PTQ-Level Efficiency, QAT-Level Accuracy | 训练后量化的效率,量化感知训练的精度
update readme for Intel xpu usage (#1441)
chen, suyuef16d57d7c1c5
diff --git a/README.md b/README.mdindex 57c1a4124..c0a643eec 100644--- a/README.md+++ b/README.md@@ -97,22 +97,31 @@ Includes [multiple gpus quantization](https://github.com/intel/auto-round/blob/m### Install from pypi```bash-# CPU/Intel GPU/CUDA+# CPU(Xeon)/GPU(CUDA)pip install auto-round-# HPU+# HPU(Gaudi)+# install inside the hpu docker container, e.g. vault.habana.ai/gaudi-docker/1.23.0/ubuntu24.04/habanalabs/pytorch-installer-2.9.0:latestpip install auto-round-hpu++# XPU(Intel GPU)+pip install torch --index-url https://download.pytorch.org/whl/xpu+pip install auto-round```<details><summary>Build from Source</summary>```bash- # CPU/Intel GPU/CUDA+ # CPU(Xeon)/GPU(CUDA)pip install .- # HPU+ # HPU(Gaudi)python setup.py install hpu++ # XPU(Intel GPU)+ pip install torch --index-url https://download.pytorch.org/whl/xpu+ pip install .```</details>diff --git a/README_CN.md b/README_CN.mdindex 54212c74d..4ca044b73 100644--- a/README_CN.md+++ b/README_CN.md@@ -86,22 +86,31 @@ AutoRound 是专为大语言模型(LLMs)和视觉-语言模型(VLMs)设### 从 PyPI 安装```shell-# CPU / Intel GPU / CUDA+# CPU(Xeon)/GPU(CUDA)pip install auto-round-# HPU+# HPU(Gaudi)+# 在 hpu docker container 中安装, e.g. vault.habana.ai/gaudi-docker/1.23.0/ubuntu24.04/habanalabs/pytorch-installer-2.9.0:latestpip install auto-round-hpu++# XPU(Intel GPU)+pip install torch --index-url https://download.pytorch.org/whl/xpu+pip install auto-round```<details><summary>从源码编译安装</summary>```bash- # CPU/Intel GPU/CUDA+ # CPU(Xeon)/GPU(CUDA)pip install .- # HPU+ # HPU(Gaudi)python setup.py install hpu++ # XPU(Intel GPU)+ pip install torch --index-url https://download.pytorch.org/whl/xpu+ pip install .```</details>diff --git a/setup.py b/setup.pyindex c78aa1d78..2e391a8f1 100644--- a/setup.py+++ b/setup.py@@ -137,7 +137,6 @@ if __name__ == "__main__":package_name = "auto-round"- # From v0.9.3, auto-round-hpu will be published to replace auto-round-lib.hpu_build = "hpu" in sys.argvif hpu_build:sys.argv.remove("hpu")