Initial snapshot: hand motion pipeline (Dyn-HaMR + dex-retargeting + SPIDER)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-09-15 14:13:49 +08:00
commit 7e4ef6f98b
7002 changed files with 1452249 additions and 0 deletions
@@ -0,0 +1,94 @@
+#!/usr/bin/env python3
+"""
+Convert MP4 videos to optimized GIFs for GitHub README display.
+This script creates smaller, optimized GIFs suitable for web display.
+"""
+import os
+import sys
+import imageio
+from pathlib import Path
+
+def convert_mp4_to_gif(mp4_path, gif_path, max_frames=150, fps=10, scale=0.5):
+    """
+    Convert MP4 to optimized GIF.
+    
+    Args:
+        mp4_path: Path to input MP4 file
+        gif_path: Path to output GIF file
+        max_frames: Maximum number of frames to include (for file size)
+        fps: Frames per second for output GIF
+        scale: Scale factor (0.5 = half size)
+    """
+    print(f"Converting {mp4_path} to {gif_path}...")
+    
+    try:
+        # Read video
+        reader = imageio.get_reader(mp4_path)
+        fps_original = reader.get_meta_data()['fps']
+        
+        # Calculate frame skip to get desired fps
+        frame_skip = max(1, int(fps_original / fps))
+        
+        frames = []
+        frame_count = 0
+        
+        for i, frame in enumerate(reader):
+            if i % frame_skip == 0 and frame_count < max_frames:
+                # Resize frame
+                import numpy as np
+                from PIL import Image
+                img = Image.fromarray(frame)
+                new_size = (int(img.width * scale), int(img.height * scale))
+                img_resized = img.resize(new_size, Image.Resampling.LANCZOS)
+                frames.append(np.array(img_resized))
+                frame_count += 1
+        
+        reader.close()
+        
+        # Write GIF
+        imageio.mimsave(gif_path, frames, fps=fps, loop=0)
+        print(f"  ✓ Created {gif_path} ({frame_count} frames, {fps} fps)")
+        
+        # Get file sizes
+        mp4_size = os.path.getsize(mp4_path) / (1024 * 1024)  # MB
+        gif_size = os.path.getsize(gif_path) / (1024 * 1024)  # MB
+        print(f"  Size: {mp4_size:.1f} MB → {gif_size:.1f} MB")
+        
+    except Exception as e:
+        print(f"  ✗ Error: {e}")
+        return False
+    
+    return True
+
+def main():
+    assets_dir = Path(__file__).parent.parent / "assets"
+    
+    videos = [
+        "droid_result.mp4",
+        "vipe_result.mp4",
+        "handedness1.mp4",
+        "handedness2.mp4",
+        "handedness3.mp4",
+        "handedness4.mp4",
+    ]
+    
+    print("Converting videos to GIFs for GitHub README...")
+    print("=" * 60)
+    
+    for video in videos:
+        mp4_path = assets_dir / video
+        gif_path = assets_dir / video.replace(".mp4", ".gif")
+        
+        if not mp4_path.exists():
+            print(f"⚠ Skipping {video} (not found)")
+            continue
+        
+        convert_mp4_to_gif(mp4_path, gif_path, max_frames=1000, fps=20, scale=0.6)
+        print()
+    
+    print("=" * 60)
+    print("Done! Update README.md to use .gif files instead of .mp4")
+
+if __name__ == "__main__":
+    main()
+
@@ -0,0 +1,34 @@
+#!/usr/bin/env bash
+set -e
+
+export CONDA_ENV_NAME=dynhamr
+
+conda create -n $CONDA_ENV_NAME python=3.10 -y
+
+conda activate $CONDA_ENV_NAME
+
+# install pytorch using pip, update with appropriate cuda drivers if necessary
+pip install torch==1.13.0 torchvision==0.14.0 --index-url https://download.pytorch.org/whl/cu117
+# uncomment if pip installation isn't working
+# conda install pytorch=1.13.0 torchvision=0.14.0 pytorch-cuda=11.7 -c pytorch -c nvidia -y
+
+# install pytorch scatter using pip, update with appropriate cuda drivers if necessary
+pip install torch-scatter -f https://data.pyg.org/whl/torch-1.13.0+cu117.html
+# uncomment if pip installation isn't working
+# conda install pytorch-scatter -c pyg -y
+
+# install remaining requirements
+pip install -r requirements.txt
+
+# install source
+pip install -e .
+
+# install DROID-SLAM/DPVO
+cd third-party/DROID-SLAM
+python setup.py install
+cd ../..
+
+# install HaMeR
+cd third-party/HaMeR
+pip install -e .[all]
+pip install -v -e third-party/ViTPose
@@ -0,0 +1,30 @@
+#!/usr/bin/env bash
+set -e
+
+echo "Creating virtual environment"
+python3.10 -m venv .dynhamr
+echo "Activating virtual environment"
+
+source $PWD/.dynhamr/bin/activate
+
+# install pytorch
+$PWD/.dynhamr/bin/pip install torch==1.13.0 torchvision==0.14.0 --index-url https://download.pytorch.org/whl/cu117
+
+# torch-scatter
+$PWD/.dynhamr/bin/pip install torch-scatter -f https://data.pyg.org/whl/torch-1.13.0+cu117.html
+
+# install source
+$PWD/.dynhamr/bin/pip install -e .
+
+# install remaining requirements
+$PWD/.dynhamr/bin/pip install -r requirements.txt
+
+# install DROID-SLAM/DPVO
+cd third-party/DROID-SLAM
+python setup.py install
+cd ../..
+
+# install HaMeR
+cd third-party/HaMeR
+pip install -e .[all]
+pip install -v -e third-party/ViTPose
@@ -0,0 +1,21 @@
+#!/usr/bin/env bash
+
+
+# Download for HaMeR
+gdown https://drive.google.com/uc?id=1mv7CUAnm73oKsEEG1xE3xH2C_oqcFSzT
+# Alternatively, you can use wget
+#wget https://www.cs.utexas.edu/~pavlakos/hamer/data/hamer_demo_data.tar.gz
+tar --warning=no-unknown-keyword --exclude=".*" -xvf hamer_demo_data.tar.gz
+rm -rf hamer_demo_data.tar.gz
+
+# Download for SLAM
+gdown https://drive.google.com/uc?id=1VD1vGhl_NPzy8mza4Fx6vvqFpnlzZ86L
+mv droid.pth ./_DATA/
+
+# Download for HMP
+gdown https://drive.google.com/uc?id=1_rp_SMxZA6Gjm6Z49M6hXam9kjMD6-GQ
+unzip hmp_model.zip
+mv hmp_model ./_DATA/
+rm -rf hmp_model.zip
+
+wget https://huggingface.co/spaces/rolpotamias/WiLoR/resolve/main/pretrained_models/detector.pt -P ./third-party/hamer/pretrained_models/