RedHatAI/Meta-Llama-3.1-70B-Instruct-quantized.w4a16 · Hugging Face

Model Overview

Model Architecture: Meta-Llama-3Input: Text

Output: Text

Model Optimizations:Weight quantization: INT4

Intended Use Cases: Intended for commercial and research use in English. Similarly to

Meta-Llama-3.1-70B-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). Use in languages other than English.

Release Date: 7/26/2024

Version: 1.0

License(s): Llama3.1

Model Developers: Neural Magic

This model is a quantized version of

Meta-Llama-3.1-70B-Instruct

. It was evaluated on a several tasks to assess the its quality in comparison to the unquatized model, including multiple-choice, math reasoning, and open-ended text generation. Meta-Llama-3.1-70B-Instruct-quantized.w4a16 achieves 100.0% recovery for the Arena-Hard evaluation, 99.4% for OpenLLM v1 (using Meta's prompting when available), 97.4% for OpenLLM v2, 101.0% for HumanEval pass@1, and 99.2% for HumanEval+ pass@1.

Model Optimizations

This model was obtained by quantizing the weights of

Meta-Llama-3.1-70B-Instruct

to INT4 data type. 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. Symmetric per-group quantization is applied, in which a linear scaling per group of 128 parameters maps the INT4 and floating point representations of the quantized weights. The

GPTQ

algorithm is applied for quantization, as implemented in the

AutoGPTQ

library. GPTQ used a 1% damping factor and 512 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-3.1-70B-Instruct-quantized.w4a16" 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 applying the

AutoGPTQ

library as presented in the code snipet below. Although AutoGPTQ was used for this particular model, Neural Magic is transitioning to using

llm-compressor

which supports several quantization schemes and models not supported by AutoGPTQ.

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-3.1-70B-Instruct" num_samples = 512 max_seq_len = 8192 tokenizer = AutoTokenizer.from_pretrained(model_id) preprocess_fn = lambda example: {"text": "Below is an instruction that describes a task. Write a response that appropriately completes the request.\n\n{text}".format_map(example)} dataset_name = "neuralmagic/LLM_compression_calibration" dataset = load_dataset(dataset_name, split="train") ds = dataset.shuffle().select(range(num_samples)) ds = ds.map(preprocess_fn) recipe = GPTQModifier( targets="Linear", scheme="W4A16", 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-3.1-70B-Instruct-quantized.w4a16") Evaluation

This model was evaluated on the well-known Arena-Hard, OpenLLM v1, OpenLLM v2, HumanEval, and HumanEval+ benchmarks. In all cases, model outputs were generated with the

vLLM

engine.

Arena-Hard evaluations were conducted using the

Arena-Hard-Auto

repository. The model generated a single answer for each prompt form Arena-Hard, and each answer was judged twice by GPT-4. We report below the scores obtained in each judgement and the average.

OpenLLM v1 and v2 evaluations were conducted using Neural Magic's fork of

lm-evaluation-harness

(branch llama_3.1_instruct). 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

and a few fixes to OpenLLM v2 tasks.

HumanEval and HumanEval+ evaluations were conducted using Neural Magic's fork of the

EvalPlus

repository.

Detailed model outputs are available as HuggingFace datasets for

Arena-Hard

,

OpenLLM v2

, and

HumanEval

.

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

Accuracy

BenchmarkMeta-Llama-3.1-70B-Instruct Meta-Llama-3.1-70B-Instruct-quantized.w4a16 (this model)RecoveryArena Hard57.0 (55.8 / 58.2) 57.0 (57.1 / 56.8) 100.0% OpenLLM v1MMLU (5-shot) 83.9 83.6 99.5% MMLU (CoT, 0-shot) 86.2 85.6 99.2% ARC Challenge (0-shot) 93.3 92.8 99.5% GSM-8K (CoT, 8-shot, strict-match) 95.4 94.4 99.0% Hellaswag (10-shot) 86.7 86.3 99.5% Winogrande (5-shot) 85.3 85.5 100.2% TruthfulQA (0-shot, mc2) 60.7 59.8 98.6% Average84.5083.9899.4%OpenLLM v2MMLU-Pro (5-shot) 48.1 47.3 98.2% IFEval (0-shot) 86.4 85.7 99.2% BBH (3-shot) 55.8 55.0 98.6% Math-|v|-5 (4-shot) 26.1 24.4 93.5% GPQA (0-shot) 15.4 13.9 89.9% MuSR (0-shot) 18.2 17.3 95.0% Average41.740.697.4%CodingHumanEval pass@1 79.7 80.5 101.0% HumanEval+ pass@1 74.8 74.2 99.2% Reproduction

The results were obtained using the following commands:

MMLU

lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/Meta-Llama-3.1-70B-Instruct-quantized.w4a16",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-3.1-70B-Instruct-quantized.w4a16",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-3.1-70B-Instruct-quantized.w4a16",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-3.1-70B-Instruct-quantized.w4a16",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-3.1-70B-Instruct-quantized.w4a16",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-3.1-70B-Instruct-quantized.w4a16",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-3.1-70B-Instruct-quantized.w4a16",dtype=auto,add_bos_token=True,max_model_len=4096,tensor_parallel_size=1 \ --tasks truthfulqa \ --num_fewshot 0 \ --batch_size auto OpenLLM v2

lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/Meta-Llama-3.1-70B-Instruct-quantized.w4a16",dtype=auto,max_model_len=4096,tensor_parallel_size=1,enable_chunked_prefill=True \ --apply_chat_template \ --fewshot_as_multiturn \ --tasks leaderboard \ --batch_size auto HumanEval and HumanEval+

Generation python3 codegen/generate.py \ --model neuralmagic/Meta-Llama-3.1-70B-Instruct-quantized.w4a16 \ --bs 16 \ --temperature 0.2 \ --n_samples 50 \ --root "." \ --dataset humaneval Sanitization python3 evalplus/sanitize.py \ humaneval/neuralmagic--Meta-Llama-3.1-70B-Instruct-quantized.w4a16_vllm_temp_0.2 Evaluation evalplus.evaluate \ --dataset humaneval \ --samples humaneval/neuralmagic--Meta-Llama-3.1-70B-Instruct-quantized.w4a16_vllm_temp_0.2-sanitized