Update app.py

app.py

@@ -12,6 +12,71 @@ from qwenimage.qwen_fa3_processor import QwenDoubleStreamAttnProcessorFA3
 import math
 import os
 
+import os
+import spaces
+import torch
+from diffusers import AutoencoderKLWan, WanPipeline, WanImageToVideoPipeline, UniPCMultistepScheduler
+from diffusers.utils import export_to_video
+import gradio as gr
+import tempfile
+from huggingface_hub import hf_hub_download
+import numpy as np
+from PIL import Image
+import random
+
+
+HF_MODEL = os.environ.get("HF_UPLOAD_REPO", "rahul7star/qwen-edit-img-repo")
+
+# --- CPU-only upload function ---
+def upload_image_and_prompt_cpu(input_image, prompt_text) -> str:
+    from datetime import datetime
+    import tempfile, os, uuid, shutil
+    from huggingface_hub import HfApi
+
+    # Instantiate the HfApi class
+    api = HfApi()
+    print(prompt_text)
+
+    today_str = datetime.now().strftime("%Y-%m-%d")
+    unique_subfolder = f"Upload-Image-{uuid.uuid4().hex[:8]}"
+    hf_folder = f"{today_str}/{unique_subfolder}"
+
+    # Save image temporarily
+    with tempfile.NamedTemporaryFile(suffix=".png", delete=False) as tmp_img:
+        if isinstance(input_image, str):
+            shutil.copy(input_image, tmp_img.name)
+        else:
+            input_image.save(tmp_img.name, format="PNG")
+        tmp_img_path = tmp_img.name
+
+    # Upload image using HfApi instance
+    api.upload_file(
+        path_or_fileobj=tmp_img_path,
+        path_in_repo=f"{hf_folder}/input_image.png",
+        repo_id=HF_MODEL,
+        repo_type="model",
+        token=os.environ.get("HUGGINGFACE_HUB_TOKEN")
+    )
+
+    # Save prompt as summary.txt
+    summary_file = tempfile.NamedTemporaryFile(delete=False, suffix=".txt").name
+    with open(summary_file, "w", encoding="utf-8") as f:
+        f.write(prompt_text)
+
+    api.upload_file(
+        path_or_fileobj=summary_file,
+        path_in_repo=f"{hf_folder}/summary.txt",
+        repo_id=HF_MODEL,
+        repo_type="model",
+        token=os.environ.get("HUGGINGFACE_HUB_TOKEN")
+    )
+
+    # Cleanup
+    os.remove(tmp_img_path)
+    os.remove(summary_file)
+
+    return hf_folder
+
 # --- Model Loading ---
 dtype = torch.bfloat16
 device = "cuda" if torch.cuda.is_available() else "cpu"
@@ -77,6 +142,11 @@ def _append_prompt(base: str, extra: str) -> str:
 
 def generate_single_view(input_images, prompt, seed, num_inference_steps, true_guidance_scale):
     generator = torch.Generator(device=device).manual_seed(seed)
+    print(prompt)
+    try:
+        upload_image_and_prompt_cpu(input_image, prompt)
+    except Exception as e:
+        print("Upload failed:", e)
     result = pipe(
         image=input_images if input_images else None,
         prompt=prompt,