--- tags: - fp8 - vllm license: gemma --- # gemma-2-2b-it-FP8 ## Model Overview - **Model Architecture:** Gemma 2 - **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 [gemma-2-2b-it](https://huggingface.co/google/gemma-2-2b-it), 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:** 8/7/2024 - **Version:** 1.0 - **License(s):** [gemma](https://ai.google.dev/gemma/terms) - **Model Developers:** Neural Magic Quantized version of [gemma-2-2b-it](https://huggingface.co/google/gemma-2-2b-it). It achieves an average score of 58.41 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 58.80. ### Model Optimizations This model was obtained by quantizing the weights and activations of [gemma-2-2b-it](https://huggingface.co/google/gemma-2-2b-it) to FP8 data type, ready for inference with vLLM built from source. This optimization reduces the number of bits per parameter from 16 to 8, reducing the disk size and GPU memory requirements by approximately 50%. In particular, this model can now be loaded and evaluated with a single node of 8xH100 GPUs, as opposed to multiple nodes. 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. [LLM Compressor](https://github.com/vllm-project/llm-compressor) 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/gemma-2-2b-it-FP8" sampling_params = SamplingParams(temperature=0.6, top_p=0.9, max_tokens=256) tokenizer = AutoTokenizer.from_pretrained(model_id) messages = [ {"role": "user", "content": "Who are you? Please respond in pirate speak!"}, ] prompts = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) llm = LLM(model=model_id) outputs = llm.generate(prompts, sampling_params) generated_text = outputs[0].outputs[0].text print(generated_text) ``` vLLM also supports OpenAI-compatible serving. See the [documentation](https://docs.vllm.ai/en/latest/) for more details. ## Creation This model was created by applying [LLM Compressor with calibration samples from UltraChat](https://github.com/vllm-project/llm-compressor/blob/sa/big_model_support/examples/big_model_offloading/big_model_w8a8_calibrate.py), as presented in the code snipet below. ```python import torch from datasets import load_dataset from transformers import AutoTokenizer from llmcompressor.transformers import SparseAutoModelForCausalLM, oneshot from llmcompressor.transformers.compression.helpers import ( calculate_offload_device_map, custom_offload_device_map, ) recipe = """ quant_stage: quant_modifiers: QuantizationModifier: ignore: ["lm_head"] config_groups: group_0: weights: num_bits: 8 type: float strategy: tensor dynamic: false symmetric: true input_activations: num_bits: 8 type: float strategy: tensor dynamic: false symmetric: true targets: ["Linear"] """ model_stub = "google/gemma-2-2b-it-FP8" model_name = model_stub.split("/")[-1] device_map = calculate_offload_device_map( model_stub, reserve_for_hessians=False, num_gpus=1, torch_dtype=torch.float16 ) model = SparseAutoModelForCausalLM.from_pretrained( model_stub, torch_dtype=torch.float16, device_map=device_map ) tokenizer = AutoTokenizer.from_pretrained(model_stub) output_dir = f"./{model_name}-FP8" DATASET_ID = "HuggingFaceH4/ultrachat_200k" DATASET_SPLIT = "train_sft" NUM_CALIBRATION_SAMPLES = 512 MAX_SEQUENCE_LENGTH = 4096 ds = load_dataset(DATASET_ID, split=DATASET_SPLIT) ds = ds.shuffle(seed=42).select(range(NUM_CALIBRATION_SAMPLES)) def preprocess(example): return { "text": tokenizer.apply_chat_template( example["messages"], tokenize=False, ) } ds = ds.map(preprocess) def tokenize(sample): return tokenizer( sample["text"], padding=False, max_length=MAX_SEQUENCE_LENGTH, truncation=True, add_special_tokens=False, ) ds = ds.map(tokenize, remove_columns=ds.column_names) oneshot( model=model, output_dir=output_dir, dataset=ds, recipe=recipe, max_seq_length=MAX_SEQUENCE_LENGTH, num_calibration_samples=NUM_CALIBRATION_SAMPLES, save_compressed=True, ) ``` ## 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: ``` lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/gemma-2-2b-it-FP8",dtype=auto,gpu_memory_utilization=0.4,add_bos_token=True,max_model_len=4096 \ --tasks openllm \ --batch_size auto ``` ### Accuracy #### Open LLM Leaderboard evaluation scores
Benchmark gemma-2-2b-it gemma-2-2b-it-FP8(this model) Recovery
MMLU (5-shot) 56.93 56.87 99.89%
ARC Challenge (25-shot) 57.76 57.25 99.12%
GSM-8K (5-shot, strict-match) 45.11 44.35 98.32%
Hellaswag (10-shot) 71.22 70.87 99.51%
Winogrande (5-shot) 68.67 68.11 99.18%
TruthfulQA (0-shot) 53.11 52.99 99.77%
Average 58.80 58.41 99.33%