--- language: - en pipeline_tag: text-generation license: mit --- # Phi-3-medium-128k-instruct-quantized.w8a16 ## Model Overview - **Model Architecture:** Phi-3 - **Input:** Text - **Output:** Text - **Model Optimizations:** - **Weight quantization:** INT8 - **Intended Use Cases:** Intended for commercial and research use in English. Similarly to [Phi-3-medium-128k-instruct](https://huggingface.co/microsoft/Phi-3-medium-128k-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:** 7/11/2024 - **Version:** 1.0 - **License(s):** MIT - **Model Developers:** Neural Magic Quantized version of [Phi-3-medium-128k-instruct](https://huggingface.co/microsoft/Phi-3-medium-128k-instruct), a 14 billion-parameter open model trained using the Phi-3 datasets. It achieves an average score of 73.99 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 73.32. ### Model Optimizations This model was obtained by quantizing the weights of [Phi-3-medium-128k-instruct](https://huggingface.co/microsoft/Phi-3-medium-128k-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](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 ### 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 (using 2 GPUs). ```python from vllm import LLM, SamplingParams from transformers import AutoTokenizer model_id = "neuralmagic/Phi-3-medium-128k-instruct-quantized.w8a16" number_gpus = 2 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, trust_remote_code=True, max_model_len=8196, tensor_parallel_size=number_gpus) 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. ### Use with transformers The following example contemplates how the model can be deployed in Transformers using the `generate()` function. ```python from transformers import AutoTokenizer, AutoModelForCausalLM model_id = "neuralmagic/Phi-3-medium-128k-instruct-quantized.w8a16" tokenizer = AutoTokenizer.from_pretrained(model_id) model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype="auto", device_map="auto", trust_remote_code=True, ) messages = [ {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"}, {"role": "user", "content": "Who are you?"}, ] input_ids = tokenizer.apply_chat_template( messages, add_generation_prompt=True, return_tensors="pt" ).to(model.device) outputs = model.generate( input_ids, max_new_tokens=256, do_sample=True, temperature=0.6, top_p=0.9, ) response = outputs[0][input_ids.shape[-1]:] print(tokenizer.decode(response, skip_special_tokens=True)) ``` ## 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 = "microsoft/Phi-3-medium-128k-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("Phi-3-medium-128k-instruct-quantized.w8a16") ``` ## 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 (using 2 GPUs): ``` lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/Phi-3-medium-128k-instruct-quantized.w8a16",dtype=auto,gpu_memory_utilization=0.4,add_bos_token=True,max_model_len=4096,tensor_parallel_size=2 \ --tasks openllm \ --batch_size auto ``` ### Accuracy #### Open LLM Leaderboard evaluation scores
Benchmark Phi-3-medium-128k-instruct Phi-3-medium-128k-instruct-quantized.w8a16(this model) Recovery
MMLU (5-shot) 75.64 76.79 101.5%
ARC Challenge (25-shot) 67.58 69.54 102.9%
GSM-8K (5-shot, strict-match) 83.32 84.31 101.2%
Hellaswag (10-shot) 84.37 84.96 100.7%
Winogrande (5-shot) 75.45 73.64 97.6%
TruthfulQA (0-shot) 53.54 54.73 102.2%
Average 73.32 73.99 100.9%