Upload 3 files

README.md.txt

@@ -0,0 +1,13 @@
+---
+title: Faceless Video Generator (Free)
+emoji: 🎥
+colorFrom: indigo
+colorTo: purple
+sdk: gradio
+sdk_version: "3.39"
+app_file: app.py
+pinned: false
+---
+
+Generate AI talking avatar videos from text and image.  
+Upload an image and enter your script, then click generate.

app.py

@@ -0,0 +1,32 @@
+from huggingface_hub import snapshot_download
+import gradio as gr
+import subprocess
+import os
+import uuid
+
+def setup_models():
+    if not os.path.exists("checkpoints"):
+        print("Downloading model...")
+        snapshot_download(repo_id="deepinsight/first-order-model", local_dir="checkpoints")
+
+setup_models()
+
+def generate(text, image):
+    session = str(uuid.uuid4())[:8]
+    os.makedirs(f"results/{session}", exist_ok=True)
+    image_path = f"results/{session}/avatar.jpg"
+    image.save(image_path)
+    audio_path = f"results/{session}/audio.wav"
+    tts_cmd = f'tts --text "{text}" --out_path {audio_path}'
+    subprocess.run(tts_cmd, shell=True, check=True)
+    video_cmd = f'python checkpoints/inference.py --driven_audio {audio_path} --source_image {image_path} --result_dir results/{session}'
+    subprocess.run(video_cmd, shell=True, check=True)
+    return f"results/{session}/video.mp4"
+
+gr.Interface(
+    fn=generate,
+    inputs=[gr.Textbox(label="Script"), gr.Image(label="Avatar Image", type="pil")],
+    outputs=gr.Video(label="Generated Video"),
+    title="Faceless Video Generator (Free)",
+    description="Type your message + upload an image. Get a talking avatar video!"
+).launch()

requirements.txt

@@ -0,0 +1,6 @@
+gradio
+huggingface_hub
+tts
+torch
+opencv-python
+numpy