fix:add makefile and dockerfile

Dockerfile

@@ -0,0 +1,34 @@
+FROM pytorch/pytorch:2.3.1-cuda12.1-cudnn8-devel
+
+# Arguments to build Docker Image using CUDA
+ARG USE_CUDA=0
+ARG TORCH_ARCH="7.0;7.5;8.0;8.6"
+
+ENV AM_I_DOCKER True
+ENV BUILD_WITH_CUDA "${USE_CUDA}"
+ENV TORCH_CUDA_ARCH_LIST="${TORCH_ARCH}"
+ENV CUDA_HOME /usr/local/cuda-12.1/
+# Ensure CUDA is correctly set up
+ENV PATH /usr/local/cuda-12.1/bin:${PATH}
+ENV LD_LIBRARY_PATH /usr/local/cuda-12.1/lib64:${LD_LIBRARY_PATH}
+
+# Install required packages and specific gcc/g++
+RUN apt-get update && apt-get install --no-install-recommends wget ffmpeg=7:* \
+    libsm6=2:* libxext6=2:* git=1:* nano vim=2:* ninja-build gcc-10 g++-10 -y \
+    && apt-get clean && apt-get autoremove && rm -rf /var/lib/apt/lists/*
+
+ENV CC=gcc-10
+ENV CXX=g++-10
+
+RUN mkdir -p /home/appuser/Grounded-SAM-2
+COPY . /home/appuser/Grounded-SAM-2/
+
+WORKDIR /home/appuser/Grounded-SAM-2
+
+
+# Install essential Python packages
+RUN python -m pip install --upgrade pip setuptools wheel numpy
+
+# Install segment_anything package in editable mode
+RUN python -m pip install -e .
+

Makefile

@@ -0,0 +1,47 @@
+# Get version of CUDA and enable it for compilation if CUDA > 11.0
+# This solves https://github.com/IDEA-Research/Grounded-Segment-Anything/issues/53
+# and https://github.com/IDEA-Research/Grounded-Segment-Anything/issues/84
+# when running in Docker
+# Check if nvcc is installed
+NVCC := $(shell which nvcc)
+ifeq ($(NVCC),)
+	# NVCC not found
+	USE_CUDA := 0
+	NVCC_VERSION := "not installed"
+else
+	NVCC_VERSION := $(shell nvcc --version | grep -oP 'release \K[0-9.]+')
+	USE_CUDA := $(shell echo "$(NVCC_VERSION) > 11" | bc -l)
+endif
+
+# Add the list of supported ARCHs
+ifeq ($(USE_CUDA), 1)
+	TORCH_CUDA_ARCH_LIST := "7.0;7.5;8.0;8.6+PTX"
+	BUILD_MESSAGE := "I will try to build the image with CUDA support"
+else
+	TORCH_CUDA_ARCH_LIST :=
+	BUILD_MESSAGE := "CUDA $(NVCC_VERSION) is not supported"
+endif
+
+
+prepare-checkpoints:
+	@echo "Preparing checkpoints..."
+	cd checkpoints && \
+	bash download_ckpts.sh && \
+	mkdir -p gdino_checkpoints && \
+	cd gdino_checkpoints && \
+	wget https://github.com/IDEA-Research/GroundingDINO/releases/download/v0.1.0-alpha/groundingdino_swint_ogc.pth && \
+	wget https://github.com/IDEA-Research/GroundingDINO/releases/download/v0.1.0-alpha2/groundingdino_swinb_cogcoor.pth
+
+
+build-image:
+	@echo $(BUILD_MESSAGE)
+	docker build --build-arg USE_CUDA=$(USE_CUDA) \
+	--build-arg TORCH_ARCH=$(TORCH_CUDA_ARCH_LIST) \
+	-t grounded_sam2:1.0 .
+run:
+	docker run --gpus all -it --rm --net=host --privileged \
+	-v /tmp/.X11-unix:/tmp/.X11-unix \
+	-v "${PWD}":/home/appuser/Grounded-SAM-2 \
+	-e DISPLAY=$DISPLAY \
+	--name=gsa \
+	--ipc=host -it grounded_sam2:1.0

gdino_checkpoints/download_ckpts.sh

@@ -0,0 +1,24 @@
+#!/bin/bash
+
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+
+# Define the URLs for the checkpoints
+BASE_URL="https://github.com/IDEA-Research/GroundingDINO/releases/download/"
+swint_ogc_url="${BASE_URL}v0.1.0-alpha/groundingdino_swint_ogc.pth"
+swinb_cogcoor_url="${BASE_URL}v0.1.0-alpha2/groundingdino_swinb_cogcoor.pth"
+
+
+
+# Download each of the four checkpoints using wget
+echo "Downloading groundingdino_swint_ogc.pth checkpoint..."
+wget $swint_ogc_url || { echo "Failed to download checkpoint from $swint_ogc_url"; exit 1; }
+
+echo "Downloading groundingdino_swinb_cogcoor.pth checkpoint..."
+wget $swinb_cogcoor_url || { echo "Failed to download checkpoint from $swinb_cogcoor_url"; exit 1; }
+
+echo "All checkpoints are downloaded successfully."

grounding_dino/setup.py

@@ -92,6 +92,10 @@ def get_extensions():
             "-D__CUDA_NO_HALF_OPERATORS__",
             "-D__CUDA_NO_HALF_CONVERSIONS__",
             "-D__CUDA_NO_HALF2_OPERATORS__",
+            "-gencode=arch=compute_70,code=sm_70",
+            "-gencode=arch=compute_75,code=sm_75",
+            "-gencode=arch=compute_80,code=sm_80",
+            "-gencode=arch=compute_86,code=sm_86",
         ]
     else:
         print("Compiling without CUDA")
@@ -208,7 +212,7 @@ if __name__ == "__main__":
         url="https://github.com/IDEA-Research/GroundingDINO",
         description="open-set object detector",
         license=license,
-        install_requires=parse_requirements("requirements.txt"),
+        # install_requires=parse_requirements("requirements.txt"),
         packages=find_packages(
             exclude=(
                 "configs",

setup.py

@@ -25,11 +25,28 @@ with open("README.md", "r") as f:
 REQUIRED_PACKAGES = [
     "torch>=2.3.1", 
     "torchvision>=0.18.1",
+    "transformers",  
     "numpy>=1.24.4",
     "tqdm>=4.66.1",
     "hydra-core>=1.3.2",
     "iopath>=0.1.10",
     "pillow>=9.4.0",  
+    "huggingface_hub",  
+    "diffusers[torch]==0.15.1",  
+    "onnxruntime==1.14.1",  
+    "onnx==1.13.1",  
+    "ipykernel==6.16.2",  
+    "scipy",  
+    "gradio",  
+    "openai",  
+    "matplotlib>=3.9.1",  
+    "opencv-python>=4.7.0",  
+    "dds_cloudapi_sdk",  
+    "addict",  
+    "yapf",  
+    "timm",  
+    "supervision>=0.22.0",  
+    "pycocotools",  
 ]
 
 EXTRA_PACKAGES = {