Migrate from Conda to uv

README.md

@@ -59,49 +59,42 @@ This breakthrough is driven by an innovative data engine that has automatically
 
 ### Prerequisites
 
+- [uv](https://github.com/astral-sh/uv) installed
 - Python 3.12 or higher
-- PyTorch 2.7 or higher
 - CUDA-compatible GPU with CUDA 12.6 or higher
 
-1. **Create a new Conda environment:**
+1. **Clone the repository:**
 
 ```bash
-conda create -n sam3 python=3.12
-conda deactivate
-conda activate sam3
+git clone https://github.com/facebookresearch/sam3.git
+cd sam3
 ```
 
-2. **Install PyTorch with CUDA support:**
+2. **Sync the environment:**
 
-```bash
-pip install torch==2.7.0 torchvision torchaudio --index-url https://download.pytorch.org/whl/cu126
-```
-
-3. **Clone the repository and install the package:**
+This will create a virtual environment, install Python 3.12, PyTorch 2.7.0 (with CUDA 12.6), and all project dependencies.
 
 ```bash
-git clone https://github.com/facebookresearch/sam3.git
-cd sam3
-pip install -e .
+uv sync --extra notebooks --extra train --extra dev
 ```
 
-4. **Install additional dependencies for example notebooks or development:**
+3. **Verify installation:**
 
 ```bash
-# For running example notebooks
-pip install -e ".[notebooks]"
-
-# For development
-pip install -e ".[train,dev]"
+uv run python -c "import torch; import sam3; print('Installation successful!')"
 ```
 
 ## Getting Started
 
 ⚠️ Before using SAM 3, please request access to the checkpoints on the SAM 3
 Hugging Face [repo](https://huggingface.co/facebook/sam3). Once accepted, you
-need to be authenticated to download the checkpoints. You can do this by running
-the following [steps](https://huggingface.co/docs/huggingface_hub/en/quick-start#authentication)
-(e.g. `hf auth login` after generating an access token.)
+need to be authenticated to download the checkpoints. You can do this by running:
+
+```bash
+uv run huggingface-cli login
+```
+
+and pasting your access token from [your settings](https://huggingface.co/settings/tokens).
 
 ### Basic Usage
 
@@ -174,11 +167,8 @@ dataset.
 To run the Jupyter notebook examples:
 
 ```bash
-# Make sure you have the notebooks dependencies installed
-pip install -e ".[notebooks]"
-
-# Start Jupyter notebook
-jupyter notebook examples/sam3_image_predictor_example.ipynb
+# Start Jupyter notebook (notebooks dependencies are installed via uv sync)
+uv run jupyter notebook examples/sam3_image_predictor_example.ipynb
 ```
 
 ## Model
@@ -347,13 +337,13 @@ We release 2 image benchmarks, [SA-Co/Gold](scripts/eval/gold/README.md) and
 To set up the development environment:
 
 ```bash
-pip install -e ".[dev,train]"
+uv sync --extra dev --extra train
 ```
 
 To format the code:
 
 ```bash
-ufmt format .
+uv run ufmt format .
 ```
 
 ## Contributing

README_TRAIN.md

@@ -7,7 +7,7 @@ This repository supports finetuning SAM3 models on custom datasets in multi-node
 
 ```bash
 cd sam3
-pip install -e ".[train]"
+uv sync --extra train
 ```
 
 ### Training Script Usage

examples/sam3_image_predictor_example.ipynb

@@ -108,18 +108,15 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": null,
    "metadata": {
     "output": {
      "id": 785101161160169,
      "loadingStatus": "loaded"
     }
    },
    "outputs": [],
-   "source": [
-    "bpe_path = f\"{sam3_root}/assets/bpe_simple_vocab_16e6.txt.gz\"\n",
-    "model = build_sam3_image_model(bpe_path=bpe_path)"
-   ]
+   "source": "model = build_sam3_image_model()"
   },
   {
    "cell_type": "code",
@@ -371,4 +368,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 4
-}
+}

examples/sample_run.py

@@ -0,0 +1,59 @@
+
+import os
+import torch
+import numpy as np
+from PIL import Image
+from sam3.model_builder import build_sam3_image_model
+from sam3.model.sam3_image_processor import Sam3Processor
+
+def run_sample():
+    print("Setting up SAM3...")
+
+    # Check if assets exist
+    # Updated to look one directory up for assets since this is in examples/
+    image_path = os.path.join(os.path.dirname(os.path.dirname(os.path.abspath(__file__))), "assets/images/test_image.jpg")
+
+    if not os.path.exists(image_path):
+        print(f"Warning: {image_path} not found. Please ensure the assets directory exists.")
+        return
+
+    try:
+        # Load the model
+        print("Loading SAM3 model (this may trigger checkpoint download)...")
+        model = build_sam3_image_model()
+        # Lower confidence threshold from default 0.5 to detect more objects
+        processor = Sam3Processor(model, confidence_threshold=0.1)
+
+        # Load an image
+        print(f"Loading image from {image_path}...")
+        image = Image.open(image_path)
+
+        # Run inference
+        print("Running inference...")
+        inference_state = processor.set_image(image)
+
+        # Prompt the model
+        prompt_text = "kid wearing a red bib" 
+        print(f"Prompting with: '{prompt_text}'")
+        output = processor.set_text_prompt(state=inference_state, prompt=prompt_text)
+
+        # Get results
+        masks = output["masks"]
+        boxes = output["boxes"]
+        scores = output["scores"]
+
+        print("\nSuccess!")
+        print(f"Found {len(masks)} masks.")
+        print(f"Scores: {scores[:5]} (showing top 5)")
+        print(f"Masks shape: {masks.shape}")
+
+    except Exception as e:
+        print("\nError occurred:")
+        print(str(e))
+        if "huggingface" in str(e).lower() or "unauthorized" in str(e).lower() or "login" in str(e).lower():
+            print("\nIt looks like an authentication error.")
+            print("Please run: uv run huggingface-cli login")
+            print("And paste your token from https://huggingface.co/settings/tokens")
+
+if __name__ == "__main__":
+    run_sample()

pyproject.toml

@@ -7,7 +7,7 @@ name = "sam3"
 dynamic = ["version"]
 description = "SAM3 (Segment Anything Model 3) implementation"
 readme = "README.md"
-requires-python = ">=3.8"
+requires-python = ">=3.9"
 license = {file = "LICENSE"}
 authors = [
     {name = "Meta AI Research"}
@@ -33,6 +33,9 @@ dependencies = [
     "iopath>=0.1.10",
     "typing_extensions",
     "huggingface_hub",
+    "torch==2.7.0",
+    "torchvision>=0.22.0",
+    "torchaudio>=2.7.0",
 ]
 
 [project.optional-dependencies]
@@ -133,3 +136,6 @@ testpaths = ["tests"]
 python_files = "test_*.py"
 python_classes = "Test*"
 python_functions = "test_*"
+
+[[tool.uv.index]]
+url = "https://download.pytorch.org/whl/cu126"