Upload New File

README.md

+# SAM
+
+## 项目简介
+
+SAM 是 Meta 推出的通用图像分割模型，通过三阶段架构实现零样本分割。它支持点、框等提示方式，能精准分割图像对象，通用性强、效果精准。
+
+## 环境部署
+
+### 1. 拉取镜像
+
+```
+docker pull image.sourcefind.cn:5000/dcu/admin/base/vllm:0.9.2-ubuntu22.04-dtk25.04.1-rc5-rocblas101839-0811-das1.6-py3.10-20250913-rc1
+```
+
+### 2. 创建容器
+
+```
+docker run -it \
+--network=host \
+--hostname=localhost \
+--name=SAM_infer \
+-v /opt/hyhal:/opt/hyhal:ro \
+-v $PWD:/workspace \
+--ipc=host \
+--device=/dev/kfd \
+--device=/dev/mkfd \
+--device=/dev/dri \
+--shm-size=512G \
+--privileged \
+--group-add video \
+--cap-add=SYS_PTRACE \
+--security-opt seccomp=unconfined \
+image.sourcefind.cn:5000/dcu/admin/base/vllm:0.9.2-ubuntu22.04-dtk25.04.1-rc5-rocblas101839-0811-das1.6-py3.10-20250913-rc1 \
+/bin/bash  
+```
+
+## 测试步骤
+
+### 1. 拉取优化后代码仓库
+
+```
+git clone http://developer.sourcefind.cn/codes/bw-bestperf/SAM.git
+cd SAM
+```
+
+### 2. 安装依赖
+
+```
+pip install pycocotools -i https://pypi.tuna.tsinghua.edu.cn/simple
+pip install onnx -i https://pypi.tuna.tsinghua.edu.cn/simple
+pip install onnxruntime-1.19.2+das.opt1.dtk25041-cp310-cp310-manylinux_2_28_x86_64.whl 
+pip install -e .
+```
+
+ ### 4. 下载模型与数据集
+
+```
+# 模型下载
+wget https://dl.fbaipublicfiles.com/segment_anything/sam_vit_h_4b8939.pth
+# 数据集下载
+wget -O  coco_2017.zip  https://ai-studio-online.bj.bcebos.com/v1/9b15388b52fa45bb8f4e66343d6e2fa690630a526df64fcd90d8ed8a124627bd?responseContentDisposition=attachment%3Bfilename%3Dmini-coco2017.zip&authorization=bce-auth-v1%2F5cfe9a5e1454405eb2a975c43eace6ec%2F2025-10-11T02%3A23%3A00Z%2F60%2F%2Fdb2d898bccb4ef8674047b3eda1aaaa094a71c8f87a7d2b51f49ee16e30c77ac
+# 数据集解压
+unzip coco_2017.zip  -d mini_coco/
+```
+
+### 5. 测试命令
+
+```
+python scripts/amg_amp_compile_v2_warmup.py --checkpoint sam_vit_h_4b8939.pth --model-type vit_h --input mini_coco/val2017 --output ./output
+```
+
+
+