diff --git a/mochi_preview/__pycache__/t2v_synth_mochi.cpython-312.pyc b/mochi_preview/__pycache__/t2v_synth_mochi.cpython-312.pyc
deleted file mode 100644
index 1be05b3..0000000
Binary files a/mochi_preview/__pycache__/t2v_synth_mochi.cpython-312.pyc and /dev/null differ
diff --git a/mochi_preview/dit/joint_model/__pycache__/asymm_models_joint.cpython-312.pyc b/mochi_preview/dit/joint_model/__pycache__/asymm_models_joint.cpython-312.pyc
deleted file mode 100644
index 68ad0e8..0000000
Binary files a/mochi_preview/dit/joint_model/__pycache__/asymm_models_joint.cpython-312.pyc and /dev/null differ
diff --git a/mochi_preview/dit/joint_model/__pycache__/mod_rmsnorm.cpython-312.pyc b/mochi_preview/dit/joint_model/__pycache__/mod_rmsnorm.cpython-312.pyc
deleted file mode 100644
index 4de68dd..0000000
Binary files a/mochi_preview/dit/joint_model/__pycache__/mod_rmsnorm.cpython-312.pyc and /dev/null differ
diff --git a/mochi_preview/t2v_synth_mochi.py b/mochi_preview/t2v_synth_mochi.py
index ab92689..33cfaa7 100644
--- a/mochi_preview/t2v_synth_mochi.py
+++ b/mochi_preview/t2v_synth_mochi.py
@@ -1,5 +1,33 @@
 import json
-from typing import Dict, List, Optional
+from typing import Dict, List, Optional, Union
+
+#temporary patch to fix bug in Windows
+def patched_write_atomic(
+    path_: str,
+    content: Union[str, bytes],
+    make_dirs: bool = False,
+    encode_utf_8: bool = False,
+) -> None:
+    # Write into temporary file first to avoid conflicts between threads
+    # Avoid using a named temporary file, as those have restricted permissions
+    from pathlib import Path
+    import os
+    import shutil
+    import threading
+    assert isinstance(
+        content, (str, bytes)
+    ), "Only strings and byte arrays can be saved in the cache"
+    path = Path(path_)
+    if make_dirs:
+        path.parent.mkdir(parents=True, exist_ok=True)
+    tmp_path = path.parent / f".{os.getpid()}.{threading.get_ident()}.tmp"
+    write_mode = "w" if isinstance(content, str) else "wb"
+    with tmp_path.open(write_mode, encoding="utf-8" if encode_utf_8 else None) as f:
+        f.write(content)
+    shutil.copy2(src=tmp_path, dst=path)
+    os.remove(tmp_path)
+import torch._inductor.codecache
+torch._inductor.codecache.write_atomic = patched_write_atomic
 
 import torch
 import torch.nn.functional as F
@@ -48,6 +76,8 @@ def unnormalize_latents(
     assert z.size(1) == mean.size(0) == std.size(0)
     return z * std.to(z) + mean.to(z)
 
+
+
 def compute_packed_indices(
     N: int,
     text_mask: List[torch.Tensor],