File size: 1,574 Bytes
3a0fb2c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
---
license: apache-2.0
tags:
- finetuned
- quantized
- 4-bit
- gptq
- transformers
- safetensors
- llama
- text-generation
- base_model:moreh/MoMo-72B-lora-1.8.7-DPO
- license:other
- autotrain_compatible
- endpoints_compatible
- text-generation-inference
- region:us
- has_space
model_name: Smaug-72B-v0.1-GPTQ
base_model: abacusai/Smaug-72B-v0.1
inference: false
model_creator: abacusai
pipeline_tag: text-generation
quantized_by: MaziyarPanahi
---
# Description
[MaziyarPanahi/Smaug-72B-v0.1-GPTQ](https://huggingface.co/MaziyarPanahi/Smaug-72B-v0.1-GPTQ) is a quantized (GPTQ) version of [abacusai/Smaug-72B-v0.1](https://huggingface.co/abacusai/Smaug-72B-v0.1)

## How to use
### Install the necessary packages

```
pip install --upgrade accelerate auto-gptq transformers
```

### Example Python code


```python
from transformers import AutoTokenizer, pipeline
from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
import torch

model_id = "MaziyarPanahi/Smaug-72B-v0.1-GPTQ"

quantize_config = BaseQuantizeConfig(
        bits=4,
        group_size=128,
        desc_act=False
    )

model = AutoGPTQForCausalLM.from_quantized(
        model_id,
        use_safetensors=True,
        device="cuda:0",
        quantize_config=quantize_config)

tokenizer = AutoTokenizer.from_pretrained(model_id)

pipe = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    max_new_tokens=512,
    temperature=0.7,
    top_p=0.95,
    repetition_penalty=1.1
)

outputs = pipe("What is a large language model?")
print(outputs[0]["generated_text"])
```