bug fix: when AutoEncoder is not set, process stops

app.py

@@ -274,7 +274,6 @@ def run_inference(config, config_history, pipeline, progress=gr.Progress(track_t
     
     if str(config["model"]) != 'None' and str(config["model"]) != 'null' and str(config["model"]) != '' and str(config["scheduler"]) != 'None':
 
-        print(config["model"])
         progress(1, desc="Initializing pipeline...")
         
         torch.cuda.empty_cache()
@@ -297,7 +296,7 @@ def run_inference(config, config_history, pipeline, progress=gr.Progress(track_t
         if str(config["attention_slicing"]).lower() == 'true': pipeline.enable_attention_slicing()
 
         # AUTO ENCODER
-        if str(config["auto_encoder"]).lower() != 'none' and str(config["auto_encoder"]).lower() != 'null':
+        if str(config["auto_encoder"]).lower() != 'none' and str(config["auto_encoder"]).lower() != 'null' and str(config["auto_encoder"]).lower() != '':
             pipeline.vae = AutoencoderKL.from_pretrained(config["auto_encoder"], torch_dtype=get_data_type(config["data_type"])).to(config["device"])
 
         if str(config["enable_vae_slicing"]).lower() != 'false': pipeline.enable_vae_slicing()
@@ -472,7 +471,7 @@ with gr.Blocks(analytics_enabled=False) as demo:
         gr.Markdown("**VAE** stands for Variational Auto Encoders. An 'autoencoder' is an artificial neural network that is able to encode input data and decode to output data to bascially recreate the input. The VAE whereas adds a couple of additional layers of complexity to create new and unique output.")
     with gr.Row():
         with gr.Column():
-            in_auto_encoders = gr.Dropdown(value="None", choices=list(auto_encoders.keys()), label="Auto encoder", info="leave empty to not add an auto encoder")
+            in_auto_encoders = gr.Dropdown(value="", choices=list(auto_encoders.keys()), label="Auto encoder", info="leave empty to not add an auto encoder")
             out_auto_encoder_description = gr.Textbox(value="", label="Description")
         with gr.Column():
             in_enable_vae_slicing = gr.Radio(label="Enable VAE slicing:", value=config.value["enable_vae_slicing"], choices=["True", "False"], info="decoding the batches of latents one image at a time, which may reduce memory usage, see https://huggingface.co/docs/diffusers/main/en/optimization/memory")

appConfig.json

@@ -65,7 +65,6 @@
     ],
     "refiners": ["stabilityai/stable-diffusion-xl-refiner-1.0"],
     "auto_encoders": {
-        "None": "",
         "stabilityai/sdxl-vae": "finetuned auto encoder for stable diffusion models, see https://huggingface.co/stabilityai/sdxl-vae",
         "madebyollin/sdxl-vae-fp16-fix": "stable diffusion models encoder with fp16 precision, see https://huggingface.co/madebyollin/sdxl-vae-fp16-fix",
         "stabilityai/sd-vae-ft-mse": "works best with CompVis/stable-diffusion-v1-4, see https://huggingface.co/stabilityai/sd-vae-ft-mse"

config.py

@@ -179,7 +179,7 @@ def assemble_code(str_config):
     if str(config["cpu_offload"]).lower() != 'false': code.append("pipeline.enable_model_cpu_offload()")
     
     # AUTO ENCODER
-    if str(config["auto_encoder"]).lower() != 'none': 
+    if str(config["auto_encoder"]).lower() != 'none' and str(config["auto_encoder"]).lower() != 'null' and str(config["auto_encoder"]).lower() != '':
         code.append(f'pipeline.vae = AutoencoderKL.from_pretrained("{config["auto_encoder"]}", torch_dtype=data_type).to(device)')
     
     if str(config["enable_vae_slicing"]).lower() != 'false': code.append("pipeline.enable_vae_slicing()")