Update README.md
Browse files
README.md
CHANGED
@@ -3,24 +3,24 @@
|
|
3 |
license: openrail
|
4 |
|
5 |
datasets:
|
6 |
-
- LinkSoul/
|
7 |
language:
|
8 |
- zh
|
9 |
- en
|
10 |
---
|
11 |
|
12 |
|
13 |
-
#
|
14 |
|
15 |
-
开源,可商用的**中英文双语语音-语言助手
|
16 |
|
17 |
<!--
|
18 |
<div align="center">
|
19 |
-
<img src="https://huggingface.co/LinkSoul/
|
20 |
</div>
|
21 |
|
22 |
-->
|
23 |
-
![
|
24 |
|
25 |
## 基础演示
|
26 |
|
@@ -29,29 +29,29 @@ language:
|
|
29 |
## 在线试玩
|
30 |
|
31 |
> Talk is cheap, Show you the Demo.
|
32 |
-
- [Demo 地址 / HuggingFace Spaces](https://huggingface.co/spaces/LinkSoul/
|
33 |
|
34 |
## 资源下载
|
35 |
|
36 |
- 模型:
|
37 |
-
- [
|
38 |
-
- [
|
39 |
|
40 |
- 语言模型:
|
41 |
- [Chinese-Llama-2-7b](https://github.com/LinkSoul-AI/Chinese-Llama-2-7b)
|
42 |
- [Baichuan-7B](https://huggingface.co/baichuan-inc/Baichuan-7B)
|
43 |
|
44 |
-
- 数据集:[
|
45 |
|
46 |
## 环境安装
|
47 |
```shell
|
48 |
# clone the repository
|
49 |
-
git clone https://github.com/LinkSoul-AI/
|
50 |
-
cd
|
51 |
|
52 |
# install package
|
53 |
-
conda create -n
|
54 |
-
conda activate
|
55 |
pip install --upgrade pip
|
56 |
pip install -e .
|
57 |
```
|
@@ -59,11 +59,11 @@ pip install -e .
|
|
59 |
## 快速测试
|
60 |
|
61 |
```shell
|
62 |
-
export
|
63 |
python infer.py \
|
64 |
--input_audio_file PATH/TO/YOUR/AUDIO \
|
65 |
-
--
|
66 |
-
--
|
67 |
--llm_type "Chinese_llama2" or "baichuan" \
|
68 |
```
|
69 |
|
@@ -80,7 +80,7 @@ python infer.py \
|
|
80 |
|
81 |
## 项目协议
|
82 |
|
83 |
-
[Apache-2.0 license](https://github.com/LinkSoul-AI/
|
84 |
|
85 |
## 微信交流群
|
86 |
<!--
|
|
|
3 |
license: openrail
|
4 |
|
5 |
datasets:
|
6 |
+
- LinkSoul/LLaSM-Audio-Instructions
|
7 |
language:
|
8 |
- zh
|
9 |
- en
|
10 |
---
|
11 |
|
12 |
|
13 |
+
# LLaSM: Large Language and Speech Model
|
14 |
|
15 |
+
开源,可商用的**中英文双语语音-语言助手 LLaSM 以及中英文语音 SFT 数据集 LLaSM-Audio-Instructions**,第一个支持中英文语音-文本多模态对话的开源可商用对话模型。
|
16 |
|
17 |
<!--
|
18 |
<div align="center">
|
19 |
+
<img src="https://huggingface.co/LinkSoul/LLaSM-Cllama2/blob/main/meta/preview.jpg" width="40%">
|
20 |
</div>
|
21 |
|
22 |
-->
|
23 |
+
![LLaSM](meta/llasm_preview.jpg)
|
24 |
|
25 |
## 基础演示
|
26 |
|
|
|
29 |
## 在线试玩
|
30 |
|
31 |
> Talk is cheap, Show you the Demo.
|
32 |
+
- [Demo 地址 / HuggingFace Spaces](https://huggingface.co/spaces/LinkSoul/LLaSM)
|
33 |
|
34 |
## 资源下载
|
35 |
|
36 |
- 模型:
|
37 |
+
- [LLaSM-Chinese-Llama-2-7B](https://huggingface.co/LinkSoul/LLaSM-Cllama2)
|
38 |
+
- [LLaSM-Baichuan-7B](https://huggingface.co/LinkSoul/LLaSM-Baichuan)
|
39 |
|
40 |
- 语言模型:
|
41 |
- [Chinese-Llama-2-7b](https://github.com/LinkSoul-AI/Chinese-Llama-2-7b)
|
42 |
- [Baichuan-7B](https://huggingface.co/baichuan-inc/Baichuan-7B)
|
43 |
|
44 |
+
- 数据集:[LLaSM-Audio-Instructions](https://huggingface.co/datasets/LinkSoul/LLaSM-Audio-Instructions)
|
45 |
|
46 |
## 环境安装
|
47 |
```shell
|
48 |
# clone the repository
|
49 |
+
git clone https://github.com/LinkSoul-AI/LLaSM
|
50 |
+
cd LLaSM
|
51 |
|
52 |
# install package
|
53 |
+
conda create -n llasm python=3.10 -y
|
54 |
+
conda activate llasm
|
55 |
pip install --upgrade pip
|
56 |
pip install -e .
|
57 |
```
|
|
|
59 |
## 快速测试
|
60 |
|
61 |
```shell
|
62 |
+
export LLASM_DEVICE="cuda:0"
|
63 |
python infer.py \
|
64 |
--input_audio_file PATH/TO/YOUR/AUDIO \
|
65 |
+
--llasm_model PATH/TO/LLaSM/MODEL \
|
66 |
+
--llasm_audio_tower PATH/TO/WHISPER/MODEL \
|
67 |
--llm_type "Chinese_llama2" or "baichuan" \
|
68 |
```
|
69 |
|
|
|
80 |
|
81 |
## 项目协议
|
82 |
|
83 |
+
[Apache-2.0 license](https://github.com/LinkSoul-AI/LLaSM/blob/main/LICENSE)
|
84 |
|
85 |
## 微信交流群
|
86 |
<!--
|