teste

app.py

@@ -207,24 +207,26 @@ loras = [
     #add new
 ]
 
-#--------------------------------------------------Model Initialization-----------------------------------------------------------------------------------------#
-from diffusers import AutoencoderKL
-
-# Definir dtype antes de usar
+# Initialize the base model
 dtype = torch.bfloat16
-
-good_vae = AutoencoderKL.from_pretrained(
-    "black-forest-labs/FLUX.1-dev",
-    subfolder="vae",
-    torch_dtype=dtype,
-    use_auth_token=True
-).to("cuda" if torch.cuda.is_available() else "cpu")
-
-
-
 device = "cuda" if torch.cuda.is_available() else "cpu"
 base_model = "black-forest-labs/FLUX.1-dev"
 
+taef1 = AutoencoderTiny.from_pretrained("madebyollin/taef1", torch_dtype=dtype).to(device)
+good_vae = AutoencoderKL.from_pretrained(base_model, subfolder="vae", torch_dtype=dtype).to(device)
+pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=dtype, vae=taef1).to(device)
+pipe_i2i = AutoPipelineForImage2Image.from_pretrained(
+    base_model,
+    vae=good_vae,
+    transformer=pipe.transformer,
+    text_encoder=pipe.text_encoder,
+    tokenizer=pipe.tokenizer,
+    text_encoder_2=pipe.text_encoder_2,
+    tokenizer_2=pipe.tokenizer_2,
+    torch_dtype=dtype
+)
+
+
 #TAEF1 is very tiny autoencoder which uses the same "latent API" as FLUX.1's VAE. FLUX.1 is useful for real-time previewing of the FLUX.1 generation process.#
 taef1 = AutoencoderTiny.from_pretrained("madebyollin/taef1", torch_dtype=dtype).to(device)
 good_vae = AutoencoderKL.from_pretrained(base_model, subfolder="vae", torch_dtype=dtype).to(device)