Add jelly blob animation video generation setup

Dockerfile.opensora

@@ -0,0 +1,37 @@
+# Use official Python runtime as base image
+FROM nvidia/cuda:12.1.1-runtime-ubuntu22.04
+
+WORKDIR /workspace
+
+# Install system dependencies
+RUN apt-get update && apt-get install -y \
+    python3.10 \
+    python3-pip \
+    git \
+    curl \
+    wget \
+    build-essential \
+    && rm -rf /var/lib/apt/lists/*
+
+# Upgrade pip
+RUN pip3 install --upgrade pip setuptools wheel
+
+# Copy requirements
+COPY requirements.txt .
+
+# Install Python dependencies
+RUN pip3 install -r requirements.txt --no-cache-dir
+
+# Install Open-Sora
+COPY . .
+RUN pip3 install -e . --no-cache-dir
+
+# Create output directory
+RUN mkdir -p outputs
+
+# Set environment variables
+ENV CUDA_VISIBLE_DEVICES=0
+ENV PYTHONUNBUFFERED=1
+
+# Default command
+CMD ["bash"]

GENERATE_VIDEO_README.md

@@ -0,0 +1,221 @@
+# 🎬 Jelly Blob Animation - Video Generation
+
+Your Pixar-quality jelly blob animation has been prepared for video generation using Open-Sora!
+
+## 📋 What's Ready
+
+✅ **Prompt prepared**: `jelly_blob_prompt.txt` - Detailed description of the animation scene
+✅ **Video configurations**: Optimized inference configs for 256x256 and 512x512 output
+✅ **Setup guides**: Multiple installation options for Windows, WSL, and Docker
+
+## 🚀 Quick Start
+
+### 💻 **Option 1: Windows Subsystem for Linux (WSL) - Recommended**
+
+**1. Install WSL2 in PowerShell (Admin):**
+```powershell
+wsl --install -d Ubuntu
+```
+
+**2. Open Ubuntu WSL terminal and run:**
+```bash
+cd /mnt/c/Users/amink/.copilot/repos/copilot-worktrees/Open-Sora/salihcamal2-art-ideal-telegram
+bash setup_wsl.sh
+```
+
+**3. Generate video:**
+```bash
+source ~/opensora-env/bin/activate
+python scripts/inference.py \
+  --config configs/opensora/inference/16x256x256.py \
+  --prompt-path jelly_blob_prompt.txt \
+  --save-dir ./outputs/jelly_blob_video
+```
+
+---
+
+### 🐳 **Option 2: Docker**
+
+**1. Build and run Docker container:**
+```bash
+cd path/to/Open-Sora
+bash run_docker.sh
+```
+
+**Requirements:**
+- Docker installed
+- Docker GPU support (`nvidia-docker`)
+- 30GB disk space
+
+---
+
+### ☁️ **Option 3: Google Colab (Free GPU)**
+
+Create a new Colab notebook and run:
+```python
+# Clone repo
+!git clone https://github.com/hpcaitech/Open-Sora.git
+%cd Open-Sora
+
+# Install dependencies
+!pip install -e . -q
+
+# Download pretrained models
+!wget https://huggingface.co/hpcaitech/Open-Sora/resolve/main/model_weights/opensora.pt -O pretrained_models/opensora.pt
+
+# Create prompt file
+!echo "YOUR_PROMPT_HERE" > prompt.txt
+
+# Generate video
+!python scripts/inference.py --config configs/opensora/inference/16x256x256.py --prompt-path prompt.txt
+```
+
+---
+
+### 🌐 **Option 4: Use Online Service**
+
+Visit: https://huggingface.co/spaces/hpcaitech/Open-Sora
+- No installation required
+- Free GPU access
+- Simple web interface
+
+---
+
+## 📝 Your Animation Prompt
+
+The prompt prepared for your jelly blob animation:
+
+```
+3D animated children's cartoon, Pixar-quality render, glossy squishy gelatin surfaces 
+with soft inner glow, extremely simple rounded shapes designed for toddlers, vibrant 
+saturated candy colors, bright cheerful lighting, smooth clean composition. 
+
+WOBBO: bright yellow translucent jelly blob kid, round teardrop shape, big glossy eyes, 
+tiny stubby arms, bouncy soft body with inner glow, extremely cute. 
+
+DOT: tiny purple grape-size baby jelly blob, smallest family member, shiny purple with 
+sparkle particles, giggly. 
+
+Fast playful pacing, DOT giggles 'Boo!' in center frame, then pops and vanishes in a 
+tiny purple sparkle burst with star particles. WOBBO gasps, his whole jelly body wobbles 
+backward in comic shock, eyes growing huge to 150% size. Soft bounce squash and stretch 
+animation. Fast playful surprise mood. House location. Bright cheerful toddler-friendly 
+composition.
+```
+
+---
+
+## ⚙️ Video Generation Parameters
+
+Default configuration (16x256x256.py):
+- **Duration**: 16 frames
+- **Resolution**: 256×256 pixels
+- **FPS**: 8 fps (adjustable)
+- **Quality**: 2 seconds video
+- **Time**: ~2-3 minutes generation
+
+For higher quality (64x512x512.py):
+- **Resolution**: 512×512 pixels  
+- **Duration**: 2-4 seconds
+- **Time**: ~5-10 minutes
+- **GPU**: Requires high-end GPU (RTX 3090, A100)
+
+---
+
+## 📊 Hardware Requirements
+
+| Component | Minimum | Recommended | Ideal |
+|-----------|---------|-------------|-------|
+| GPU | RTX 2080 | RTX 3090 | RTX 4090/A100 |
+| VRAM | 8GB | 24GB | 40GB+ |
+| System RAM | 16GB | 32GB | 64GB |
+| Disk Space | 50GB | 100GB | 200GB |
+| Generation Time | 10-15 min | 2-5 min | <2 min |
+
+---
+
+## 📁 Output Files
+
+After generation, you'll find:
+
+```
+outputs/jelly_blob_video/
+├── sample_0.mp4        # Main video file
+├── sample_0.gif        # Preview GIF
+├── sample_0_frames/    # Individual frames (optional)
+└── sample_0_info.txt   # Generation metadata
+```
+
+---
+
+## 🔧 Custom Generation Commands
+
+**Higher resolution (512×512):**
+```bash
+python scripts/inference.py \
+  --config configs/opensora/inference/64x512x512.py \
+  --prompt-path jelly_blob_prompt.txt
+```
+
+**Longer video (more frames):**
+```bash
+python scripts/inference.py \
+  --config configs/opensora/inference/16x256x256.py \
+  --prompt-path jelly_blob_prompt.txt \
+  --num-frames 24
+```
+
+**Custom settings:**
+```bash
+python scripts/inference.py \
+  --config configs/opensora/inference/16x256x256.py \
+  --prompt "your custom prompt here" \
+  --batch-size 1 \
+  --seed 42 \
+  --num-sampling-steps 50
+```
+
+---
+
+## 🐛 Troubleshooting
+
+### "ColossalAI not found"
+→ Use WSL, Docker, or Colab instead of native Windows
+
+### "CUDA out of memory"  
+→ Reduce resolution or batch size:
+```bash
+--image-size 256 256 --batch-size 1
+```
+
+### "Model weights not found"
+→ Download from HuggingFace:
+```bash
+python -c "from opensora.registry import MODELS; MODELS['STDiT-XL/2'].from_pretrained('hpcaitech/Open-Sora')"
+```
+
+### "Slow generation on CPU"
+→ Install GPU drivers and CUDA toolkit for 10x+ speedup
+
+---
+
+## 📚 Additional Resources
+
+- **Open-Sora Docs**: https://github.com/hpcaitech/Open-Sora/tree/main/docs
+- **Model Weights**: https://huggingface.co/hpcaitech/Open-Sora
+- **Community**: https://github.com/hpcaitech/Open-Sora/discussions
+
+---
+
+## ✨ Next Steps
+
+1. **Choose installation method** (WSL recommended for Windows)
+2. **Run setup script** for your chosen method
+3. **Generate video** using commands above
+4. **Enjoy your jelly blob animation!** 🎉
+
+---
+
+Generated on: 2026-07-19
+Prompt file: `jelly_blob_prompt.txt`
+Setup guide: `JELLY_BLOB_VIDEO_GENERATION.md`

JELLY_BLOB_VIDEO_GENERATION.md

@@ -0,0 +1,120 @@
+# Jelly Blob Animation Video Generation Guide
+
+## Issue
+Open-Sora requires ColossalAI, which does not support native Windows installation. You need to use one of these approaches:
+
+---
+
+## Option 1: Windows Subsystem for Linux (WSL) - Recommended
+### Setup WSL2 with Ubuntu
+```bash
+# In PowerShell (Admin):
+wsl --install -d Ubuntu
+
+# Then in WSL bash:
+cd /mnt/c/Users/amink/.copilot/repos/copilot-worktrees/Open-Sora/salihcamal2-art-ideal-telegram
+
+# Install Python dependencies
+sudo apt-get update
+sudo apt-get install -y python3.10 python3-pip cuda-toolkit
+
+# Install Open-Sora
+pip install -e . --no-build-isolation
+
+# Generate video
+python scripts/inference.py \
+  --config configs/opensora/inference/16x256x256.py \
+  --prompt-path jelly_blob_prompt.txt
+```
+
+---
+
+## Option 2: Docker Container
+### Build and run Docker container
+```bash
+# Create Dockerfile in repo root
+docker build -f Dockerfile.opensora -t opensora:latest .
+
+# Run container
+docker run --gpus all -it \
+  -v C:/Users/amink/.copilot/repos/copilot-worktrees/Open-Sora:/workspace \
+  opensora:latest \
+  python scripts/inference.py \
+    --config configs/opensora/inference/16x256x256.py \
+    --prompt-path jelly_blob_prompt.txt
+```
+
+---
+
+## Option 3: Cloud GPU Services
+Use services like:
+- **Google Colab** (free GPU access)
+- **Paperspace Gradient**
+- **Lambda Labs**
+- **AWS EC2 with GPU**
+
+Upload the repo and run the inference script.
+
+---
+
+## Option 4: Use Pre-trained Models Online
+Use existing Open-Sora services:
+- Visit: https://huggingface.co/spaces/hpcaitech/Open-Sora
+- Upload your prompt and generate video
+
+---
+
+## Video Generation Command (Once Installed)
+
+```bash
+# Basic inference
+python scripts/inference.py \
+  --config configs/opensora/inference/16x256x256.py \
+  --prompt-path jelly_blob_prompt.txt \
+  --save-dir ./outputs/jelly_blob_video
+
+# With custom parameters
+python scripts/inference.py \
+  --config configs/opensora/inference/16x256x256.py \
+  --prompt "Your prompt here" \
+  --num-frames 24 \
+  --image-size 512 512 \
+  --batch-size 1 \
+  --save-dir ./outputs/jelly_blob_video
+```
+
+---
+
+## Prompt Used for Your Video
+
+The following prompt was prepared:
+
+```
+3D animated children's cartoon, Pixar-quality render, glossy squishy gelatin surfaces with soft inner glow, extremely simple rounded shapes designed for toddlers, vibrant saturated candy colors, bright cheerful lighting, smooth clean composition. WOBBO: bright yellow translucent jelly blob kid, round teardrop shape, big glossy eyes, tiny stubby arms, bouncy soft body with inner glow, extremely cute. DOT: tiny purple grape-size baby jelly blob, smallest family member, shiny purple with sparkle particles, giggly. Fast playful pacing, DOT giggles 'Boo!' in center frame, then pops and vanishes in a tiny purple sparkle burst with star particles. WOBBO gasps, his whole jelly body wobbles backward in comic shock, eyes growing huge to 150% size. Soft bounce squash and stretch animation. Fast playful surprise mood. House location. Bright cheerful toddler-friendly composition.
+```
+
+This prompt has been saved to: `jelly_blob_prompt.txt`
+
+---
+
+## Hardware Requirements
+
+- **GPU**: NVIDIA GPU with CUDA (RTX 3090, A100, H100 recommended)
+- **RAM**: 32GB+ (preferably 64GB)
+- **Disk**: 100GB+ for models and outputs
+- **Time**: 2-5 minutes per video (varies by resolution/frames)
+
+---
+
+## Next Steps
+
+1. Choose installation method above (WSL recommended for Windows users)
+2. Follow setup instructions for your choice
+3. Run the video generation command
+4. Video will be saved to `./outputs/jelly_blob_video/`
+
+Generated videos will include:
+- MP4 format for compatibility
+- GIF preview
+- Frame quality: 256x256 to 512x512
+- Duration: 16-24 frames at 8-24 fps