Update app.py

app.py

@@ -1,83 +1,42 @@
 import gradio as gr
-from PIL import Image
-import torch
-from diffusers import StableDiffusionPipeline
-from io import BytesIO
-import requests
-import tempfile
-import base64
+from diffusers import StableDiffusionPipeline, DiffusionPipeline
+import torch, imageio
 
-# ========================
-# Image Generation Model
-# ========================
-pipe_image = StableDiffusionPipeline.from_pretrained(
-    "stabilityai/stable-diffusion-2-1-base"
-)
-pipe_image = pipe_image.to("cpu")  # CPU friendly
+# Load Image Generator
+img_model = "runwayml/stable-diffusion-v1-5"
+img_pipe = StableDiffusionPipeline.from_pretrained(img_model, torch_dtype=torch.float32)
+img_pipe = img_pipe.to("cpu")
 
+# Load Video Generator
+vid_model = "damo-vilab/text-to-video-ms-1.7b"
+vid_pipe = DiffusionPipeline.from_pretrained(vid_model, torch_dtype=torch.float32)
+vid_pipe = vid_pipe.to("cpu")
+
+# Image generation function
 def generate_image(prompt):
-    image = pipe_image(prompt, num_inference_steps=10).images[0]
+    image = img_pipe(prompt).images[0]
     return image
 
-# ========================
-# Video Generation (Placeholder)
-# ========================
-# CPU তে ভিডিও Diffusion চালানো কঠিন, তাই placeholder
+# Video generation function
 def generate_video(prompt):
-    # ছোট লুপিং red color ভিডিও তৈরি
-    import numpy as np
-    import cv2
-    height, width = 128, 128
-    fourcc = cv2.VideoWriter_fourcc(*'mp4v')
-    temp_file = tempfile.NamedTemporaryFile(suffix=".mp4", delete=False)
-    out = cv2.VideoWriter(temp_file.name, fourcc, 5, (width, height))
-    for i in range(10):
-        frame = np.zeros((height, width, 3), dtype=np.uint8)
-        frame[:, :, 0] = int(25*i)  # Blue channel gradient
-        frame[:, :, 1] = int(255-25*i)  # Green channel gradient
-        frame[:, :, 2] = 128  # Constant Red
-        out.write(frame)
-    out.release()
-    return temp_file.name
-
-# ========================
-# Voice / Text-to-Speech
-# ========================
-# CPU compatible HuggingFace TTS model
-from TTS.api import TTS
-
-tts_model = TTS(model_name="tts_models/en/ljspeech/tacotron2-DDC", progress_bar=False)
-
-def generate_voice(text):
-    temp_file = tempfile.NamedTemporaryFile(suffix=".wav", delete=False)
-    tts_model.tts_to_file(text=text, file_path=temp_file.name)
-    return temp_file.name
+    video_frames = vid_pipe(prompt, num_frames=8).frames
+    output_path = "output.mp4"
+    imageio.mimsave(output_path, video_frames, fps=8)
+    return output_path
 
-# ========================
-# Gradio UI
-# ========================
+# Gradio Tabs UI
 with gr.Blocks() as demo:
-    gr.Markdown("## 🌊 Ultimate AI App (Image, Video, Voice)")
-
-    # Image Tab
-    with gr.Tab("Image Generation"):
-        prompt_img = gr.Textbox(label="Enter image prompt")
-        out_img = gr.Image(label="Generated Image")
-        btn_img = gr.Button("Generate Image")
-        btn_img.click(generate_image, inputs=prompt_img, outputs=out_img)
-
-    # Video Tab
-    with gr.Tab("Video Generation"):
-        prompt_vid = gr.Textbox(label="Enter video prompt")
-        out_vid = gr.Video(label="Generated Video")
-        btn_vid = gr.Button("Generate Video")
-        btn_vid.click(generate_video, inputs=prompt_vid, outputs=out_vid)
-
-    # Voice Tab
-    with gr.Tab("Voice Generation"):
-        prompt_voice = gr.Textbox(label="Enter text for voice")
-        out_voice = gr.Audio(label="Generated Voice")
-        btn_voice = gr.Button("Generate Voice")
-        btn_voice.click(generate_voice, inputs=prompt_voice, outputs=out_voice)
+    gr.Markdown("## 🎨 Free AI Image + Video Generator")
+    with gr.Tab("🖼️ Image Generator"):
+        text_in = gr.Textbox(label="ছবির জন্য প্রম্পট লিখুন")
+        img_out = gr.Image(type="pil")
+        btn1 = gr.Button("Generate Image")
+        btn1.click(generate_image, inputs=text_in, outputs=img_out)
+
+    with gr.Tab("🎬 Video Generator"):
+        text_in2 = gr.Textbox(label="ভিডিওর জন্য প্রম্পট লিখুন")
+        vid_out = gr.Video()
+        btn2 = gr.Button("Generate Video")
+        btn2.click(generate_video, inputs=text_in2, outputs=vid_out)
 
 demo.launch()