File size: 6,107 Bytes
61dc0fd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3de5e69
 
 
61dc0fd
 
 
 
3de5e69
61dc0fd
 
 
3de5e69
 
 
 
 
 
 
 
 
 
 
 
61dc0fd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3de5e69
61dc0fd
 
 
 
 
3de5e69
 
 
d9b36fe
3de5e69
 
 
 
 
 
 
 
61dc0fd
 
 
 
c4f9132
61dc0fd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c4f9132
 
61dc0fd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
---
license: other
language:
- en
- de
- fr
- it
- pt
- hi
- es
- th
library_name: transformers
pipeline_tag: text-generation
tags:
- llama-3.1
- meta
- autoawq
---

> [!IMPORTANT]
> This repository is a community-driven quantized version of the original model [`meta-llama/Meta-Llama-3.1-405B-Instruct`](https://huggingface.co/meta-llama/Meta-Llama-3.1-405B-Instruct) which is the FP16 half-precision official version released by Meta AI.

## Model Information

The Meta Llama 3.1 collection of multilingual large language models (LLMs) is a collection of pretrained and instruction tuned generative models in 8B, 70B and 405B sizes (text in/text out). The Llama 3.1 instruction tuned text only models (8B, 70B, 405B) are optimized for multilingual dialogue use cases and outperform many of the available open source and closed chat models on common industry benchmarks.

This repository contains [`meta-llama/Meta-Llama-3.1-405B-Instruct`](https://huggingface.co/meta-llama/Meta-Llama-3.1-405B-Instruct) quantized using [AutoAWQ](https://github.com/casperhansen/AutoAWQ) from FP16 down to INT4 using the GEMM kernels performing zero-point quantization with a group size of 128.

## Model Usage

> [!NOTE]
> In order to run the inference with Llama 3.1 405B Instruct AWQ in INT4, around 203 GiB of VRAM are needed only for loading the model checkpoint, without including the KV cache or the CUDA graphs, meaning that there should be a bit over that VRAM available.

In order to use the current quantized model, support is offered for different solutions as `transformers`, `autoawq`, or `text-generation-inference`.

### 🤗 transformers

In order to run the inference with Llama 3.1 405B Instruct AWQ in INT4, both `torch` and `autoawq` need to be installed as:

```bash
pip install "torch>=2.2.0,<2.3.0" autoawq --upgrade
```

Then, the latest version of `transformers` need to be installed, being 4.43.0 or higher, as:

```bash
pip install "transformers[accelerate]>=4.43.0" --upgrade
```

To run the inference on top of Llama 3.1 405B Instruct AWQ in INT4 precision, the AWQ model can be instantiated as any other causal language modeling model via `AutoModelForCausalLM` and run the inference normally.

```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "hugging-quants/Meta-Llama-3.1-405B-Instruct-AWQ-INT4"
prompt = [
  {"role": "system", "content": "You are a helpful assistant, that responds as a pirate."},
  {"role": "user", "content": "What's Deep Learning?"},
]

tokenizer = AutoTokenizer.from_pretrained(model_id)

inputs = tokenizer.apply_chat_template(prompt, tokenize=True, add_generation_prompt=True, return_tensors="pt").cuda()

model = AutoModelForCausalLM.from_pretrained(
  model_id,
  torch_dtype=torch.float16,
  low_cpu_mem_usage=True,
  device_map="auto",
)

outputs = model.generate(inputs, do_sample=True, max_new_tokens=256)
print(tokenizer.batch_decode(outputs, skip_special_tokens=True))
```

### AutoAWQ

In order to run the inference with Llama 3.1 405B Instruct AWQ in INT4, both `torch` and `autoawq` need to be installed as:

```bash
pip install "torch>=2.2.0,<2.3.0" autoawq --upgrade
```

Then, the latest version of `transformers` need to be installed, being 4.43.0 or higher, as:

```bash
pip install "transformers[accelerate]>=4.43.0" --upgrade
```

Alternatively, one may want to run that via `AutoAWQ` even though it's built on top of 🤗 `transformers`, which is the recommended approach instead as described above.

```python
import torch
from awq import AutoAWQForCausalLM
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "hugging-quants/Meta-Llama-3.1-405B-Instruct-AWQ-INT4"
prompt = [
  {"role": "system", "content": "You are a helpful assistant, that responds as a pirate."},
  {"role": "user", "content": "What's Deep Learning?"},
]

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoAWQForCausalLM.from_pretrained(
  model_id,
  torch_dtype=torch.float16,
  low_cpu_mem_usage=True,
  device_map="auto",
)

inputs = tokenizer.apply_chat_template(prompt, tokenize=True, add_generation_prompt=True, return_tensors="pt", return_dict=True).to('cuda')
outputs = model.generate(**inputs, do_sample=True, max_new_tokens=256)
print(tokenizer.batch_decode(outputs, skip_special_tokens=True))
```

The AutoAWQ script has been adapted from [AutoAWQ/examples/generate.py](https://github.com/casper-hansen/AutoAWQ/blob/main/examples/generate.py).

### 🤗 Text Generation Inference (TGI)

Coming soon!

## Quantization Reproduction

> [!NOTE]
> In order to quantize Llama 3.1 405B Instruct using AutoAWQ, you will need to use an instance with at least enough CPU RAM to fit the whole model i.e. ~800GiB, and an NVIDIA GPU with 80GiB of VRAM to quantize it.

In order to quantize Llama 3.1 405B Instruct, first install `torch` and `autoawq` as follows:

```bash
pip install "torch>=2.2.0,<2.3.0" autoawq --upgrade
```

Otherwise the quantization may fail, since the AutoAWQ kernels are built with PyTorch 2.2.1, meaning that those will break with PyTorch 2.3.0.

Then install the latest version of `transformers` as follows:

```bash
pip install "transformers>=4.43.0" --upgrade
```

And then, run the following script, adapted from [`AutoAWQ/examples/quantize.py`](https://github.com/casper-hansen/AutoAWQ/blob/main/examples/quantize.py) as follows:

```python
from awq import AutoAWQForCausalLM
from transformers import AutoTokenizer

model_path = "meta-llama/Meta-Llama-3.1-405B-Instruct"
quant_path = "hugging-quants/Meta-Llama-3.1-405B-Instruct-AWQ-INT4"
quant_config = {
  "zero_point": True,
  "q_group_size": 128,
  "w_bit": 4,
  "version": "GEMM",
}

# Load model
model = AutoAWQForCausalLM.from_pretrained(
  model_path, **{"low_cpu_mem_usage": True, "use_cache": False}
)
tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)

# Quantize
model.quantize(tokenizer, quant_config=quant_config)

# Save quantized model
model.save_quantized(quant_path)
tokenizer.save_pretrained(quant_path)

print(f'Model is quantized and saved at "{quant_path}"')
```