llm-grounded-diffusion / generation.py

Commit History

Use fast schedule for per-box generation to speed up
d871568

Tony Lian commited on

Apply batching to SAM to reduce the memory cost with many objects
93de48e

Tony Lian commited on

Empty cache after each generation
0cbad80

Tony Lian commited on

Empty GPU cache at the end of run
1aae4d6

Tony Lian commited on

Add batched single object generation
61ac46b

Tony Lian commited on

Add autocast
2335a8f

Tony Lian commited on

Allow overriding the overall prompt
6007e4c

Tony Lian commited on

Allow using different schedulers and negative prompts
ec7f11c

Tony Lian commited on

Add examples, adjustable number of steps, and custom template
9668cda

Tony Lian commited on

Add stage 2
1f39cf9

Tony Lian commited on