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 SD3 diffusers lora.
comfyanonymousac151ac16986
diff --git a/comfy/lora.py b/comfy/lora.pyindex 096285bba..37254b03f 100644--- a/comfy/lora.py+++ b/comfy/lora.py@@ -29,6 +29,7 @@ def load_lora(lora, to_load):regular_lora = "{}.lora_up.weight".format(x)diffusers_lora = "{}_lora.up.weight".format(x)+ diffusers2_lora = "{}.lora_B.weight".format(x)transformers_lora = "{}.lora_linear_layer.up.weight".format(x)A_name = None@@ -40,6 +41,10 @@ def load_lora(lora, to_load):A_name = diffusers_loraB_name = "{}_lora.down.weight".format(x)mid_name = None+ elif diffusers2_lora in lora.keys():+ A_name = diffusers2_lora+ B_name = "{}.lora_A.weight".format(x)+ mid_name = Noneelif transformers_lora in lora.keys():A_name = transformers_loraB_name ="{}.lora_linear_layer.down.weight".format(x)@@ -164,6 +169,7 @@ def load_lora(lora, to_load):for x in lora.keys():if x not in loaded_keys:logging.warning("lora key not loaded: {}".format(x))+return patch_dictdef model_lora_keys_clip(model, key_map={}):@@ -217,7 +223,8 @@ def model_lora_keys_clip(model, key_map={}):return key_mapdef model_lora_keys_unet(model, key_map={}):- sdk = model.state_dict().keys()+ sd = model.state_dict()+ sdk = sd.keys()for k in sdk:if k.startswith("diffusion_model.") and k.endswith(".weight"):@@ -238,4 +245,17 @@ def model_lora_keys_unet(model, key_map={}):if diffusers_lora_key.endswith(".to_out.0"):diffusers_lora_key = diffusers_lora_key[:-2]key_map[diffusers_lora_key] = unet_key++ if isinstance(model, comfy.model_base.SD3): #Diffusers lora SD3+ for i in range(model.model_config.unet_config.get("depth", 0)):+ k = "transformer.transformer_blocks.{}.attn.".format(i)+ qkv = "diffusion_model.joint_blocks.{}.x_block.attn.qkv.weight".format(i)+ proj = "diffusion_model.joint_blocks.{}.x_block.attn.proj.weight".format(i)+ if qkv in sd:+ offset = sd[qkv].shape[0] // 3+ key_map["{}to_q".format(k)] = (qkv, (0, 0, offset))+ key_map["{}to_k".format(k)] = (qkv, (0, offset, offset))+ key_map["{}to_v".format(k)] = (qkv, (0, offset * 2, offset))+ key_map["{}to_out.0".format(k)] = proj+return key_map