From 3b7e01b1640c2be1c16bc7c35dda03a6e86bce72 Mon Sep 17 00:00:00 2001
From: shiyutang <34859558+shiyutang@users.noreply.github.com>
Date: Mon, 10 Jan 2022 17:25:48 +0800
Subject: [PATCH] add_torch2paddlepy (#5460)

---
 .../mobilenetv3_prod/Step1-5/torch2paddle.py  | 34 +++++++++++++++++++
 .../ArticleReproduction_CV.md                 |  5 ++-
 2 files changed, 36 insertions(+), 3 deletions(-)
 create mode 100644 tutorials/mobilenetv3_prod/Step1-5/torch2paddle.py

diff --git a/tutorials/mobilenetv3_prod/Step1-5/torch2paddle.py b/tutorials/mobilenetv3_prod/Step1-5/torch2paddle.py
new file mode 100644
index 00000000..2a1afe3f
--- /dev/null
+++ b/tutorials/mobilenetv3_prod/Step1-5/torch2paddle.py
@@ -0,0 +1,34 @@
+import numpy as np
+import torch
+import paddle
+
+
+def torch2paddle():
+    torch_path = "./data/mobilenet_v3_small-047dcff4.pth"
+    paddle_path = "./data/mv3_small_paddle.pdparams"
+    torch_state_dict = torch.load(torch_path)
+    fc_names = ["classifier"]
+    paddle_state_dict = {}
+    for k in torch_state_dict:
+        if "num_batches_tracked" in k:
+            continue
+        v = torch_state_dict[k].detach().cpu().numpy()
+        flag = [i in k for i in fc_names]
+        if any(flag) and "weight" in k:  # ignore bias
+            new_shape = [1, 0] + list(range(2, v.ndim))
+            print(
+                f"name: {k}, ori shape: {v.shape}, new shape: {v.transpose(new_shape).shape}"
+            )
+            v = v.transpose(new_shape)
+        k = k.replace("running_var", "_variance")
+        k = k.replace("running_mean", "_mean")
+        # if k not in model_state_dict:
+        if False:
+            print(k)
+        else:
+            paddle_state_dict[k] = v
+    paddle.save(paddle_state_dict, paddle_path)
+
+
+if __name__ == "__main__":
+    torch2paddle()
diff --git a/tutorials/tipc/train_infer_python/ArticleReproduction_CV.md b/tutorials/tipc/train_infer_python/ArticleReproduction_CV.md
index b79c2849..81c68e12 100644
--- a/tutorials/tipc/train_infer_python/ArticleReproduction_CV.md
+++ b/tutorials/tipc/train_infer_python/ArticleReproduction_CV.md
@@ -218,14 +218,13 @@ MobilnetV3网络结构的PyTorch实现: [mobilenetv3_prod/Step1-5/mobilenetv3_re
 将mobilenetv3-torch的[模型参数](https://download.pytorch.org/models/mobilenet_v3_small-047dcff4.pth)保存在本地之后，就可以通过下面的权重转换示例进行转换：
 
 ```python
-
 import numpy as np
 import torch
 import paddle
 
 def torch2paddle():
-    torch_path = "./mobilenet_v3_small-047dcff4.pth"
-    paddle_path = "./mv3_small_paddle.pdparams"
+    torch_path = "./data/mobilenet_v3_small-047dcff4.pth"
+    paddle_path = "./data/mv3_small_paddle.pdparams"
     torch_state_dict = torch.load(torch_path)
     fc_names = ["classifier"]
     paddle_state_dict = {}
-- 
GitLab