LinkSoul-AI/Chinese-Llama-2-7b

Issues

• 华视回放

  #61 opened a month ago by full55555
  1

• chinese-llama-2-7b.ipynb出错求助

  #52 opened 10 months ago by leoring123
  1

• 请问支持微调吗

  #60 opened 6 months ago by hzlirz
  0

• 有人遇到过这个报错吗，我的flash_attn版本是2.3.0，跪求大佬指点。ImportError: cannot import name 'flash_attn_unpadded_qkvpacked_func' from 'flash_attn.flash_attn_interface'

  #58 opened 9 months ago by SoseloX
  1

• requirements.txt

  #59 opened 7 months ago by jiangbinbin-code
  0

• 使用项目中的api.py启动模型，没做4bit量化，使用v100的卡，推理特别慢，经常卡住

  #51 opened a year ago by jcxian
  2

• 4张p40的显卡，装载模型报错CUDA out of memory，但是每张卡24G呀，这个该怎么办

  #56 opened 9 months ago by GzRichChen
  1

• 训练的数据集格式是啥

  #53 opened 10 months ago by 631068264
  2

• ileNotFoundError: Directory LinkSoul/instruction_merge_set is neither a `Dataset` directory nor a `DatasetDict` directory.

  #28 opened a year ago by zyxcambridge
  2

• datatsetsbug

  #30 opened a year ago by zyxcambridge
  1

• 训练时的prompt是什么格式的呢

  #57 opened 9 months ago by zozoteacher
  0

• 继续微调

  #55 opened 10 months ago by hallucination11
  0

• 你好，请问怎么实现公网的api呢

  #54 opened 10 months ago by 15899885850
  0

• 4bit量化后的模型为什么这么大

  #31 opened a year ago by liuyukid
  2

• 制作镜像卡住

  #50 opened a year ago by MUFCLK
  0

• 快速测试代码段OOM

  #36 opened a year ago by LiuZhihhxx
  2

• 请教一下用open-api-server跟inference-hf 回答问题不一致的问题

  #49 opened a year ago by ibmxiang
  0

• 为什么7b模型有26G，原始llama2 7b模型只有13G

  #44 opened a year ago by secain
  0

• 怎么从本地直接加载模型啊

  #46 opened a year ago by JACKzhuz
  1

• 请教一下公开模型的训练细节

  #47 opened a year ago by shichao-wang
  0

• when I try to train the model, Can't load the configuration of '/PATH/TO/TRANSFORMERS/VERSION/LLAMA2'

  #43 opened a year ago by githubtianya
  0

• 调lora需要多少资源

  #42 opened a year ago by liaoweiguo
  0

• 有更大的模型吗，33B，70B

  #41 opened a year ago by liaoweiguo
  0

• 这个是不是还没进行词表扩充

  #19 opened a year ago by longgui0318
  0

• huggingface上模型是openrail协议，项目上是apache2，不一致在哪

  #39 opened a year ago by wzg-zhuo
  0

• 有流的返回方式吗？

  #34 opened a year ago by tianlichunhong
  2

• 能提供測試分數嗎

  #35 opened a year ago by CRGBS
  0

• 能否做个API接口程序呢？非常感谢！

  #8 opened a year ago by stuarthe
  5

• OSError: We couldn't connect to 'https://huggingface.co' to load this file, couldn't find it in the cached files and it looks like LinkSoul/Chinese-Llama-2-7b is not the path to a directory containing a file named config.json. Checkout your internet connection or see how to run the library in offline mode at 'https://huggingface.co/docs/transformers/installation#offline-mode'.

  #33 opened a year ago by hxujal
  0

• could not select device driver "" with capabilities: [[gpu]]

  #32 opened a year ago by hxujal
  0

• 数据集中中文样本数量

  #29 opened a year ago by pangwenfeng
  0

• 卡住了

  #27 opened a year ago by zyxcambridge
  0

• 训练资源

  #18 opened a year ago by DavidYanAnDe
  1

• 加不了群

  #20 opened a year ago by inkoku
  1

• Traditional Chinese

  #12 opened a year ago by compustar
  1

• 您好，关于 训练和推理，有建议的服务器配置吗？

  #2 opened a year ago by willqq
  6

• 请问flash attention 用的是v1吗，是什么版本啊

  #16 opened a year ago by lllyyyqqq
  0

• ggmlv

  #7 opened a year ago by 13331112522
  4

• 什么时候开源训练代码？

  #1 opened a year ago by Young2019
  1