The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
Support segmind vega model.
comfyanonymousb454a67bb964
diff --git a/comfy/model_detection.py b/comfy/model_detection.pyindex c682c3e1a..e3af422a3 100644--- a/comfy/model_detection.py+++ b/comfy/model_detection.py@@ -289,7 +289,13 @@ def unet_config_from_diffusers_unet(state_dict, dtype):'channel_mult': [1, 2, 4], 'transformer_depth_middle': -1, 'use_linear_in_transformer': True, 'context_dim': 2048, 'num_head_channels': 64,'use_temporal_attention': False, 'use_temporal_resblock': False}- supported_models = [SDXL, SDXL_refiner, SD21, SD15, SD21_uncliph, SD21_unclipl, SDXL_mid_cnet, SDXL_small_cnet, SDXL_diffusers_inpaint, SSD_1B]+ Segmind_Vega = {'use_checkpoint': False, 'image_size': 32, 'out_channels': 4, 'use_spatial_transformer': True, 'legacy': False,+ 'num_classes': 'sequential', 'adm_in_channels': 2816, 'dtype': dtype, 'in_channels': 4, 'model_channels': 320,+ 'num_res_blocks': [2, 2, 2], 'transformer_depth': [0, 0, 1, 1, 2, 2], 'transformer_depth_output': [0, 0, 0, 1, 1, 1, 2, 2, 2],+ 'channel_mult': [1, 2, 4], 'transformer_depth_middle': -1, 'use_linear_in_transformer': True, 'context_dim': 2048, 'num_head_channels': 64,+ 'use_temporal_attention': False, 'use_temporal_resblock': False}++ supported_models = [SDXL, SDXL_refiner, SD21, SD15, SD21_uncliph, SD21_unclipl, SDXL_mid_cnet, SDXL_small_cnet, SDXL_diffusers_inpaint, SSD_1B, Segmind_Vega]for unet_config in supported_models:matches = Truediff --git a/comfy/supported_models.py b/comfy/supported_models.pyindex 455323b96..2f2dee871 100644--- a/comfy/supported_models.py+++ b/comfy/supported_models.py@@ -217,6 +217,16 @@ class SSD1B(SDXL):"use_temporal_attention": False,}+class Segmind_Vega(SDXL):+ unet_config = {+ "model_channels": 320,+ "use_linear_in_transformer": True,+ "transformer_depth": [0, 0, 1, 1, 2, 2],+ "context_dim": 2048,+ "adm_in_channels": 2816,+ "use_temporal_attention": False,+ }+class SVD_img2vid(supported_models_base.BASE):unet_config = {"model_channels": 320,@@ -242,5 +252,5 @@ class SVD_img2vid(supported_models_base.BASE):def clip_target(self):return None-models = [SD15, SD20, SD21UnclipL, SD21UnclipH, SDXLRefiner, SDXL, SSD1B]+models = [SD15, SD20, SD21UnclipL, SD21UnclipH, SDXLRefiner, SDXL, SSD1B, Segmind_Vega]models += [SVD_img2vid]