p71924506
/
CPM-9G-8B
forked from p83651209/CPM-9G-8B
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Update README.md

This commit is contained in:
p04896573 2024-08-19 11:08:17 +08:00
parent e4da44fcdc
commit 97539a1f44
1 changed files with 4 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
README.md
View File

@ -10,6 +10,9 @@
- 由于新架构中多数据集验证发现2B模型进行lora训练效果不及全参数微调 - 由于新架构中多数据集验证发现2B模型进行lora训练效果不及全参数微调
- 2B模型采用全参数微调训练，我们在[QUICK START](https://www.osredm.com/jiuyuan/CPM-9G-8B/tree/FM_9G/quick_start_clean/readmes/quick_start.md) 中更新了更多关于微调训练的信息 - 2B模型采用全参数微调训练，我们在[QUICK START](https://www.osredm.com/jiuyuan/CPM-9G-8B/tree/FM_9G/quick_start_clean/readmes/quick_start.md) 中更新了更多关于微调训练的信息
- 8B模型LORA微调在master分支进行训练 - 8B模型LORA微调在master分支进行训练
## 版本更新内容 ## 版本更新内容
具体的迭代信息如下： 具体的迭代信息如下：
- 训练：升级了训练代码，提升GPU利用率和并行化，并且2B模型能兼容transformers中的tokenizer(LlamaTokenizerFast) - 训练：升级了训练代码，提升GPU利用率和并行化，并且2B模型能兼容transformers中的tokenizer(LlamaTokenizerFast)