Update README.md
Browse files
README.md
CHANGED
|
@@ -8,6 +8,7 @@ tags:
|
|
| 8 |
- conversational
|
| 9 |
- generation
|
| 10 |
- qlora
|
|
|
|
| 11 |
---
|
| 12 |
# Anima
|
| 13 |
|
|
@@ -46,16 +47,30 @@ Anima模型基于QLoRA开源的[33B guanaco](https://huggingface.co/timdettmers/
|
|
| 46 |
|
| 47 |
#### 如何训练
|
| 48 |
|
| 49 |
-
|
| 50 |
-
|
| 51 |
-
```
|
| 52 |
-
# 1. install dependencies
|
| 53 |
-
pip install -r requirements.txt
|
| 54 |
-
# 2.
|
| 55 |
-
cd training
|
| 56 |
-
./run_Amina_training.sh
|
| 57 |
-
```
|
| 58 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
|
| 60 |
|
| 61 |
## 📊验证评估
|
|
|
|
| 8 |
- conversational
|
| 9 |
- generation
|
| 10 |
- qlora
|
| 11 |
+
pipeline_tag: conversational
|
| 12 |
---
|
| 13 |
# Anima
|
| 14 |
|
|
|
|
| 47 |
|
| 48 |
#### 如何训练
|
| 49 |
|
| 50 |
+
1. 重现Anima的模型训练过程:使用以下步骤可以重现Anima 33B模型(单卡80GB H100或双卡 40GB A100均测试过可运行):
|
| 51 |
+
|
| 52 |
+
```bash
|
| 53 |
+
# 1. install dependencies
|
| 54 |
+
pip install -r requirements.txt
|
| 55 |
+
# 2.
|
| 56 |
+
cd training
|
| 57 |
+
./run_Amina_training.sh
|
| 58 |
+
```
|
| 59 |
+
|
| 60 |
+
2. 基于Anima finetune训练其他model:
|
| 61 |
+
|
| 62 |
+
```bash
|
| 63 |
+
# 1. install dependencies
|
| 64 |
+
pip install -r requirements.txt
|
| 65 |
+
# 2.
|
| 66 |
+
cd training
|
| 67 |
+
./run_finetune_raining_based_on_Anima.sh
|
| 68 |
+
```
|
| 69 |
+
注:可以修改run_finetune_raining_based_on_Anima.sh中的--dataset和--dataset_format参数使用其他训练数据dataset。
|
| 70 |
+
|
| 71 |
+
#### 多卡训练
|
| 72 |
+
由于使用Hugging Face Accelerate,天然支持多卡训练。
|
| 73 |
+
我们测试过双卡40GB的A100,可以直接运行。
|
| 74 |
|
| 75 |
|
| 76 |
## 📊验证评估
|