-
Google 刚公布了 Gemma 开源模型,请问有计划是否可以添加到 Xtuner,本人想为 xtuner 贡献 `Gemma` 相关配置和实现
huggingface: https://huggingface.co/google/gemma-7b
technical report: https://blog.google/technology/developers/gemma-open…
-
### System Info
(Possible duplicate: #10256)
I have written a custom tokenizer that builds on top of `BertTokenizer` (returns one extra list of ids that will later be embedded in a custom model). …
-
### 描述该错误
```
self = , model_name = 'internlm/internlm2-chat-7b-sft'
@pytest.mark.parametrize("model_name", [
"internlm/internlm2-chat-7b",
"internlm/internlm2-chat-7b-sft",…
-
### Checklist
- [X] 1. I have searched related issues but cannot get the expected help.
- [X] 2. The bug has not been fixed in the latest version.
### Describe the bug
lmdeploy 0.2.2在 internlm2-cha…
-
**功能描述 **
梳理文章的大纲
HHH16 updated
4 months ago
-
https://github.com/Dao-AILab/flash-attention
Flash attention v2 was released claiming 2x speedups. Making an issue to remind myself to have a look at it. And also if anyone else wants to try implem…
-
### Prerequisite
- [X] I have searched [Issues](https://github.com/open-compass/opencompass/issues/) and [Discussions](https://github.com/open-compass/opencompass/discussions) but cannot get the expe…
-
![image](https://github.com/InternLM/xtuner/assets/43675899/ec7c4671-6603-4223-97fa-2d4c7cd33de6)
llava interlm2chat + qlora 推理mmb数据集报错tensor不在一个设备上,这个挺奇怪的我看代码没有to(device)的操作全是cuda()咋还会报错呢
-
现在无论怎么改py样例文件,微调的时候只能用到1个gpu
如:
把 internlm2_chat_7b_qlora_oasst1_e3_copy.py 的 batch_size 为1的时候,执行:xtuner train ./internlm2_chat_7b_qlora_oasst1_e3_copy.py --deepspeed deepspeed_zero2
![image](http…
-
### Checklist
- [X] 1. I have searched related issues but cannot get the expected help.
- [X] 2. The bug has not been fixed in the latest version.
### Describe the bug
![6b857676a91e886304f529295b8…