# Mistral-Nemo-Instruct-2407-FP8 ## Model Overview - **Model Architecture:** Mistral-Nemo - **Input:** Text - **Output:** Text - **Model Optimizations:** - **Weight quantization:** FP8 - **Activation quantization:** FP8 - **Intended Use Cases:** Intended for commercial and research use in English. Similarly to [Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-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). Use in languages other than English. - **Release Date:** 7/18/2024 - **Version:** 1.0 - **License(s):** [apache-2.0](https://huggingface.co/datasets/choosealicense/licenses/blob/main/markdown/apache-2.0.md) - **Model Developers:** Neural Magic Quantized version of [Mistral-Nemo-Instruct-2407](https://huggingface.co/mistralai/Mistral-Nemo-Instruct-2407). It achieves an average score of 71.28 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 71.61. ### Model Optimizations This model was obtained by quantizing the weights and activations of [Mistral-Nemo-Instruct-2407](https://huggingface.co/mistralai/Mistral-Nemo-Instruct-2407) to FP8 data type, ready for inference with vLLM >= 0.5.0. 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 and activations of the linear operators within transformers blocks are quantized. Symmetric per-tensor quantization is applied, in which a single linear scaling maps the FP8 representations of the quantized weights and activations. [AutoFP8](https://github.com/neuralmagic/AutoFP8) is used for quantization with 512 sequences of UltraChat. ## 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 vllm import LLM, SamplingParams from transformers import AutoTokenizer model_id = "neuralmagic/Mistral-Nemo-Instruct-2407-FP8" sampling_params = SamplingParams(temperature=0.3, 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, tokenize=False) llm = LLM(model=model_id, max_model_len=4096) 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 applying [AutoFP8 with calibration samples from ultrachat](https://github.com/neuralmagic/AutoFP8/blob/147fa4d9e1a90ef8a93f96fc7d9c33056ddc017a/example_dataset.py), as presented in the code snipet below. Although AutoFP8 was used for this particular model, Neural Magic is transitioning to using [llm-compressor](https://github.com/vllm-project/llm-compressor) which supports several quantization schemes and models not supported by AutoFP8. Note that ```transformers``` must be built from source. ```python from datasets import load_dataset from transformers import AutoTokenizer from auto_fp8 import AutoFP8ForCausalLM, BaseQuantizeConfig pretrained_model_dir = "mistralai/Mistral-Nemo-Instruct-2407" quantized_model_dir = "Mistral-Nemo-Instruct-2407-FP8" tokenizer = AutoTokenizer.from_pretrained(pretrained_model_dir, use_fast=True, model_max_length=4096) tokenizer.pad_token = tokenizer.eos_token ds = load_dataset("mgoin/ultrachat_2k", split="train_sft").select(range(512)) examples = [tokenizer.apply_chat_template(batch["messages"], tokenize=False) for batch in ds] examples = tokenizer(examples, padding=True, truncation=True, return_tensors="pt").to("cuda") quantize_config = BaseQuantizeConfig( quant_method="fp8", activation_scheme="static" ignore_patterns=["re:.*lm_head"], ) model = AutoFP8ForCausalLM.from_pretrained( pretrained_model_dir, quantize_config=quantize_config ) model.quantize(examples) model.save_quantized(quantized_model_dir) ``` ## Evaluation The model was evaluated on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) leaderboard tasks (version 1) with the [lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness/tree/383bbd54bc621086e05aa1b030d8d4d5635b25e6) (commit 383bbd54bc621086e05aa1b030d8d4d5635b25e6) and the [vLLM](https://docs.vllm.ai/en/stable/) engine, using the following command. Note that ```vllm``` must be built from source. ``` lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/Mistral-Nemo-Instruct-2407-FP8",dtype=auto,gpu_memory_utilization=0.4,max_model_len=4096 \ --tasks openllm \ --batch_size auto ``` ### Accuracy #### Open LLM Leaderboard evaluation scores

Benchmark	Mistral-Nemo-Instruct-2407	Mistral-Nemo-Instruct-2407-FP8(this model)	Recovery
MMLU (5-shot)	68.35	68.50	100.2%
ARC Challenge (25-shot)	65.53	64.68	98.70%
GSM-8K (5-shot, strict-match)	74.45	73.01	98.06%
Hellaswag (10-shot)	84.32	84.18	99.83%
Winogrande (5-shot)	82.16	82.32	100.1%
TruthfulQA (0-shot)	54.85	54.96	100.2%
Average	71.61	71.28	99.53%