Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,88 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: llama2
|
3 |
+
language:
|
4 |
+
- ko
|
5 |
+
pipeline_tag: text-generation
|
6 |
+
tags:
|
7 |
+
- ' llama'
|
8 |
+
- facebook
|
9 |
+
- ' meta'
|
10 |
+
- llama-2
|
11 |
+
- kollama
|
12 |
+
- llama-2-ko
|
13 |
+
- llama-2-ko-chat
|
14 |
+
- text-generation-inference
|
15 |
+
---
|
16 |
+
|
17 |
+
# 💻MAC os Compatible💻
|
18 |
+
|
19 |
+
# Llama 2 ko 7B - GGUF
|
20 |
+
- Model creator: [Meta](https://huggingface.co/meta-llama)
|
21 |
+
- Original model: [Llama 2 7B Chat](https://huggingface.co/meta-llama/Llama-2-7b-chat)
|
22 |
+
- Original Llama-2-Ko-Chat model: [Llama 2 ko 7B Chat](https://huggingface.co/kfkas/Llama-2-ko-7b-Chat)
|
23 |
+
- Reference: [Llama 2 7B GGUF](https://huggingface.co/TheBloke/Llama-2-7B-GGUF)
|
24 |
+
|
25 |
+
<!-- description start -->
|
26 |
+
## Download
|
27 |
+
```shell
|
28 |
+
pip3 install huggingface-hub>=0.17.1
|
29 |
+
```
|
30 |
+
|
31 |
+
Then you can download any individual model file to the current directory, at high speed, with a command like this:
|
32 |
+
|
33 |
+
```shell
|
34 |
+
huggingface-cli download 24bean/Llama-2-ko-7B-Chat-GGUF llama-2-ko-7b-chat-q8-0.gguf --local-dir . --local-dir-use-symlinks False
|
35 |
+
```
|
36 |
+
|
37 |
+
Or you can download llama-2-ko-7b.gguf, non-quantized model by
|
38 |
+
|
39 |
+
```shell
|
40 |
+
huggingface-cli download 24bean/Llama-2-ko-7B-Chat-GGUF llama-2-ko-7b-chat.gguf --local-dir . --local-dir-use-symlinks False
|
41 |
+
```
|
42 |
+
|
43 |
+
## Example `llama.cpp` command
|
44 |
+
|
45 |
+
Make sure you are using `llama.cpp` from commit [d0cee0d36d5be95a0d9088b674dbb27354107221](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
|
46 |
+
|
47 |
+
```shell
|
48 |
+
./main -ngl 32 -m llama-2-ko-7b-chat-q8-0.gguf --color -c 4096 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "{prompt}"
|
49 |
+
```
|
50 |
+
|
51 |
+
# How to run from Python code
|
52 |
+
|
53 |
+
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries.
|
54 |
+
|
55 |
+
## How to load this model from Python using ctransformers
|
56 |
+
|
57 |
+
### First install the package
|
58 |
+
|
59 |
+
```bash
|
60 |
+
# Base ctransformers with no GPU acceleration
|
61 |
+
pip install ctransformers>=0.2.24
|
62 |
+
# Or with CUDA GPU acceleration
|
63 |
+
pip install ctransformers[cuda]>=0.2.24
|
64 |
+
# Or with ROCm GPU acceleration
|
65 |
+
CT_HIPBLAS=1 pip install ctransformers>=0.2.24 --no-binary ctransformers
|
66 |
+
# Or with Metal GPU acceleration for macOS systems
|
67 |
+
CT_METAL=1 pip install ctransformers>=0.2.24 --no-binary ctransformers
|
68 |
+
```
|
69 |
+
|
70 |
+
### Simple example code to load one of these GGUF models
|
71 |
+
|
72 |
+
```python
|
73 |
+
from ctransformers import AutoModelForCausalLM
|
74 |
+
|
75 |
+
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
|
76 |
+
llm = AutoModelForCausalLM.from_pretrained("24bean/Llama-2-ko-7B-Chat-GGUF", model_file="llama-2-7b-chat-q8-0.gguf", model_type="llama", gpu_layers=50)
|
77 |
+
|
78 |
+
print(llm("인공지능은"))
|
79 |
+
```
|
80 |
+
|
81 |
+
## How to use with LangChain
|
82 |
+
|
83 |
+
Here's guides on using llama-cpp-python or ctransformers with LangChain:
|
84 |
+
|
85 |
+
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
|
86 |
+
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
|
87 |
+
|
88 |
+
<!-- README_GGUF.md-how-to-run end -->
|