a22cdea371
Provide an initial support to FP8 computation. This PR is inspired by HuggingFace TGI: huggingface/text-generation-inference#1726 This feature can be enabled with --quantization fp8 or -q fp8 when launching an engine. Algorithm: We still load a model checkpoint in FP16/BF16. After the weights are loaded, Fp8LinearMethod calculates the per-tensor scaling factor of weights and quantizes the weights accordingly. The scaling factor will then be stored for future use. Meanwhile, the per-tensor scaling factor for activations is calculated in every forward pass. Initial Results: Currently tested Mistral-7B on 1xH100. With prompt length ~5 and decoding length 128: BF16: 1.47s FP8: 1.66s I'll try to use larger models and try to find more performance bottleneck. Meanwhile, you're welcome to try this code.
25 lines
872 B
Python
25 lines
872 B
Python
"""Tests whether FP8 computation is enabled correctly.
|
|
|
|
Run `pytest tests/quantization/test_fp8.py --forked`.
|
|
"""
|
|
import pytest
|
|
import torch
|
|
|
|
from vllm.model_executor.layers.quantization import QUANTIZATION_METHODS
|
|
from vllm.model_executor.layers.quantization.fp8 import Fp8LinearMethod
|
|
|
|
capability = torch.cuda.get_device_capability()
|
|
capability = capability[0] * 10 + capability[1]
|
|
|
|
|
|
@pytest.mark.skipif(
|
|
capability < QUANTIZATION_METHODS["fp8"].get_min_capability(),
|
|
reason="FP8 is not supported on this GPU type.")
|
|
def test_load_fp16_model(vllm_runner) -> None:
|
|
llm = vllm_runner("facebook/opt-125m", quantization="fp8")
|
|
|
|
model = llm.model.llm_engine.model_executor.driver_worker.model_runner.model
|
|
fc1 = model.model.decoder.layers[0].fc1
|
|
assert isinstance(fc1.linear_method, Fp8LinearMethod)
|
|
assert fc1.weight.dtype == torch.float8_e4m3fn
|