File size: 5,014 Bytes
32a91b9
 
 
 
 
3dd4a87
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9a45a0e
3dd4a87
 
 
 
 
63af3ac
 
 
 
 
 
 
 
 
 
 
 
 
dfa3560
 
 
63af3ac
 
3dd4a87
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7ae908f
3dd4a87
7ae908f
3dd4a87
76e52ca
3dd4a87
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2b928ed
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
datasets:
- NeelNanda/pile-10k
---


## Model Details: Mistral-7B-v0.1-int4-inc

This model is an int4 model with group_size 128 of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1)  generated by [intel/auto-round](https://github.com/intel/auto-round). 

## How To Use

### Reproduce the model

Here is the sample command to reproduce the model

```bash
git clone https://github.com/intel/auto-round
cd auto-round/examples/language-modeling
pip install -r requirements.txt
python3 main.py \
--model_name  mistralai/Mistral-7B-v0.1 \
--device 0 \
--group_size 128 \
--bits 4 \
--iters 1000 \
--deployment_device 'gpu' \
--output_dir "./tmp_autoround"
```



### Use the model
### INT4 Inference with ITREX on CPU
Install the latest [intel-extension-for-transformers](https://github.com/intel/intel-extension-for-transformers)
```python
from intel_extension_for_transformers.transformers import AutoModelForCausalLM
from transformers import AutoTokenizer
quantized_model_dir = "Intel/Mistral-7B-v0.1-int4-inc"
model = AutoModelForCausalLM.from_pretrained(quantized_model_dir,
                                             device_map="auto",
                                             trust_remote_code=False,
                                             use_neural_speed=False,
                                             )
tokenizer = AutoTokenizer.from_pretrained(quantized_model_dir, use_fast=True)
print(tokenizer.decode(model.generate(**tokenizer("There is a girl who likes adventure,", return_tensors="pt").to(model.device),max_new_tokens=50)[0]))
"""
<s> There is a girl who likes adventure, and she is a little bit crazy. She is a little bit crazy because she likes to do things that are dangerous. She likes to climb mountains, and she likes to go on long hikes. She also likes to go on long bike rides
"""
```


### INT4 Inference with AutoGPTQ

Install [AutoGPTQ](https://github.com/AutoGPTQ/AutoGPTQ) from source first

```python
from transformers import AutoModelForCausalLM, AutoTokenizer
quantized_model_dir = "Intel/Mistral-7B-v0.1-int4-inc"
model = AutoModelForCausalLM.from_pretrained(quantized_model_dir,
                                             device_map="auto",
                                             trust_remote_code=False,
                                             )
tokenizer = AutoTokenizer.from_pretrained(quantized_model_dir, use_fast=True)
print(tokenizer.decode(model.generate(**tokenizer("There is a girl who likes adventure,", return_tensors="pt").to(model.device),max_new_tokens=50)[0]))
```



### Evaluate the model 

Install [lm-eval-harness](https://github.com/EleutherAI/lm-evaluation-harness.git) from source, we used the git id f3b7917091afba325af3980a35d8a6dcba03dc3f

```bash
lm_eval  --model hf --model_args pretrained="Intel/Mistral-7B-v0.1-int4-inc",autogptq=True,gptq_use_triton=True --device cuda:0 --tasks lambada_openai,hellaswag,piqa,winogrande,truthfulqa_mc1,openbookqa,boolq,rte,arc_easy,arc_challenge,mmlu  --batch_size 128
```

| Metric         | BF16   | INT4   |
| -------------- | ------ | ------ |
| Avg.           | 0.6306 | 0.6308 |
| mmlu           | 0.5961 | 0.5880 |
| lambada_openai | 0.7561 | 0.7551 |
| hellaswag      | 0.6128 | 0.6079 |
| winogrande     | 0.7443 | 0.7451 |
| piqa           | 0.8079 | 0.8014 |
| truthfulqa_mc1 | 0.2803 | 0.2889 |
| openbookqa     | 0.3280 | 0.3300 |
| boolq          | 0.8373 | 0.8278 |
| rte            | 0.6643 | 0.6968 |
| arc_easy       | 0.8085 | 0.8060 |
| arc_challenge  | 0.5009 | 0.4915 |



## Ethical Considerations and Limitations

The model can produce factually incorrect output, and should not be relied on to produce factually accurate information. Because of the limitations of the pretrained model and the finetuning datasets, it is possible that this model could generate lewd, biased or otherwise offensive outputs.

Therefore, before deploying any applications of the model, developers should perform safety testing.

## Caveats and Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model.

Here are a couple of useful links to learn more about Intel's AI software:

* Intel Neural Compressor [link](https://github.com/intel/neural-compressor)
* Intel Extension for Transformers [link](https://github.com/intel/intel-extension-for-transformers)

## Disclaimer

The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. Please consult an attorney before using this model for commercial purposes.


## Cite

@article{cheng2023optimize,
  title={Optimize weight rounding via signed gradient descent for the quantization of llms},
  author={Cheng, Wenhua and Zhang, Weiwei and Shen, Haihao and Cai, Yiyang and He, Xin and Lv, Kaokao},
  journal={arXiv preprint arXiv:2309.05516},
  year={2023}
}

[arxiv](https://arxiv.org/abs/2309.05516) [github](https://github.com/intel/auto-round)