Update app.py

app.py

@@ -30,61 +30,42 @@ def start_async_upload(local_dir, hf_repo, output_log):
     threading.Thread(target=runner, daemon=True).start()
 
 
+from huggingface_hub import upload_folder, HfFolder
+
 async def async_upload_model(local_dir, hf_repo, output_log, max_retries=3):
-    """Uploads a local model directory to HF Hub asynchronously with detailed logging."""
+    """
+    Uploads a local model directory to HF Hub asynchronously using HTTP API.
+    """
     try:
         token = HfFolder.get_token()
-        api = HfApi()
-        api.create_repo(repo_id=hf_repo, exist_ok=True)
         output_log.append(f"[INFO] ☁️ Preparing to upload to repo: {hf_repo}")
 
-        # List files to upload
-        all_files = []
-        total_size = 0
-        for root, dirs, files in os.walk(local_dir):
-            for file in files:
-                filepath = os.path.join(root, file)
-                filesize = os.path.getsize(filepath)
-                total_size += filesize
-                rel_path = os.path.relpath(filepath, local_dir)
-                all_files.append((rel_path, filepath, filesize))
-
-        output_log.append(f"[INFO] 📂 Total {len(all_files)} files to upload, total size: {total_size / 1024:.2f} KB")
-        for f, _, size in all_files:
-            output_log.append(f"   - {f} ({size / 1024:.2f} KB)")
-
-        # Upload files individually with retries
-        with tempfile.TemporaryDirectory() as tmpdir:
-            repo = Repository(local_dir=tmpdir, clone_from=hf_repo, use_auth_token=token)
-
-            # Copy files to temp repo folder
-            for rel_path, src_path, size in all_files:
-                dst_path = os.path.join(tmpdir, rel_path)
-                os.makedirs(os.path.dirname(dst_path), exist_ok=True)
-                shutil.copy2(src_path, dst_path)
-                output_log.append(f"[INFO] Copied {rel_path} to temp repo folder ({size / 1024:.2f} KB)")
-
-            # Push to hub with retries
-            attempt = 0
-            while attempt < max_retries:
-                try:
-                    output_log.append(f"[INFO] 🔄 Attempt {attempt+1} to push to HF Hub...")
-                    repo.push_to_hub(commit_message="Upload fine-tuned model")
-                    output_log.append("[SUCCESS] ✅ Model successfully uploaded to HF Hub!")
-                    break
-                except Exception as e:
-                    attempt += 1
-                    output_log.append(f"[ERROR] Upload attempt {attempt} failed: {e}")
-                    if attempt >= max_retries:
-                        output_log.append("[ERROR] ❌ Max retries reached. Upload failed.")
-                    else:
-                        output_log.append("[INFO] Retrying upload in 5 seconds...")
-                        await asyncio.sleep(5)
+        attempt = 0
+        while attempt < max_retries:
+            try:
+                output_log.append(f"[INFO] 🔄 Attempt {attempt+1} to upload folder via HTTP API...")
+                upload_folder(
+                    folder_path=local_dir,
+                    repo_id=hf_repo,
+                    repo_type="model",
+                    token=token,
+                    ignore_patterns=["*.lock", "*.tmp"],  # ignore temp files
+                    create_pr=False,
+                )
+                output_log.append("[SUCCESS] ✅ Model successfully uploaded to HF Hub!")
+                break
+            except Exception as e:
+                attempt += 1
+                output_log.append(f"[ERROR] Upload attempt {attempt} failed: {e}")
+                if attempt >= max_retries:
+                    output_log.append("[ERROR] ❌ Max retries reached. Upload failed.")
+                else:
+                    output_log.append("[INFO] Retrying upload in 5 seconds...")
+                    await asyncio.sleep(5)
 
     except Exception as e:
         output_log.append(f"[ERROR] ❌ Unexpected error during upload: {e}")
 
-
 # ==== GPU check ====
 def check_gpu_status():
     return "🚀 Zero GPU Ready - GPU will be allocated when training starts"