Commit

app.py

@@ -40,7 +40,7 @@ def infer(prompt, seed=42, randomize_seed=False, aspect_ratio="4:3 landscape 115
     generator = torch.Generator().manual_seed(seed)
     
     # Load the selected LoRA weight and fuse it
-    lora_weight_path = os.path.join("lora_weights", lora_weight)
+    lora_weight_path = os.path.join("loras", lora_weight)
     # pipe.load_lora_weights(weight_path)
     # pipe.fuse_lora()
     torch.cuda.empty_cache()
@@ -52,6 +52,7 @@ def infer(prompt, seed=42, randomize_seed=False, aspect_ratio="4:3 landscape 115
         aspect_ratio,
         seed,
         num_inference_steps,
+        lora_weight,
     )
     # Generate images
     # for img in pipe.flux_pipe_call_that_returns_an_iterable_of_images(

python.py

@@ -183,12 +183,6 @@ vaedecode = NODE_CLASS_MAPPINGS["VAEDecode"]()
 saveimage = NODE_CLASS_MAPPINGS["SaveImage"]()
 
 def model_loading():
-    # loraloadermodelonly = NODE_CLASS_MAPPINGS["LoraLoaderModelOnly"]()
-    # loraloadermodelonly_72 = loraloadermodelonly.load_lora_model_only(
-    #     lora_name=lora_weight_path,
-    #     strength_model=1,
-    #     model=get_value_at_index(unetloader_12, 0),
-    # )
     model_loaders = [dualcliploader_11, vaeloader_10, unetloader_12, loraloadermodelonly_72]
     valid_models = [
     getattr(loader[0], 'patcher', loader[0]) 
@@ -206,7 +200,9 @@ def generate_image(prompt,
         aspect_ratio,
         seed,
         num_inference_steps,
+        lora_weight,
         ):
+    print(seed)
     cliptextencode = NODE_CLASS_MAPPINGS["CLIPTextEncode"]()
     cliptextencode_6 = cliptextencode.encode(
         text=prompt,
@@ -221,6 +217,14 @@ def generate_image(prompt,
         upscale_factor=1.5,
         batch_size=1,
     )
+    loraloadermodelonly = NODE_CLASS_MAPPINGS["LoraLoaderModelOnly"]()
+    loraloadermodelonly_72 = loraloadermodelonly.load_lora_model_only(
+        lora_name=lora_weight_path,
+        strength_model=1,
+        model=get_value_at_index(unetloader_12, 0),
+    )
+    randomnoise = NODE_CLASS_MAPPINGS["RandomNoise"]()
+    randomnoise_25 = randomnoise.get_noise(noise_seed='42')
     with torch.inference_mode():
         for q in range(1):
             modelsamplingflux_61 = modelsamplingflux.patch(
@@ -261,7 +265,7 @@ def generate_image(prompt,
             )
 
             saveimage_9 = saveimage.save_images(
-                filename_prefix="MarkuryFLUX", images=get_value_at_index(vaedecode_8, 0)
+                filename_prefix="image", images=get_value_at_index(vaedecode_8, 0)
             )
             saved_path = f"output/{saveimage_9['ui']['images'][0]['filename']}"
             return saved_path, seed