Add further comments + fixes

demo_gr.py

@@ -28,6 +28,7 @@ class FluxGenerator:
     def __init__(self, model_name: str, device: str, offload: bool):
         self.device = torch.device(device)
         self.offload = offload
+        self.model_name = model_name
         self.is_schnell = model_name == "flux-schnell"
         self.model, self.ae, self.t5, self.clip, self.nsfw_classifier = get_models(
             model_name,
@@ -49,6 +50,7 @@ def generate_image(
         image2image_strength=0.0,
         add_sampling_metadata=True,
     ):
+        seed = int(seed)
         if seed == -1:
             seed = None
 
@@ -152,9 +154,9 @@ def generate_image(
 
             img.save(filename, format="png", quality=100)
 
-            return img, opts.seed, filename, None
+            return img, str(opts.seed), filename, None
         else:
-            return None, opts.seed, None, "Your generated image may contain NSFW content."
+            return None, str(opts.seed), None, "Your generated image may contain NSFW content."
 
 def create_demo(model_name: str, device: str = "cuda" if torch.cuda.is_available() else "cpu", offload: bool = False):
     generator = FluxGenerator(model_name, device, offload)
@@ -166,15 +168,16 @@ def create_demo(model_name: str, device: str = "cuda" if torch.cuda.is_available
         with gr.Row():
             with gr.Column():
                 prompt = gr.Textbox(label="Prompt", value="a photo of a forest with mist swirling around the tree trunks. The word \"FLUX\" is painted over it in big, red brush strokes with visible texture")
+                do_img2img = gr.Checkbox(label="Image to Image", value=False, interactive=not is_schnell)
+                init_image = gr.Image(label="Input Image", visible=False)
+                image2image_strength = gr.Slider(0.0, 1.0, 0.8, step=0.1, label="Noising strength", visible=False)
+
                 with gr.Accordion("Advanced Options", open=False):
                     width = gr.Slider(128, 8192, 1360, step=16, label="Width")
                     height = gr.Slider(128, 8192, 768, step=16, label="Height")
                     num_steps = gr.Slider(1, 50, 4 if is_schnell else 50, step=1, label="Number of steps")
                     guidance = gr.Slider(1.0, 10.0, 3.5, step=0.1, label="Guidance", interactive=not is_schnell)
-                    seed = gr.Number(-1, label="Seed (-1 for random)", precision=0)
-                    do_img2img = gr.Checkbox(label="Image to Image", value=False, interactive=not is_schnell)
-                    init_image = gr.Image(label="Input Image", visible=False)
-                    image2image_strength = gr.Slider(0.0, 1.0, 0.8, step=0.1, label="Noising strength", visible=False)
+                    seed = gr.Textbox(-1, label="Seed (-1 for random)")
                     add_sampling_metadata = gr.Checkbox(label="Add sampling parameters to metadata?", value=True)
 
                 generate_btn = gr.Button("Generate")

pyproject.toml

@@ -28,8 +28,12 @@ streamlit = [
   "streamlit",
   "streamlit-keyup",
 ]
+gradio = [
+  "gradio",
+]
 all = [
   "flux[streamlit]",
+  "flux[gradio]",
 ]
 
 [project.scripts]