Update app.py

app.py

@@ -14,8 +14,8 @@ from live_preview_helpers import calculate_shift, retrieve_timesteps, flux_pipe_
 from io import BytesIO
 import base64
 from diffusers.pipelines.flux.pipeline_flux import FluxPipeline
-# For voice transcription
-import speech_recognition as sr
+import torch.nn.functional as F
+import torch.nn as nn
 os.environ['HF_HUB_DOWNLOAD_TIMEOUT'] = '120'
 dtype = torch.bfloat16
 device = "cuda" if torch.cuda.is_available() else "cpu"
@@ -149,65 +149,56 @@ Lighting: Time of day, intensity, direction (e.g., backlighting).
     except Exception as e:
          enhanced = f"Error enhancing prompt: {str(e)}"
     return enhanced
-# --- Voice Transcription Function ---
-def transcribe_audio(audio_file):
-    r = sr.Recognizer()
-    try:
-        with sr.AudioFile(audio_file) as source:
-            audio_data = r.record(source)
-        text = r.recognize_google(audio_data)
-    except Exception as e:
-        text = f"Transcription error: {str(e)}"
-    return text
-
-# --- Gradio Interface with Enhanced UI and Voice Recognition ---
-custom_css = """
+def fake_typing():
+    for _ in range(3):
+        yield {"visible": False}, {"visible": True}
+        time.sleep(0.5)
+css = """
 #col-container {
     margin: 0 auto;
-    max-width: 600px;
-    padding: 20px;
+    max-width: 520px;
 }
-.user-msg { background: #e3f2fd; border-radius: 15px; padding: 10px; margin: 5px; }
-.bot-msg { background: #f5f5f5; border-radius: 15px; padding: 10px; margin: 5px; }
 """
 
-with gr.Blocks(css=custom_css, title="FLUX.1 [dev] Enhanced UI with Voice Recognition") as demo:
-    gr.Markdown("# FLUX.1 [dev] with Enhanced UI and Voice Recognition")
-    # Using Tabs to separate functionalities
-    with gr.Tabs():
-        with gr.Tab("Prompt Enhancement"):
-            gr.Markdown("### Step 1: Enhance Your Prompt")
-            original_prompt = gr.Textbox(label="Original Prompt", placeholder="Enter your creative idea here...", lines=3)
-            enhance_button = gr.Button("Enhance Prompt", variant="secondary")
-            enhanced_prompt = gr.Textbox(label="Enhanced Prompt (Editable)", placeholder="Enhanced prompt appears here...", lines=3)
-            enhance_button.click(enhance_prompt, original_prompt, enhanced_prompt)
-
-        with gr.Tab("Voice Recognition"):
-            gr.Markdown("### Step 1A: Record Your Prompt")
-            audio_input = gr.Audio(source=["microphone"], type="filepath", label="Record your prompt")
-            transcribe_button = gr.Button("Transcribe Audio", variant="secondary")
-            voice_text = gr.Textbox(label="Transcribed Prompt", placeholder="Your spoken prompt will appear here...", lines=3)
-            transcribe_button.click(transcribe_audio, audio_input, voice_text)
-
-        with gr.Tab("Generate Image"):
-            gr.Markdown("### Step 2: Generate Image")
-            with gr.Row():
-                run_button = gr.Button("Generate Image", variant="primary")
-                clear_history_button = gr.Button("Clear History", variant="secondary")
-            result = gr.Image(label="Generated Image", show_label=False)
-            with gr.Accordion("Advanced Settings", open=False):
-                seed = gr.Slider(0, MAX_SEED, value=0, label="Seed", info="Seed for reproducibility")
-                randomize_seed = gr.Checkbox(True, label="Randomize Seed")
-                with gr.Row():
-                    width = gr.Slider(256, MAX_IMAGE_SIZE, 1024, step=32, label="Width")
-                    height = gr.Slider(256, MAX_IMAGE_SIZE, 1024, step=32, label="Height")
-                with gr.Row():
-                    guidance_scale = gr.Slider(1, 15, 3.5, step=0.1, label="Guidance Scale")
-                    num_inference_steps = gr.Slider(1, 50, 28, step=1, label="Inference Steps")
-            with gr.Accordion("Generation History", open=False):
-                history_display = gr.HTML("<p style='margin: 20px;'>No generations yet</p>")
-            # State to track generation history
-            history_state = gr.State([])
+with gr.Blocks(css="""
+.user-msg { background: #e3f2fd; border-radius: 15px; padding: 10px; margin: 5px; }
+.bot-msg { background: #f5f5f5; border-radius: 15px; padding: 10px; margin: 5px; }
+""") as demo:
+    chatbot = gr.Chatbot(elem_classes=["user-msg", "bot-msg"])
+    history_state = gr.State([])
+    with gr.Column(elem_id="col-container"):
+         gr.Markdown("# FLUX.1 [dev] with History Tracking")
+         gr.Markdown("### Step 1: Enhance Your Prompt")
+         original_prompt = gr.Textbox(label="Original Prompt", lines=2)
+         enhance_button = gr.Button("Enhance Prompt")
+         enhanced_prompt = gr.Textbox(label="Enhanced Prompt (Editable)", lines=2)
+         enhance_button.click(enhance_prompt, original_prompt, enhanced_prompt)
+         gr.Markdown("### Step 2: Generate Image")
+         with gr.Row():
+              run_button = gr.Button("Generate Image", variant="primary")
+         result = gr.Image(label="Result", show_label=False)
+         with gr.Accordion("Advanced Settings"):
+              seed = gr.Slider(0, MAX_SEED, value=0, label="Seed")
+              randomize_seed = gr.Checkbox(True, label="Randomize seed")
+              with gr.Row():
+                   width = gr.Slider(256, MAX_IMAGE_SIZE, 1024, step=32, label="Width")
+                   height = gr.Slider(256, MAX_IMAGE_SIZE, 1024, step=32, label="Height")
+              with gr.Row():
+                   guidance_scale = gr.Slider(1, 15, 3.5, step=0.1, label="Guidance Scale")
+                   num_inference_steps = gr.Slider(1, 50, 28, step=1, label="Inference Steps")
+         with gr.Accordion("Generation History", open=False):
+              history_display = gr.HTML("<p style='margin: 20px;'>No generations yet</p>")
+         gr.Examples(
+              examples=[
+                   "a tiny astronaut hatching from an egg on the moon",
+                   "a cat holding a sign that says hello world",
+                   "an anime illustration of a wiener schnitzel",
+              ],
+              inputs=enhanced_prompt,
+              outputs=[result, seed],
+              fn=infer,
+              cache_examples="lazy"
+         )
     generation_event = run_button.click(
          fn=infer,
          inputs=[enhanced_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps],