File size: 4,591 Bytes
3779a9d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
cd020fe
3779a9d
 
 
 
 
 
8ba19c3
3779a9d
8ba19c3
3779a9d
 
 
 
 
 
06d3986
3779a9d
 
1c78c94
3779a9d
 
 
6bfceae
3779a9d
 
 
 
 
24a5819
 
 
 
209b46a
3779a9d
 
 
 
 
 
 
 
6bfceae
3779a9d
 
 
 
 
 
5fd8819
3779a9d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
209b46a
b6ba501
3779a9d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
---
license: cc-by-sa-3.0
tags:
- MosaicML
- AWQ
inference: false
---

# MPT-30B-Instruct (4-bit 128g AWQ Quantized)
[MPT-30B-Instruct](https://huggingface.co/mosaicml/mpt-30b-chat) is a model for short-form instruction following. 

This model is a 4-bit 128 group size AWQ quantized model. For more information about AWQ quantization, please click [here](https://github.com/mit-han-lab/llm-awq).

## Model Date

July 5, 2023

## Model License

Please refer to original MPT model license ([link](https://huggingface.co/mosaicml/mpt-30b-chat)).

Please refer to the AWQ quantization license ([link](https://github.com/llm-awq/blob/main/LICENSE)).

## CUDA Version

This model was successfully tested on CUDA driver v530.30.02 and runtime v11.7 with Python v3.10.11. Please note that AWQ requires NVIDIA GPUs with compute capability of `8.0` or higher.

For Docker users, the `nvcr.io/nvidia/pytorch:23.06-py3` image is runtime v12.1 but otherwise the same as the configuration above and has also been verified to work.

## How to Use

```bash
git clone https://github.com/mit-han-lab/llm-awq \
&& cd llm-awq \
&& git checkout ce4a6bb1c238c014a06672cb74f6865573494d66 \
&& pip install -e . \
&& cd awq/kernels \
&& python setup.py install
```

```python
import time
import torch
from awq.quantize.quantizer import real_quantize_model_weight
from transformers import AutoModelForCausalLM, AutoConfig, AutoTokenizer, TextStreamer
from accelerate import init_empty_weights, load_checkpoint_and_dispatch
from huggingface_hub import snapshot_download

model_name = "abhinavkulkarni/mosaicml-mpt-30b-chat-w4-g128-awq"

# Config
config = AutoConfig.from_pretrained(model_name, trust_remote_code=True)

# Tokenizer
try:
    tokenizer = AutoTokenizer.from_pretrained(config.tokenizer_name, trust_remote_code=True)
except:
    tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=False, trust_remote_code=True)
streamer = TextStreamer(tokenizer, skip_special_tokens=True)

# Model
w_bit = 4
q_config = {
    "zero_point": True,
    "q_group_size": 128,
}

load_quant = snapshot_download(model_name)

with init_empty_weights():
    model = AutoModelForCausalLM.from_config(config=config, 
                                                 torch_dtype=torch.float16, trust_remote_code=True)

real_quantize_model_weight(model, w_bit=w_bit, q_config=q_config, init_only=True)
model.tie_weights()

model = load_checkpoint_and_dispatch(model, load_quant, device_map="balanced")

# Inference
prompt = f'''What is the difference between nuclear fusion and fission?
###Response:'''

input_ids = tokenizer(prompt, return_tensors='pt').input_ids.cuda()
output = model.generate(
    inputs=input_ids, 
    temperature=0.7,
    max_new_tokens=512,
    top_p=0.15,
    top_k=0,
    repetition_penalty=1.1,
    eos_token_id=tokenizer.eos_token_id,
    streamer=streamer)
```

## Evaluation

This evaluation was done using [LM-Eval](https://github.com/EleutherAI/lm-evaluation-harness).

[MPT-30B-Instruct](https://huggingface.co/mosaicml/mpt-30b-chat)

|  Task  |Version|    Metric     | Value |   |Stderr|
|--------|------:|---------------|------:|---|------|
|wikitext|      1|word_perplexity|11.5609|   |      |
|        |       |byte_perplexity| 1.5805|   |      |
|        |       |bits_per_byte  | 0.6603|   |      |

[MPT-30B-Instruct (4-bit 128-group AWQ)](https://huggingface.co/abhinavkulkarni/mosaicml-mpt-30b-chat-w4-g128-awq)

|  Task  |Version|    Metric     | Value |   |Stderr|
|--------|------:|---------------|------:|---|------|
|wikitext|      1|word_perplexity|11.6649|   |      |
|        |       |byte_perplexity| 1.5831|   |      |
|        |       |bits_per_byte  | 0.6628|   |      |

## Acknowledgements

The MPT model was originally finetuned by Sam Havens and the MosaicML NLP team. Please cite this model using the following format:

```
@online{MosaicML2023Introducing,
    author    = {MosaicML NLP Team},
    title     = {Introducing MPT-30B: A New Standard for Open-Source, Commercially Usable LLMs},
    year      = {2023},
    url       = {www.mosaicml.com/blog/mpt-30b},
    note      = {Accessed: 2023-03-28}, % change this date
    urldate   = {2023-03-28} % change this date
}
```


The model was quantized with AWQ technique. If you find AWQ useful or relevant to your research, please kindly cite the paper:

```
@article{lin2023awq,
  title={AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration},
  author={Lin, Ji and Tang, Jiaming and Tang, Haotian and Yang, Shang and Dang, Xingyu and Han, Song},
  journal={arXiv},
  year={2023}
}
```