diff --git a/Docker/download.py b/Docker/download.py
index 234fd069..ed077660 100644
--- a/Docker/download.py
+++ b/Docker/download.py
@@ -1,5 +1,27 @@
 # Download moda ASR related models
 from modelscope import snapshot_download
-model_dir = snapshot_download('damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch',revision="v2.0.4")
-model_dir = snapshot_download('damo/speech_fsmn_vad_zh-cn-16k-common-pytorch',revision="v2.0.4")
-model_dir = snapshot_download('damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch',revision="v2.0.4")
+import shutil
+import os
+import time
+from modelscope.hub.snapshot_download import snapshot_download
+
+def safe_download(model_id, revision='v2.0.4', max_retries=3):
+    cache_base = os.path.expanduser(f'~/.cache/modelscope/hub/{model_id.replace("/", "/")}')
+    if os.path.exists(cache_base):
+        shutil.rmtree(cache_base)  # Remove possibly corrupted model cache
+
+    for attempt in range(max_retries):
+        try:
+            print(f"Attempting download for: {model_id} (Attempt {attempt + 1})")
+            model_dir = snapshot_download(model_id, revision=revision)
+            print(f"✅ Downloaded {model_id} to {model_dir}")
+            return model_dir
+        except Exception as e:
+            print(f"❌ Failed on attempt {attempt + 1} for {model_id}: {e}")
+            time.sleep(5)
+    raise RuntimeError(f"❌ All attempts failed for {model_id}")
+
+# Model downloads
+safe_download('damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch')
+safe_download('damo/speech_fsmn_vad_zh-cn-16k-common-pytorch')
+safe_download('damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch')