Mask2Former / tools /convert-pretrained-swin-model-to-d2.py
Ahsen Khaliq
add files
16aee22
raw
history blame contribute delete
856 Bytes
#!/usr/bin/env python
# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved
import pickle as pkl
import sys
import torch
"""
Usage:
# download pretrained swin model:
wget https://github.com/SwinTransformer/storage/releases/download/v1.0.0/swin_tiny_patch4_window7_224.pth
# run the conversion
./convert-pretrained-model-to-d2.py swin_tiny_patch4_window7_224.pth swin_tiny_patch4_window7_224.pkl
# Then, use swin_tiny_patch4_window7_224.pkl with the following changes in config:
MODEL:
WEIGHTS: "/path/to/swin_tiny_patch4_window7_224.pkl"
INPUT:
FORMAT: "RGB"
"""
if __name__ == "__main__":
input = sys.argv[1]
obj = torch.load(input, map_location="cpu")["model"]
res = {"model": obj, "__author__": "third_party", "matching_heuristics": True}
with open(sys.argv[2], "wb") as f:
pkl.dump(res, f)