Hugging Face's logo

neuralmagic
/
Meta-Llama-3.1-8B-Instruct-quantized.w8a16

Text Generation
Transformers
Safetensors
llama
int8
vllm
conversational
text-generation-inference
Inference Endpoints
compressed-tensors
Model card Files Community
3
Meta-Llama-3.1-8B-Instruct-quantized.w8a16 / README.md
alexmarques's picture
alexmarques
Update README.md
38e03ba verified
preview code
|
raw
history blame contribute delete
11.3 kB
metadata
tags:
  - int8
  - vllm
language:
  - en
  - de
  - fr
  - it
  - pt
  - hi
  - es
  - th
pipeline_tag: text-generation
license: llama3.1
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct

Meta-Llama-3.1-8B-Instruct-quantized.w8a16

Model Overview

  • Model Architecture: Meta-Llama-3
    • Input: Text
    • Output: Text
  • Model Optimizations:
    • Weight quantization: INT8
  • Intended Use Cases: Intended for commercial and research use multiple languages. Similarly to Meta-Llama-3.1-8B-Instruct, this models is intended for assistant-like chat.
  • Out-of-scope: Use in any manner that violates applicable laws or regulations (including trade compliance laws).
  • Release Date: 7/23/2024
  • Version: 1.0
  • License(s): Llama3.1
  • Model Developers: Neural Magic

Quantized version of Meta-Llama-3.1-8B-Instruct. It achieves scores within 1% of the scores of the unquantized model for MMLU, ARC-Challenge, GSM-8k, Hellaswag, Winogrande and TruthfulQA.

Model Optimizations

This model was obtained by quantizing the weights of Meta-Llama-3.1-8B-Instruct to INT8 data type. This optimization reduces the number of bits per parameter from 16 to 8, reducing the disk size and GPU memory requirements by approximately 50%.

Only the weights of the linear operators within transformers blocks are quantized. Symmetric per-channel quantization is applied, in which a linear scaling per output dimension maps the INT8 and floating point representations of the quantized weights. The GPTQ algorithm is applied for quantization, as implemented in the llm-compressor library. GPTQ used a 1% damping factor and 256 sequences of 8,192 random tokens.

Deployment

This model can be deployed efficiently using the vLLM backend, as shown in the example below.

from vllm import LLM, SamplingParams
from transformers import AutoTokenizer

model_id = "neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16"
number_gpus = 1
max_model_len = 8192

sampling_params = SamplingParams(temperature=0.6, top_p=0.9, max_tokens=256)

tokenizer = AutoTokenizer.from_pretrained(model_id)

messages = [
    {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
    {"role": "user", "content": "Who are you?"},
]

prompts = tokenizer.apply_chat_template(messages, add_generation_prompt=True, tokenize=False)

llm = LLM(model=model_id, tensor_parallel_size=number_gpus, max_model_len=max_model_len)

outputs = llm.generate(prompts, sampling_params)

generated_text = outputs[0].outputs[0].text
print(generated_text)

vLLM aslo supports OpenAI-compatible serving. See the documentation for more details.

Creation

This model was created by using the llm-compressor library as presented in the code snipet below.

from transformers import AutoTokenizer
from datasets import Dataset
from llmcompressor.transformers import SparseAutoModelForCausalLM, oneshot
from llmcompressor.modifiers.quantization import GPTQModifier
import random

model_id = "meta-llama/Meta-Llama-3.1-8B-Instruct"

num_samples = 256
max_seq_len = 8192

tokenizer = AutoTokenizer.from_pretrained(model_id)

max_token_id = len(tokenizer.get_vocab()) - 1
input_ids = [[random.randint(0, max_token_id) for _ in range(max_seq_len)] for _ in range(num_samples)]
attention_mask = num_samples * [max_seq_len * [1]]
ds = Dataset.from_dict({"input_ids": input_ids, "attention_mask": attention_mask})

recipe = GPTQModifier(
  targets="Linear",
  scheme="W8A16",
  ignore=["lm_head"],
  dampening_frac=0.01,
)

model = SparseAutoModelForCausalLM.from_pretrained(
  model_id,
  device_map="auto",
  trust_remote_code=True,
)

oneshot(
  model=model,
  dataset=ds,
  recipe=recipe,
  max_seq_length=max_seq_len,
  num_calibration_samples=num_samples,
)
model.save_pretrained("Meta-Llama-3.1-8B-Instruct-quantized.w8a16")

Evaluation

The model was evaluated on MMLU, ARC-Challenge, GSM-8K, Hellaswag, Winogrande and TruthfulQA. Evaluation was conducted using the Neural Magic fork of lm-evaluation-harness (branch llama_3.1_instruct) and the vLLM engine. This version of the lm-evaluation-harness includes versions of MMLU, ARC-Challenge and GSM-8K that match the prompting style of Meta-Llama-3.1-Instruct-evals.

Note: Results have been updated after Meta modified the chat template.

Accuracy

Category Benchmark Meta-Llama-3.1-8B-Instruct Meta-Llama-3.1-8B-Instruct-quantized.w8a16 (this model) Recovery
OpenLLM v1
Hellaswag (10-shot) 80.47 80.48 100.0%
Winogrande (5-shot) 78.06 77.51 99.3%
TruthfulQA (0-shot, mc2) 54.48 54.41 99.9%
Average 74.05 74.12 100.1%
Multilingual Portuguese MMLU (5-shot) 59.96 59.79 99.8%
Spanish MMLU (5-shot) 60.25 59.92 99.4%
Italian MMLU (5-shot) 59.23 59.25 100.0%
German MMLU (5-shot) 58.63 58.31 99.5%
French MMLU (5-shot) 59.65 59.57 99.9%
Hindi MMLU (5-shot) 50.10 49.97 99.7%
Thai MMLU (5-shot) 49.12 49.09 99.9%

Reproduction

The results were obtained using the following commands:

MMLU 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,max_model_len=3850,max_gen_toks=10,tensor_parallel_size=1 \
  --tasks mmlu_llama_3.1_instruct \
  --fewshot_as_multiturn \
  --apply_chat_template \
  --num_fewshot 5 \
  --batch_size auto

MMLU-CoT 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,max_model_len=4064,max_gen_toks=1024,tensor_parallel_size=1 \
  --tasks mmlu_cot_0shot_llama_3.1_instruct \
  --apply_chat_template \
  --num_fewshot 0 \
  --batch_size auto

ARC-Challenge 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,max_model_len=3940,max_gen_toks=100,tensor_parallel_size=1 \
  --tasks arc_challenge_llama_3.1_instruct \
  --apply_chat_template \
  --num_fewshot 0 \
  --batch_size auto

GSM-8K 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,max_model_len=4096,max_gen_toks=1024,tensor_parallel_size=1 \
  --tasks gsm8k_cot_llama_3.1_instruct \
  --fewshot_as_multiturn \
  --apply_chat_template \
  --num_fewshot 8 \
  --batch_size auto

Hellaswag 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,add_bos_token=True,max_model_len=4096,tensor_parallel_size=1 \
  --tasks hellaswag \
  --num_fewshot 10 \
  --batch_size auto

Winogrande 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,add_bos_token=True,max_model_len=4096,tensor_parallel_size=1 \
  --tasks winogrande \
  --num_fewshot 5 \
  --batch_size auto

TruthfulQA 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,add_bos_token=True,max_model_len=4096,tensor_parallel_size=1 \
  --tasks truthfulqa \
  --num_fewshot 0 \
  --batch_size auto

MMLU Portuguese 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,max_model_len=3850,max_gen_toks=10,tensor_parallel_size=1 \
  --tasks mmlu_pt_llama_3.1_instruct \
  --fewshot_as_multiturn \
  --apply_chat_template \
  --num_fewshot 5 \
  --batch_size auto

MMLU Spanish 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,max_model_len=3850,max_gen_toks=10,tensor_parallel_size=1 \
  --tasks mmlu_es_llama_3.1_instruct \
  --fewshot_as_multiturn \
  --apply_chat_template \
  --num_fewshot 5 \
  --batch_size auto

MMLU Italian 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,max_model_len=3850,max_gen_toks=10,tensor_parallel_size=1 \
  --tasks mmlu_it_llama_3.1_instruct \
  --fewshot_as_multiturn \
  --apply_chat_template \
  --num_fewshot 5 \
  --batch_size auto

MMLU German 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,max_model_len=3850,max_gen_toks=10,tensor_parallel_size=1 \
  --tasks mmlu_de_llama_3.1_instruct \
  --fewshot_as_multiturn \
  --apply_chat_template \
  --num_fewshot 5 \
  --batch_size auto

MMLU French 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,max_model_len=3850,max_gen_toks=10,tensor_parallel_size=1 \
  --tasks mmlu_fr_llama_3.1_instruct \
  --fewshot_as_multiturn \
  --apply_chat_template \
  --num_fewshot 5 \
  --batch_size auto

MMLU Hindi 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,max_model_len=3850,max_gen_toks=10,tensor_parallel_size=1 \
  --tasks mmlu_hi_llama_3.1_instruct \
  --fewshot_as_multiturn \
  --apply_chat_template \
  --num_fewshot 5 \
  --batch_size auto

MMLU Thai 

lm_eval \
  --model vllm \
  --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a16",dtype=auto,max_model_len=3850,max_gen_toks=10,tensor_parallel_size=1 \
  --tasks mmlu_th_llama_3.1_instruct \
  --fewshot_as_multiturn \
  --apply_chat_template \
  --num_fewshot 5 \
  --batch_size auto