Skip to content

The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.

Commits on master

convert nodes_openai.py to V3 schema (#10604)
Alexander Piskun committed 1 year ago
e617cdd
convert nodes_hypernetwork.py to V3 schema (#10583)
Alexander Piskun committed 1 year ago
1f3f7a2
fix(caching): treat bytes as hashable (#10567)
EverNebula committed 1 year ago
88df172
Clarify help text for --fast argument (#10609)
comfyanonymous committed 1 year ago
97ff9fa
Fix issue with pinned memory. (#10597)
comfyanonymous committed 1 year ago
44869ff
convert StabilityAI to use new API client (#10582)
Alexander Piskun committed 1 year ago
20182a3
Fix torch compile regression on fp8 ops. (#10580)
comfyanonymous committed 1 year ago
c58c13b
ScaleROPE now works on Lumina models. (#10578)
comfyanonymous committed 1 year ago
7f374e4
Fix rope scaling. (#10560)
comfyanonymous committed 1 year ago
27d1bd8
Add RAM Pressure cache mode (#10454)
rattus committed 1 year ago
513b0c4
fix img2img operation in Dall2 node (#10552)
Alexander Piskun committed 1 year ago
dfac946
163b629
Fix case of weights not being unpinned. (#10533)
comfyanonymous committed 1 year ago
ec4fc2a
Reduce memory usage for fp8 scaled op. (#10531)
comfyanonymous committed 1 year ago
1a58087
use new API client in Luma and Minimax nodes (#10528)
Alexander Piskun committed 1 year ago
6c14f3a
Fix issue. (#10527)
comfyanonymous committed 1 year ago
e525673
Speed up offloading using pinned memory. (#10526)
comfyanonymous committed 1 year ago
3fa7a5c
convert nodes_recraft.py to V3 schema (#10507)
Alexander Piskun committed 1 year ago
210f7a1
Mixed Precision Quantization System (#10498)
contentis committed 1 year ago
8817f8f