Upload 3 files

HighResolutionNet_int.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:91f7b7d3405ada242836477f1f528fab245d66ad0a7c8a97cb51781597c6730f
-size 263764182
+oid sha256:1daf04ac2d732c753c4bd79759d6942d62b55acb4cedc30e22ecbfa7647a5c22
+size 263764409

hrnet_quantized_onnx_eval.py

@@ -127,7 +127,11 @@ def run_onnx_inference(ort_session, img):
         ndarray: Model inference result.
     """
     pre_img, pad_h, pad_w = preprocess(img)
+    # transform chw into hwc format 
+
     img = np.expand_dims(pre_img, 0)
+    img = np.transpose(img, (0,2,3,1))
+
     ort_inputs = {ort_session.get_inputs()[0].name: img}
     o1 = ort_session.run(None, ort_inputs)[0]
     h, w = o1.shape[-2:]
@@ -160,6 +164,8 @@ def testval(ort_session, root, list_path):
         image = image.numpy()[0]
         out = run_onnx_inference(ort_session, image) 
         size = label.size()
+        # for hwc output
+        out = out.transpose(0, 3, 1, 2)
         if out.shape[2] != size[1] or out.shape[3] != size[2]:
             out = torch.from_numpy(out).cpu()
             pred = F.interpolate(

hrnet_quantized_onnx_inference.py

@@ -37,6 +37,7 @@ def run_onnx_inference(ort_session, img):
     """
     pre_img, pad_h, pad_w = preprocess(img)
     img = np.expand_dims(pre_img, 0)
+    img = np.transpose(img, (0,2,3,1))
     ort_inputs = {ort_session.get_inputs()[0].name: img}
     o1 = ort_session.run(None, ort_inputs)[0]
     h, w = o1.shape[-2:]
@@ -52,6 +53,7 @@ def vis(out, image, save_path='color_.png'):
                220, 220, 0, 107, 142, 35, 152, 251, 152, 70, 130, 180, 220, 20, 60, 255, 0, 0, 0, 0, 142, 0, 0, 70,
                0, 60, 100, 0, 80, 100, 0, 0, 230, 119, 11, 32 ]
     # out = out[0]
+    out = out.transpose(0, 3, 1, 2)
     if out.shape[2] != image.shape[0] or out.shape[3] != image.shape[1]:
         out = torch.from_numpy(out).cpu()
         out = F.interpolate(