Commit History

Fix: ensure_on_device before each inference to prevent text_encoder drift
d95102e
verified

samwell commited on

Explicitly move all pipeline components to CUDA, use inference_mode
850ecac
verified

samwell commited on

Remove generator entirely to fix CUDA/CPU device mismatch
0c995b9
verified

samwell commited on

Fix generator - remove device spec, use official example pattern
0e4048f
verified

samwell commited on

Fix CUDA device mismatch - resize image and add autocast
2b2bf3c
verified

samwell commited on

Update README.md
aed4085
verified

samwell commited on

Use public pytorch image, install cosmos-guardrail with --no-deps
1c3a936
verified

samwell commited on

Add FastAPI handler
4233b79
verified

samwell commited on

Add Dockerfile using NVIDIA cosmos container
331d022
verified

samwell commited on

initial commit
20462f9
verified

samwell commited on