|
|
--- |
|
|
tags: |
|
|
- fp8 |
|
|
- vllm |
|
|
--- |
|
|
|
|
|
# Qwen2-0.5B-Instruct-FP8 |
|
|
|
|
|
## Model Overview |
|
|
* <h3 style="display: inline;">Model Architecture:</h3> Based on and identical to the Qwen2-0.5B-Instruct architecture |
|
|
* <h3 style="display: inline;">Model Optimizations:</h3> Weights and activations quantized to FP8 |
|
|
* <h3 style="display: inline;">Release Date:</h3> June 14, 2024 |
|
|
* <h3 style="display: inline;">Model Developers:</h3> Neural Magic |
|
|
|
|
|
Qwen2-0.5B-Instruct quantized to FP8 weights and activations using per-tensor quantization through the AutoFP8 repository, ready for inference with vLLM >= 0.5.0. |
|
|
Calibrated with 512 UltraChat samples to achieve 99% performance recovery on the Open LLM Benchmark evaluations. |
|
|
Reduces space on disk by ~30%. |
|
|
Part of the FP8 LLMs for vLLM collection. |
|
|
|
|
|
|
|
|
|
|
|
## Usage and Creation |
|
|
Produced using [AutoFP8 with calibration samples from ultrachat](https://github.com/neuralmagic/AutoFP8/blob/147fa4d9e1a90ef8a93f96fc7d9c33056ddc017a/example_dataset.py). |
|
|
|
|
|
```python |
|
|
from datasets import load_dataset |
|
|
from transformers import AutoTokenizer |
|
|
|
|
|
from auto_fp8 import AutoFP8ForCausalLM, BaseQuantizeConfig |
|
|
|
|
|
pretrained_model_dir = "Qwen/Qwen2-0.5B-Instruct" |
|
|
quantized_model_dir = "Qwen2-0.5B-Instruct-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") |
|
|
|
|
|
model = AutoFP8ForCausalLM.from_pretrained( |
|
|
pretrained_model_dir, quantize_config=quantize_config |
|
|
) |
|
|
model.quantize(examples) |
|
|
model.save_quantized(quantized_model_dir) |
|
|
``` |
|
|
|
|
|
Evaluated through vLLM with the following script: |
|
|
|
|
|
``` |
|
|
#!/bin/bash |
|
|
|
|
|
# Example usage: |
|
|
# CUDA_VISIBLE_DEVICES=0 ./eval_openllm.sh "neuralmagic/Qwen2-0.5B-Instruct-FP8" "tensor_parallel_size=1,max_model_len=4096,add_bos_token=True,gpu_memory_utilization=0.7" |
|
|
|
|
|
export MODEL_DIR=${1} |
|
|
export MODEL_ARGS=${2} |
|
|
|
|
|
declare -A tasks_fewshot=( |
|
|
["arc_challenge"]=25 |
|
|
["winogrande"]=5 |
|
|
["truthfulqa_mc2"]=0 |
|
|
["hellaswag"]=10 |
|
|
["mmlu"]=5 |
|
|
["gsm8k"]=5 |
|
|
) |
|
|
|
|
|
declare -A batch_sizes=( |
|
|
["arc_challenge"]="auto" |
|
|
["winogrande"]="auto" |
|
|
["truthfulqa_mc2"]="auto" |
|
|
["hellaswag"]="auto" |
|
|
["mmlu"]=1 |
|
|
["gsm8k"]="auto" |
|
|
) |
|
|
|
|
|
for TASK in "${!tasks_fewshot[@]}"; do |
|
|
NUM_FEWSHOT=${tasks_fewshot[$TASK]} |
|
|
BATCH_SIZE=${batch_sizes[$TASK]} |
|
|
lm_eval --model vllm \ |
|
|
--model_args pretrained=$MODEL_DIR,$MODEL_ARGS \ |
|
|
--tasks ${TASK} \ |
|
|
--num_fewshot ${NUM_FEWSHOT} \ |
|
|
--write_out \ |
|
|
--show_config \ |
|
|
--device cuda \ |
|
|
--batch_size ${BATCH_SIZE} \ |
|
|
--output_path="results/${TASK}" |
|
|
done |
|
|
``` |
|
|
|
|
|
## Evaluation |
|
|
|
|
|
Evaluated on the Open LLM Leaderboard evaluations through vLLM. |
|
|
|
|
|
### Open LLM Leaderboard evaluation scores |
|
|
| | Qwen2-0.5B-Instruct | Qwen2-0.5B-Instruct-FP8<br>(this model) | |
|
|
| :------------------: | :----------------------: | :------------------------------------------------: | |
|
|
| arc-c<br>25-shot | 31.74 | 32.00 | |
|
|
| hellaswag<br>10-shot | 49.45 | 49.21 | |
|
|
| mmlu<br>5-shot | 43.87 | 43.63 | |
|
|
| truthfulqa<br>0-shot | 39.37 | 39.33 | |
|
|
| winogrande<br>5-shot | 55.49 | 56.59 | |
|
|
| gsm8k<br>5-shot | 37.83 | 36.85 | |
|
|
| **Average<br>Accuracy** | **42.96** | **42.94** | |
|
|
| **Recovery** | **100%** | **99.95%** | |
|
|
|
|
|
|
