You are viewing v4.43.3 version.
A newer version
v4.46.3 is available.
EETQ
The EETQ library supports int8 per-channel weight-only quantization for NVIDIA GPUS. The high-performance GEMM and GEMV kernels are from FasterTransformer and TensorRT-LLM. It requires no calibration dataset and does not need to pre-quantize your model. Moreover, the accuracy degradation is negligible owing to the per-channel quantization.
Make sure you have eetq installed from the relase page
pip install --no-cache-dir https://github.com/NetEase-FuXi/EETQ/releases/download/v1.0.0/EETQ-1.0.0+cu121+torch2.1.2-cp310-cp310-linux_x86_64.whlor via the source code https://github.com/NetEase-FuXi/EETQ. EETQ requires CUDA capability <= 8.9 and >= 7.0
git clone https://github.com/NetEase-FuXi/EETQ.git
cd EETQ/
git submodule update --init --recursive
pip install .An unquantized model can be quantized via “from_pretrained”.
from transformers import AutoModelForCausalLM, EetqConfig
path = "/path/to/model"
quantization_config = EetqConfig("int8")
model = AutoModelForCausalLM.from_pretrained(path, device_map="auto", quantization_config=quantization_config)A quantized model can be saved via “saved_pretrained” and be reused again via the “from_pretrained”.
quant_path = "/path/to/save/quantized/model"
model.save_pretrained(quant_path)
model = AutoModelForCausalLM.from_pretrained(quant_path, device_map="auto")