Further refine PyTorch verification command in Dockerfile for robust syntax parsing

Install MoGe from Git repository as specified in INSTALL.md

Add 'moge' package to cosmos-predict1.yaml dependencies

Fix PyTorch verification command in Dockerfile to avoid f-string syntax error

Temporarily disable 'set -u' around conda activate to resolve MKL unbound variable error

Fix pip FileNotFoundError by using absolute path in cosmos-predict1.yaml

Resolve Conda environment conflicts by simplifying CUDA dependencies and using stable PyTorch

Fix 'source: not found' error by using '.' in Dockerfile and start.sh

Explicitly accept conda TOS in Dockerfile

Configure Dockerfile with provided cosmos-predict1.yaml and install pip deps

Add directory listings for conda diagnosis in start.sh

Attempt to fix conda path by adding /opt/conda/bin to PATH

Robustly locate and activate conda environment in start.sh

Activate conda environment 'cosmos-predict1' in start.sh

Fix ModuleNotFoundError: Add /app/gui/api to PYTHONPATH

Ensured gui/requirements.txt is explicitly added and committed

Configure Space to run GEN3C FastAPI inference server

Updated start.sh for detailed debugging and robust file saving

Save generated video to /data/output

Added COPY . /app to Dockerfile to include project files

Added COPY . /app to Dockerfile to include project files

Added COPY . /app to Dockerfile to include project files

Final Dockerfile and start.sh for application launch

Updated start.sh for A100 large 80GB with single GPU command and offloading

Minimal Dockerfile for pre-built image, removed redundant build steps

Updated Dockerfile to pull pre-built elungky/gen3c:latest image

Merged remote changes and resolved conflicts

Merged remote changes and resolved conflicts

Initial commit for new Space - pre-built Docker image

initial commit