Remove 4-bit quantization and use regular model loading
9eddb40
Joash
commited on
Spaces:
Sleeping
Sleeping
Commit History
Fix offline mode and improve model loading
69455b9
Joash
commited on
Add comprehensive memory optimizations for model and Docker
a307172
Joash
commited on
Optimize memory usage with 4-bit quantization and Docker settings
5f0bb6b
Joash
commited on
Add detailed logging and improve error handling in model manager
1f37a6a
Joash
commited on
Fix model_manager to use actual model inference instead of mock response
93aa8dc
Joash
commited on
Fix Dockerfile permissions and user management
087ce88
Joash
commited on