--- 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.w8a8 ## Model Overview - **Model Architecture:** Meta-Llama-3 - **Input:** Text - **Output:** Text - **Model Optimizations:** - **Activation quantization:** INT8 - **Weight quantization:** INT8 - **Intended Use Cases:** Intended for commercial and research use multiple languages. Similarly to [Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/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:** 7/11/2024 - **Version:** 1.0 - **License(s):** Llama3.1 - **Model Developers:** Neural Magic This model is a quantized version of [Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-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-3.1-8B-Instruct-quantized.w8a8 achieves 105.4% recovery for the Arena-Hard evaluation, 100.3% for OpenLLM v1 (using Meta's prompting when available), 101.5% for OpenLLM v2, 99.7% for HumanEval pass@1, and 98.8% for HumanEval+ pass@1. ### Model Optimizations This model was obtained by quantizing the weights of [Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) to INT8 data type. This optimization reduces the number of bits used to represent weights and activations from 16 to 8, reducing GPU memory requirements (by approximately 50%) and increasing matrix-multiply compute throughput (by approximately 2x). Weight quantization also reduces disk size requirements by approximately 50%. Only weights and activations of the linear operators within transformers blocks are quantized. Weights are quantized with a symmetric static per-channel scheme, where a fixed linear scaling factor is applied between INT8 and floating point representations for each output channel dimension. Activations are quantized with a symmetric dynamic per-token scheme, computing a linear scaling factor at runtime for each token between INT8 and floating point representations. The [GPTQ](https://arxiv.org/abs/2210.17323) algorithm is applied for quantization, as implemented in the [llm-compressor](https://github.com/vllm-project/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](https://docs.vllm.ai/en/latest/) backend, as shown in the example below. ```python from vllm import LLM, SamplingParams from transformers import AutoTokenizer model_id = "neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a8" 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](https://docs.vllm.ai/en/latest/) for more details. ## Creation This model was created by using the [llm-compressor](https://github.com/vllm-project/llm-compressor) library as presented in the code snipet below. ```python 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="W8A8", ignore=["lm_head"], dampening_frac=0.01, ) model = SparseAutoModelForCausalLM.from_pretrained( model_id, device_map="auto", ) 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.w8a8") ``` ## 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](https://docs.vllm.ai/en/stable/) engine. Arena-Hard evaluations were conducted using the [Arena-Hard-Auto](https://github.com/lmarena/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](https://github.com/neuralmagic/lm-evaluation-harness/tree/llama_3.1_instruct) (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](https://huggingface.co/datasets/meta-llama/Meta-Llama-3.1-8B-Instruct-evals) and a few fixes to OpenLLM v2 tasks. HumanEval and HumanEval+ evaluations were conducted using Neural Magic's fork of the [EvalPlus](https://github.com/neuralmagic/evalplus) repository. Detailed model outputs are available as HuggingFace datasets for [Arena-Hard](https://huggingface.co/datasets/neuralmagic/quantized-llama-3.1-arena-hard-evals), [OpenLLM v2](https://huggingface.co/datasets/neuralmagic/quantized-llama-3.1-leaderboard-v2-evals), and [HumanEval](https://huggingface.co/datasets/neuralmagic/quantized-llama-3.1-humaneval-evals). **Note:** Results have been updated after Meta modified the chat template. ### Accuracy <table> <tr> <td><strong>Category</strong> </td> <td><strong>Benchmark</strong> </td> <td><strong>Meta-Llama-3.1-8B-Instruct </strong> </td> <td><strong>Meta-Llama-3.1-8B-Instruct-quantized.w8a8 (this model)</strong> </td> <td><strong>Recovery</strong> </td> </tr> <tr> <td rowspan="1" ><strong>LLM as a judge</strong> </td> <td>Arena Hard </td> <td>25.8 (25.1 / 26.5) </td> <td>27.2 (27.6 / 26.7) </td> <td>105.4% </td> </tr> <tr> <td rowspan="8" ><strong>OpenLLM v1</strong> </td> <td>MMLU (5-shot) </td> <td>68.3 </td> <td>67.8 </td> <td>99.3% </td> </tr> <tr> <td>MMLU (CoT, 0-shot) </td> <td>72.8 </td> <td>72.2 </td> <td>99.1% </td> </tr> <tr> <td>ARC Challenge (0-shot) </td> <td>81.4 </td> <td>81.7 </td> <td>100.3% </td> </tr> <tr> <td>GSM-8K (CoT, 8-shot, strict-match) </td> <td>82.8 </td> <td>84.8 </td> <td>102.5% </td> </tr> <tr> <td>Hellaswag (10-shot) </td> <td>80.5 </td> <td>80.3 </td> <td>99.8% </td> </tr> <tr> <td>Winogrande (5-shot) </td> <td>78.1 </td> <td>78.5 </td> <td>100.5% </td> </tr> <tr> <td>TruthfulQA (0-shot, mc2) </td> <td>54.5 </td> <td>54.7 </td> <td>100.3% </td> </tr> <tr> <td><strong>Average</strong> </td> <td><strong>74.1</strong> </td> <td><strong>74.3</strong> </td> <td><strong>100.3%</strong> </td> </tr> <tr> <td rowspan="7" ><strong>OpenLLM v2</strong> </td> <td>MMLU-Pro (5-shot) </td> <td>30.8 </td> <td>30.9 </td> <td>100.3% </td> </tr> <tr> <td>IFEval (0-shot) </td> <td>77.9 </td> <td>78.0 </td> <td>100.1% </td> </tr> <tr> <td>BBH (3-shot) </td> <td>30.1 </td> <td>31.0 </td> <td>102.9% </td> </tr> <tr> <td>Math-lvl-5 (4-shot) </td> <td>15.7 </td> <td>15.5 </td> <td>98.9% </td> </tr> <tr> <td>GPQA (0-shot) </td> <td>3.7 </td> <td>5.4 </td> <td>146.2% </td> </tr> <tr> <td>MuSR (0-shot) </td> <td>7.6 </td> <td>7.6 </td> <td>100.0% </td> </tr> <tr> <td><strong>Average</strong> </td> <td><strong>27.6</strong> </td> <td><strong>28.0</strong> </td> <td><strong>101.5%</strong> </td> </tr> <tr> <td rowspan="2" ><strong>Coding</strong> </td> <td>HumanEval pass@1 </td> <td>67.3 </td> <td>67.1 </td> <td>99.7% </td> </tr> <tr> <td>HumanEval+ pass@1 </td> <td>60.7 </td> <td>60.0 </td> <td>98.8% </td> </tr> <tr> <td rowspan="9" ><strong>Multilingual</strong> </td> <td>Portuguese MMLU (5-shot) </td> <td>59.96 </td> <td>59.36 </td> <td>99.0% </td> </tr> <tr> <td>Spanish MMLU (5-shot) </td> <td>60.25 </td> <td>59.77 </td> <td>99.2% </td> </tr> <tr> <td>Italian MMLU (5-shot) </td> <td>59.23 </td> <td>58.61 </td> <td>99.0% </td> </tr> <tr> <td>German MMLU (5-shot) </td> <td>58.63 </td> <td>58.23 </td> <td>99.3% </td> </tr> <tr> <td>French MMLU (5-shot) </td> <td>59.65 </td> <td>58.70 </td> <td>98.4% </td> </tr> <tr> <td>Hindi MMLU (5-shot) </td> <td>50.10 </td> <td>49.33 </td> <td>98.5% </td> </tr> <tr> <td>Thai MMLU (5-shot) </td> <td>49.12 </td> <td>48.09 </td> <td>97.9% </td> </tr> </table> ### 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.w8a8",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.w8a8",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.w8a8",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.w8a8",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.w8a8",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.w8a8",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.w8a8",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-3.1-8B-Instruct-quantized.w8a8",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 ``` #### MMLU Portuguese ``` lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",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.w8a8",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.w8a8",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.w8a8",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.w8a8",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.w8a8",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.w8a8",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 ``` #### HumanEval and HumanEval+ ##### Generation ``` python3 codegen/generate.py \ --model neuralmagic/Meta-Llama-3.1-8B-Instruct-quantized.w8a8 \ --bs 16 \ --temperature 0.2 \ --n_samples 50 \ --root "." \ --dataset humaneval ``` ##### Sanitization ``` python3 evalplus/sanitize.py \ humaneval/neuralmagic--Meta-Llama-3.1-8B-Instruct-quantized.w8a8_vllm_temp_0.2 ``` ##### Evaluation ``` evalplus.evaluate \ --dataset humaneval \ --samples humaneval/neuralmagic--Meta-Llama-3.1-8B-Instruct-quantized.w8a8_vllm_temp_0.2-sanitized ```