Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
inference: false
|
3 |
+
license: other
|
4 |
+
tags:
|
5 |
+
- llama-2
|
6 |
+
- llama2
|
7 |
+
- gptq
|
8 |
+
- auto-gptq
|
9 |
+
- 7b
|
10 |
+
- llama
|
11 |
+
- 4bit
|
12 |
+
---
|
13 |
+
|
14 |
+
# Get Started
|
15 |
+
This model should use [AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ) so you need to use `auto-gptq`
|
16 |
+
|
17 |
+
```py
|
18 |
+
from transformers import AutoTokenizer, pipeline, LlamaForCausalLM, LlamaTokenizer
|
19 |
+
from auto_gptq import AutoGPTQForCausalLM
|
20 |
+
|
21 |
+
model_id = 'seonglae/llama-2-7b-chat-hf-gptq'
|
22 |
+
tokenizer = AutoTokenizer.from_pretrained(model_id, use_fast=True)
|
23 |
+
model = AutoGPTQForCausalLM.from_quantized(
|
24 |
+
model_id,
|
25 |
+
model_basename=model_basename,
|
26 |
+
trust_remote_code=True,
|
27 |
+
device='cuda:0',
|
28 |
+
use_triton=False,
|
29 |
+
use_safetensors=True,
|
30 |
+
)
|
31 |
+
|
32 |
+
pipe = pipeline(
|
33 |
+
"text-generation",
|
34 |
+
model=model,
|
35 |
+
tokenizer=tokenizer,
|
36 |
+
temperature=0.5,
|
37 |
+
top_p=0.95,
|
38 |
+
max_new_tokens=100,
|
39 |
+
repetition_penalty=1.15,
|
40 |
+
)
|
41 |
+
prompt = "USER: Are you AI?\nASSISTANT:"
|
42 |
+
pipe(prompt)
|
43 |
+
```
|