File size: 9,699 Bytes
59eb0db |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 |
---
license: llama3.1
language:
- en
tags:
- nvidia
- llama3.1
- w4a16
- int4
- vllm
base_model: meta-llama/Llama-3.1-70B-Instruct
library_name: transformers
---
# Llama-3.1-Nemotron-70B-Instruct-HF-quantized.w4a16
## Model Overview
- **Model Architecture:** Llama-3.1-70B-Instruct
- **Input:** Text
- **Output:** Text
- **Model Optimizations:**
- **Weight quantization:** INT4
- **Activation quantization:** None
- **Release Date:** 3/1/2025
- **Version:** 1.0
- **Model Developers:** Neural Magic
Quantized version of [Llama-3.1-Nemotron-70B-Instruct-Hf](https://huggingface.co/nvidia/Llama-3.1-Nemotron-70B-Instruct-HF).
It achieves an average score of 79.48 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 78.92.
### Model Optimizations
This model was obtained by only quantizing the weights to INT4 data type, ready for inference with vLLM >= 0.5.2.
This optimization reduces the number of bits per parameter from 16 to 4, reducing the disk size and GPU memory requirements by approximately 75%. Only the weights of the linear operators within transformers blocks are quantized.
## Deployment
### Use with vLLM
This model can be deployed efficiently using the [vLLM](https://docs.vllm.ai/en/latest/) backend, as shown in the example below.
```python
from transformers import AutoTokenizer
from vllm import LLM, SamplingParams
max_model_len, tp_size = 4096, 2
model_name = "neuralmagic-ent/Llama-3.1-Nemotron-70B-Instruct-HF-quantized.w4a16"
tokenizer = AutoTokenizer.from_pretrained(model_name)
llm = LLM(model=model_name, tensor_parallel_size=tp_size, max_model_len=max_model_len, trust_remote_code=True)
sampling_params = SamplingParams(temperature=0.3, max_tokens=256, stop_token_ids=[tokenizer.eos_token_id])
messages_list = [
[{"role": "user", "content": "Who are you? Please respond in pirate speak!"}],
]
prompt_token_ids = [tokenizer.apply_chat_template(messages, add_generation_prompt=True) for messages in messages_list]
outputs = llm.generate(prompt_token_ids=prompt_token_ids, sampling_params=sampling_params)
generated_text = [output.outputs[0].text for output in outputs]
print(generated_text)
```
vLLM also supports OpenAI-compatible serving. See the [documentation](https://docs.vllm.ai/en/latest/) for more details.
## Creation
This model was created with [llm-compressor](https://github.com/vllm-project/llm-compressor) by running the code snippet below with the following command:
```bash
python quantize.py --model_path nvidia/Llama-3.1-Nemotron-70B-Instruct-HF --quant_path "output_dir" --calib_size 512 --dampening_frac 0.01 --observer mse --actorder group
```
```python
from datasets import load_dataset
from transformers import AutoTokenizer
from llmcompressor.modifiers.quantization import GPTQModifier
from llmcompressor.transformers import SparseAutoModelForCausalLM, oneshot, apply
import argparse
from compressed_tensors.quantization import QuantizationScheme, QuantizationArgs, QuantizationType, QuantizationStrategy
def parse_actorder(value):
# Interpret the input value for --actorder
if value.lower() == "false":
return False
elif value.lower() == "group":
return "group"
else:
raise argparse.ArgumentTypeError("Invalid value for --actorder. Use 'group' or 'False'.")
parser = argparse.ArgumentParser()
parser.add_argument('--model_path', type=str)
parser.add_argument('--quant_path', type=str)
parser.add_argument('--num_bits', type=int, default=4)
parser.add_argument('--sequential_update', type=bool, default=True)
parser.add_argument('--calib_size', type=int, default=256)
parser.add_argument('--dampening_frac', type=float, default=0.05)
parser.add_argument('--observer', type=str, default="minmax")
parser.add_argument(
'--actorder',
type=parse_actorder,
default=False, # Default value is False
help="Specify actorder as 'group' (string) or False (boolean)."
)
args = parser.parse_args()
model = SparseAutoModelForCausalLM.from_pretrained(
args.model_path,
device_map="auto",
torch_dtype="auto",
use_cache=False,
)
tokenizer = AutoTokenizer.from_pretrained(args.model_path)
NUM_CALIBRATION_SAMPLES = args.calib_size
DATASET_ID = "garage-bAInd/Open-Platypus"
DATASET_SPLIT = "train"
ds = load_dataset(DATASET_ID, split=DATASET_SPLIT)
ds = ds.shuffle(seed=42).select(range(NUM_CALIBRATION_SAMPLES))
def preprocess(example):
concat_txt = example["instruction"] + "\n" + example["output"]
return {"text": concat_txt}
ds = ds.map(preprocess)
def tokenize(sample):
return tokenizer(
sample["text"],
padding=False,
truncation=False,
add_special_tokens=True,
)
ds = ds.map(tokenize, remove_columns=ds.column_names)
quant_scheme = QuantizationScheme(
targets=["Linear"],
weights=QuantizationArgs(
num_bits=args.num_bits,
type=QuantizationType.INT,
symmetric=True,
group_size=128,
strategy=QuantizationStrategy.GROUP,
observer=args.observer,
actorder=args.actorder
),
input_activations=None,
output_activations=None,
)
recipe = [
GPTQModifier(
targets=["Linear"],
ignore=["lm_head"],
sequential_update=args.sequential_update,
dampening_frac=args.dampening_frac,
config_groups={"group_0": quant_scheme},
)
]
oneshot(
model=model,
dataset=ds,
recipe=recipe,
num_calibration_samples=args.calib_size,
)
# Save to disk compressed.
SAVE_DIR = args.quant_path
model.save_pretrained(SAVE_DIR, save_compressed=True)
tokenizer.save_pretrained(SAVE_DIR)
```
## Evaluation
The model was evaluated on OpenLLM Leaderboard [V1](https://huggingface.co/spaces/open-llm-leaderboard-old/open_llm_leaderboard) and [V2](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard#/), using the following commands:
OpenLLM Leaderboard V1:
```
lm_eval \
--model vllm \
--model_args pretrained="neuralmagic-ent/Llama-3.1-Nemotron-70B-Instruct-HF-quantized.w4a16",dtype=auto,add_bos_token=True,max_model_len=4096,tensor_parallel_size=1,gpu_memory_utilization=0.8,enable_chunked_prefill=True,trust_remote_code=True \
--tasks openllm \
--write_out \
--batch_size auto \
--output_path output_dir \
--show_config
```
OpenLLM Leaderboard V2:
```
lm_eval \
--model vllm \
--model_args pretrained="neuralmagic-ent/Llama-3.1-Nemotron-70B-Instruct-HF-quantized.w4a16",dtype=auto,add_bos_token=False,max_model_len=4096,tensor_parallel_size=1,gpu_memory_utilization=0.8,enable_chunked_prefill=True,trust_remote_code=True \
--apply_chat_template \
--fewshot_as_multiturn \
--tasks leaderboard \
--write_out \
--batch_size auto \
--output_path output_dir \
--show_config
```
### Accuracy
#### OpenLLM Leaderboard V1 evaluation scores
| Metric | nvidia/Llama-3.1-Nemotron-70B-Instruct-HF | neuralmagic-ent/Llama-3.1-Nemotron-70B-Instruct-HF-quantized.w4a16 |
|-----------------------------------------|:---------------------------------:|:-------------------------------------------:|
| ARC-Challenge (Acc-Norm, 25-shot) | 71.76 | 72.53 |
| GSM8K (Strict-Match, 5-shot) | 82.94 | 86.88 |
| HellaSwag (Acc-Norm, 10-shot) | 87.61 | 87.09 |
| MMLU (Acc, 5-shot) | 82.56 | 81.97 |
| TruthfulQA (MC2, 0-shot) | 64.85 | 64.01 |
| Winogrande (Acc, 5-shot) | 83.82 | 84.37 |
| **Average Score** | **78.92** | **79.48** |
| **Recovery** | **100.00** | **100.71** |
#### OpenLLM Leaderboard V2 evaluation scores
| Metric | nvidia/Llama-3.1-Nemotron-70B-Instruct-HF | neuralmagic-ent/Llama-3.1-Nemotron-70B-Instruct-HF-quantized.w4a16 |
|---------------------------------------------------------|:---------------------------------:|:-------------------------------------------:|
| IFEval (Inst-and-Prompt Level Strict Acc, 0-shot) | 74.3 | 72.02 |
| BBH (Acc-Norm, 3-shot) | 47.39 | 48.33 |
| Math-Hard (Exact-Match, 4-shot) | 23.21 | 25.69 |
| GPQA (Acc-Norm, 0-shot) | 1.2 | 1.96 |
| MUSR (Acc-Norm, 0-shot) | 13.2 | 12.5 |
| MMLU-Pro (Acc, 5-shot) | 43.45 | 43.57 |
| **Average Score** | **33.79** | **34.01** |
| **Recovery** | **100.00** | **100.65** |
|