg98
  1

• [Bug] LMdeploy跑出来的demo有问题

  #684 opened 4 months ago by duliangang
  5

• 加载internlm2-7b-chat报错[Bug]

  #677 opened 4 months ago by xxg98
  2

• api_server无法添加--session-len参数[Bug]

  #678 opened 4 months ago by xxg98
  2

• [QA] 模型在推理时输出长度有限，是否有相关的参数能控制输出长度

  #675 opened 4 months ago by ChingKwanCheung
  4

• 直接使用NTK支持到200K吗？是否在200k的上下文上做过训练？有没有超长文本的评测报告

  #676 opened 4 months ago by lvjianxin
  1

• [Feature] MOE多模态模型 许愿

  #670 opened 4 months ago by sanbuphy
  1

• [QA] transformers>= 4.34 高版本执行报错CUDA error: device-side assert triggered

  #659 opened 5 months ago by zhulinJulia24
  4

• 长上文的微调之后还会出教程吗？比如100k的指令微调语料大概需要多少显存，训练策略要选择什么？[QA]

  #661 opened 5 months ago by Labmem009
  3

• [Bug] ImportError: InternLM2Converter requires the protobuf library

  #656 opened 5 months ago by kkscilife
  4

• [Bug] internlm2-20B 推理时出现乱码；internlm2-7B能够正常推理。

  #657 opened 5 months ago by zexuanqiu
  4