Add_Yolov4_onnx_model

yolov4/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

yolov4/README.md

@@ -0,0 +1,105 @@
+# YOLOv4, YOLOv4-tiny and YOLOv4x-mish ONNX Conversion
+
+## Prerequisites
+
+### 1. Model files (cfg + weights)
+
+The Darknet `.cfg` files are already provided in [opencv_extra/testdata/dnn](https://github.com/opencv/opencv_extra/tree/master/testdata/dnn) (`yolov4.cfg`, `yolov4-tiny-2020-12.cfg`, `yolov4x-mish.cfg`).
+
+Download the matching `.weights` files using the OpenCV test data download script:
+
+```bash
+git clone https://github.com/opencv/opencv_extra.git
+cd opencv_extra/testdata/dnn
+python download_models.py YOLOv4 YOLOv4-tiny-2020-12 YOLOv4x-mish
+```
+
+### 2. Python environment for `pytorch-YOLOv4`
+
+The conversion uses [`pytorch-YOLOv4`](https://github.com/Tianxiaomo/pytorch-YOLOv4). Create a Python environment with the required dependencies:
+
+Supported Python versions: **3.7 – 3.10**.
+
+```bash
+conda create -n <env_name> python=<3.7-3.10> -y
+conda activate <env_name>
+pip install "torch<2.4" "torchvision<0.19" "numpy<2" onnx onnxruntime onnxscript
+```
+
+---
+
+## Conversion of YOLOv4 to ONNX
+
+```bash
+git clone https://github.com/Tianxiaomo/pytorch-YOLOv4.git
+cd pytorch-YOLOv4
+
+# Convert (dynamic batch, batch_size=0)
+python -c "from tool.darknet2onnx import transform_to_onnx; transform_to_onnx('yolov4.cfg', 'yolov4.weights', 0)"
+```
+
+---
+
+## Conversion of YOLOv4-tiny to ONNX
+
+```bash
+git clone https://github.com/Tianxiaomo/pytorch-YOLOv4.git
+cd pytorch-YOLOv4
+
+# Convert YOLOv4-tiny (dynamic batch, batch_size=0)
+python -c "from tool.darknet2onnx import transform_to_onnx; transform_to_onnx('yolov4-tiny-2020-12.cfg', 'yolov4-tiny.weights', 0)"
+```
+
+---
+
+## Conversion of YOLOv4x-mish to ONNX
+
+### Why it differs from YOLOv4
+The `yolov4x-mish.cfg` uses `new_coords=1` — a Scaled-YOLOv4 optimization. The network is trained to output values directly in the `[0, 1]` range (no sigmoid needed) and uses a squared formula `(t_w * 2)² * anchor` for width/height instead of `exp(t_w) * anchor`. This is more numerically stable for large models.
+
+### The Issue
+The `pytorch-YOLOv4` converter was written for the original YOLOv4 (`new_coords=0`) and always applies `sigmoid` + `exp`. With `new_coords=1` weights, `sigmoid` gets applied on top of already-activated values, squishing confidences from ~0.93 down to ~0.36. As a result, the model produces garbage detections.
+
+### The Fix (Modified scripts provided)
+To resolve this, modified versions of **`darknet2pytorch.py`** and **`yolo_layer.py`** are provided in this repository. These scripts include the following patches:
+
+* **`darknet2pytorch.py`**: Properly reads the `new_coords` flag from the `.cfg`.
+* **`yolo_layer.py`**: Skips the redundant sigmoid activation for `xy/obj/cls` and implements the squared `wh` formula when `new_coords=1` is detected.
+
+### Conversion Steps
+
+```bash
+git clone https://github.com/Tianxiaomo/pytorch-YOLOv4.git
+cd pytorch-YOLOv4
+
+# [!] Replace tool/darknet2pytorch.py and tool/yolo_layer.py with the patched
+#     versions from this repository before running the conversion.
+
+# Convert YOLOv4x-mish (dynamic batch, batch_size=0)
+python -c "from tool.darknet2onnx import transform_to_onnx; transform_to_onnx('yolov4x-mish.cfg', 'yolov4x-mish.weights', 0)"
+```
+
+---
+
+## Usage
+
+A demo script is provided to run inference using OpenCV DNN:
+
+```bash
+# YOLOv4 (input size: 608x608)
+python demo.py --model yolov4.onnx --image example_outputs/input.jpg --output example_outputs/yolov4_output.jpg
+
+# YOLOv4-tiny (input size: 416x416)
+python demo.py --model yolov4-tiny.onnx --image example_outputs/input.jpg --output example_outputs/yolov4-tiny_output.jpg
+
+# YOLOv4x-mish (input size: 640x640)
+python demo.py --model yolov4x-mish.onnx --image example_outputs/input.jpg --output example_outputs/yolov4x-mish_output.jpg
+```
+
+The demo prints the detected COCO classes, confidence scores, and bounding boxes, and saves an annotated output image.
+
+---
+
+## License
+
+See [License.txt](./License.txt) — This conversion tool is based on [pytorch-YOLOv4](https://github.com/Tianxiaomo/pytorch-YOLOv4) (Apache-2.0). Original YOLOv4 model weights and configuration are released by Alexey Bochkovskiy ([AlexeyAB/darknet](https://github.com/AlexeyAB/darknet)).

yolov4/darknet2pytorch.py

@@ -0,0 +1,537 @@
+import torch.nn as nn
+import torch.nn.functional as F
+import numpy as np
+from tool.region_loss import RegionLoss
+from tool.yolo_layer import YoloLayer
+from tool.config import *
+from tool.torch_utils import *
+
+
+class Mish(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+
+    def forward(self, x):
+        x = x * (torch.tanh(torch.nn.functional.softplus(x)))
+        return x
+
+
+class MaxPoolDark(nn.Module):
+    def __init__(self, size=2, stride=1):
+        super(MaxPoolDark, self).__init__()
+        self.size = size
+        self.stride = stride
+
+    def forward(self, x):
+        '''
+        darknet output_size = (input_size + p - k) / s +1
+        p : padding = k - 1
+        k : size
+        s : stride
+        torch output_size = (input_size + 2*p -k) / s +1
+        p : padding = k//2
+        '''
+        p = self.size // 2
+        if ((x.shape[2] - 1) // self.stride) != ((x.shape[2] + 2 * p - self.size) // self.stride):
+            padding1 = (self.size - 1) // 2
+            padding2 = padding1 + 1
+        else:
+            padding1 = (self.size - 1) // 2
+            padding2 = padding1
+        if ((x.shape[3] - 1) // self.stride) != ((x.shape[3] + 2 * p - self.size) // self.stride):
+            padding3 = (self.size - 1) // 2
+            padding4 = padding3 + 1
+        else:
+            padding3 = (self.size - 1) // 2
+            padding4 = padding3
+        x = F.max_pool2d(F.pad(x, (padding3, padding4, padding1, padding2), mode='replicate'),
+                         self.size, stride=self.stride)
+        return x
+
+
+class Upsample_expand(nn.Module):
+    def __init__(self, stride=2):
+        super(Upsample_expand, self).__init__()
+        self.stride = stride
+
+    def forward(self, x):
+        assert (x.data.dim() == 4)
+        
+        x = x.view(x.size(0), x.size(1), x.size(2), 1, x.size(3), 1).\
+            expand(x.size(0), x.size(1), x.size(2), self.stride, x.size(3), self.stride).contiguous().\
+            view(x.size(0), x.size(1), x.size(2) * self.stride, x.size(3) * self.stride)
+
+        return x
+
+
+class Upsample_interpolate(nn.Module):
+    def __init__(self, stride):
+        super(Upsample_interpolate, self).__init__()
+        self.stride = stride
+
+    def forward(self, x):
+        assert (x.data.dim() == 4)
+
+        out = F.interpolate(x, size=(x.size(2) * self.stride, x.size(3) * self.stride), mode='nearest')
+        return out
+
+
+class Reorg(nn.Module):
+    def __init__(self, stride=2):
+        super(Reorg, self).__init__()
+        self.stride = stride
+
+    def forward(self, x):
+        stride = self.stride
+        assert (x.data.dim() == 4)
+        B = x.data.size(0)
+        C = x.data.size(1)
+        H = x.data.size(2)
+        W = x.data.size(3)
+        assert (H % stride == 0)
+        assert (W % stride == 0)
+        ws = stride
+        hs = stride
+        x = x.view(B, C, H / hs, hs, W / ws, ws).transpose(3, 4).contiguous()
+        x = x.view(B, C, H / hs * W / ws, hs * ws).transpose(2, 3).contiguous()
+        x = x.view(B, C, hs * ws, H / hs, W / ws).transpose(1, 2).contiguous()
+        x = x.view(B, hs * ws * C, H / hs, W / ws)
+        return x
+
+
+class GlobalAvgPool2d(nn.Module):
+    def __init__(self):
+        super(GlobalAvgPool2d, self).__init__()
+
+    def forward(self, x):
+        N = x.data.size(0)
+        C = x.data.size(1)
+        H = x.data.size(2)
+        W = x.data.size(3)
+        x = F.avg_pool2d(x, (H, W))
+        x = x.view(N, C)
+        return x
+
+
+# for route, shortcut and sam
+class EmptyModule(nn.Module):
+    def __init__(self):
+        super(EmptyModule, self).__init__()
+
+    def forward(self, x):
+        return x
+
+
+# support route shortcut and reorg
+class Darknet(nn.Module):
+    def __init__(self, cfgfile, inference=False):
+        super(Darknet, self).__init__()
+        self.inference = inference
+        self.training = not self.inference
+
+        self.blocks = parse_cfg(cfgfile)
+        self.width = int(self.blocks[0]['width'])
+        self.height = int(self.blocks[0]['height'])
+
+        self.models = self.create_network(self.blocks)  # merge conv, bn,leaky
+        self.loss = self.models[len(self.models) - 1]
+
+        if self.blocks[(len(self.blocks) - 1)]['type'] == 'region':
+            self.anchors = self.loss.anchors
+            self.num_anchors = self.loss.num_anchors
+            self.anchor_step = self.loss.anchor_step
+            self.num_classes = self.loss.num_classes
+
+        self.header = torch.IntTensor([0, 0, 0, 0])
+        self.seen = 0
+
+    def forward(self, x):
+        ind = -2
+        self.loss = None
+        outputs = dict()
+        out_boxes = []
+        for block in self.blocks:
+            ind = ind + 1
+            # if ind > 0:
+            #    return x
+
+            if block['type'] == 'net':
+                continue
+            elif block['type'] in ['convolutional', 'maxpool', 'reorg', 'upsample', 'avgpool', 'softmax', 'connected']:
+                x = self.models[ind](x)
+                outputs[ind] = x
+            elif block['type'] == 'route':
+                layers = block['layers'].split(',')
+                layers = [int(i) if int(i) > 0 else int(i) + ind for i in layers]
+                if len(layers) == 1:
+                    if 'groups' not in block.keys() or int(block['groups']) == 1:
+                        x = outputs[layers[0]]
+                        outputs[ind] = x
+                    else:
+                        groups = int(block['groups'])
+                        group_id = int(block['group_id'])
+                        _, b, _, _ = outputs[layers[0]].shape
+                        x = outputs[layers[0]][:, b // groups * group_id:b // groups * (group_id + 1)]
+                        outputs[ind] = x
+                elif len(layers) == 2:
+                    x1 = outputs[layers[0]]
+                    x2 = outputs[layers[1]]
+                    x = torch.cat((x1, x2), 1)
+                    outputs[ind] = x
+                elif len(layers) == 4:
+                    x1 = outputs[layers[0]]
+                    x2 = outputs[layers[1]]
+                    x3 = outputs[layers[2]]
+                    x4 = outputs[layers[3]]
+                    x = torch.cat((x1, x2, x3, x4), 1)
+                    outputs[ind] = x
+                else:
+                    print("rounte number > 2 ,is {}".format(len(layers)))
+
+            elif block['type'] == 'shortcut':
+                from_layer = int(block['from'])
+                activation = block['activation']
+                from_layer = from_layer if from_layer > 0 else from_layer + ind
+                x1 = outputs[from_layer]
+                x2 = outputs[ind - 1]
+                x = x1 + x2
+                if activation == 'leaky':
+                    x = F.leaky_relu(x, 0.1, inplace=True)
+                elif activation == 'relu':
+                    x = F.relu(x, inplace=True)
+                outputs[ind] = x
+            elif block['type'] == 'sam':
+                from_layer = int(block['from'])
+                from_layer = from_layer if from_layer > 0 else from_layer + ind
+                x1 = outputs[from_layer]
+                x2 = outputs[ind - 1]
+                x = x1 * x2
+                outputs[ind] = x
+            elif block['type'] == 'region':
+                continue
+                if self.loss:
+                    self.loss = self.loss + self.models[ind](x)
+                else:
+                    self.loss = self.models[ind](x)
+                outputs[ind] = None
+            elif block['type'] == 'yolo':
+                # if self.training:
+                #     pass
+                # else:
+                #     boxes = self.models[ind](x)
+                #     out_boxes.append(boxes)
+                boxes = self.models[ind](x)
+                out_boxes.append(boxes)
+            elif block['type'] == 'cost':
+                continue
+            else:
+                print('unknown type %s' % (block['type']))
+
+        if self.training:
+            return out_boxes
+        else:
+            return get_region_boxes(out_boxes)
+
+    def print_network(self):
+        print_cfg(self.blocks)
+
+    def create_network(self, blocks):
+        models = nn.ModuleList()
+
+        prev_filters = 3
+        out_filters = []
+        prev_stride = 1
+        out_strides = []
+        conv_id = 0
+        for block in blocks:
+            if block['type'] == 'net':
+                prev_filters = int(block['channels'])
+                continue
+            elif block['type'] == 'convolutional':
+                conv_id = conv_id + 1
+                batch_normalize = int(block['batch_normalize'])
+                filters = int(block['filters'])
+                kernel_size = int(block['size'])
+                stride = int(block['stride'])
+                is_pad = int(block['pad'])
+                pad = (kernel_size - 1) // 2 if is_pad else 0
+                activation = block['activation']
+                model = nn.Sequential()
+                if batch_normalize:
+                    model.add_module('conv{0}'.format(conv_id),
+                                     nn.Conv2d(prev_filters, filters, kernel_size, stride, pad, bias=False))
+                    model.add_module('bn{0}'.format(conv_id), nn.BatchNorm2d(filters))
+                    # model.add_module('bn{0}'.format(conv_id), BN2d(filters))
+                else:
+                    model.add_module('conv{0}'.format(conv_id),
+                                     nn.Conv2d(prev_filters, filters, kernel_size, stride, pad))
+                if activation == 'leaky':
+                    model.add_module('leaky{0}'.format(conv_id), nn.LeakyReLU(0.1, inplace=True))
+                elif activation == 'relu':
+                    model.add_module('relu{0}'.format(conv_id), nn.ReLU(inplace=True))
+                elif activation == 'mish':
+                    model.add_module('mish{0}'.format(conv_id), Mish())
+                elif activation == 'linear':
+                    model.add_module('linear{0}'.format(conv_id), nn.Identity())
+                elif activation == 'logistic':
+                    model.add_module('sigmoid{0}'.format(conv_id), nn.Sigmoid())
+                else:
+                    print("No convolutional activation named {}".format(activation))
+
+                prev_filters = filters
+                out_filters.append(prev_filters)
+                prev_stride = stride * prev_stride
+                out_strides.append(prev_stride)
+                models.append(model)
+            elif block['type'] == 'maxpool':
+                pool_size = int(block['size'])
+                stride = int(block['stride'])
+                if stride == 1 and pool_size % 2:
+                    # You can use Maxpooldark instead, here is convenient to convert onnx.
+                    # Example: [maxpool] size=3 stride=1
+                    model = nn.MaxPool2d(kernel_size=pool_size, stride=stride, padding=pool_size // 2)
+                elif stride == pool_size:
+                    # You can use Maxpooldark instead, here is convenient to convert onnx.
+                    # Example: [maxpool] size=2 stride=2
+                    model = nn.MaxPool2d(kernel_size=pool_size, stride=stride, padding=0)
+                else:
+                    model = MaxPoolDark(pool_size, stride)
+                out_filters.append(prev_filters)
+                prev_stride = stride * prev_stride
+                out_strides.append(prev_stride)
+                models.append(model)
+            elif block['type'] == 'avgpool':
+                model = GlobalAvgPool2d()
+                out_filters.append(prev_filters)
+                models.append(model)
+            elif block['type'] == 'softmax':
+                model = nn.Softmax()
+                out_strides.append(prev_stride)
+                out_filters.append(prev_filters)
+                models.append(model)
+            elif block['type'] == 'cost':
+                if block['_type'] == 'sse':
+                    model = nn.MSELoss(reduction='mean')
+                elif block['_type'] == 'L1':
+                    model = nn.L1Loss(reduction='mean')
+                elif block['_type'] == 'smooth':
+                    model = nn.SmoothL1Loss(reduction='mean')
+                out_filters.append(1)
+                out_strides.append(prev_stride)
+                models.append(model)
+            elif block['type'] == 'reorg':
+                stride = int(block['stride'])
+                prev_filters = stride * stride * prev_filters
+                out_filters.append(prev_filters)
+                prev_stride = prev_stride * stride
+                out_strides.append(prev_stride)
+                models.append(Reorg(stride))
+            elif block['type'] == 'upsample':
+                stride = int(block['stride'])
+                out_filters.append(prev_filters)
+                prev_stride = prev_stride // stride
+                out_strides.append(prev_stride)
+
+                models.append(Upsample_expand(stride))
+                # models.append(Upsample_interpolate(stride))
+
+            elif block['type'] == 'route':
+                layers = block['layers'].split(',')
+                ind = len(models)
+                layers = [int(i) if int(i) > 0 else int(i) + ind for i in layers]
+                if len(layers) == 1:
+                    if 'groups' not in block.keys() or int(block['groups']) == 1:
+                        prev_filters = out_filters[layers[0]]
+                        prev_stride = out_strides[layers[0]]
+                    else:
+                        prev_filters = out_filters[layers[0]] // int(block['groups'])
+                        prev_stride = out_strides[layers[0]] // int(block['groups'])
+                elif len(layers) == 2:
+                    assert (layers[0] == ind - 1 or layers[1] == ind - 1)
+                    prev_filters = out_filters[layers[0]] + out_filters[layers[1]]
+                    prev_stride = out_strides[layers[0]]
+                elif len(layers) == 4:
+                    assert (layers[0] == ind - 1)
+                    prev_filters = out_filters[layers[0]] + out_filters[layers[1]] + out_filters[layers[2]] + \
+                                   out_filters[layers[3]]
+                    prev_stride = out_strides[layers[0]]
+                else:
+                    print("route error!!!")
+
+                out_filters.append(prev_filters)
+                out_strides.append(prev_stride)
+                models.append(EmptyModule())
+            elif block['type'] == 'shortcut':
+                ind = len(models)
+                prev_filters = out_filters[ind - 1]
+                out_filters.append(prev_filters)
+                prev_stride = out_strides[ind - 1]
+                out_strides.append(prev_stride)
+                models.append(EmptyModule())
+            elif block['type'] == 'sam':
+                ind = len(models)
+                prev_filters = out_filters[ind - 1]
+                out_filters.append(prev_filters)
+                prev_stride = out_strides[ind - 1]
+                out_strides.append(prev_stride)
+                models.append(EmptyModule())
+            elif block['type'] == 'connected':
+                filters = int(block['output'])
+                if block['activation'] == 'linear':
+                    model = nn.Linear(prev_filters, filters)
+                elif block['activation'] == 'leaky':
+                    model = nn.Sequential(
+                        nn.Linear(prev_filters, filters),
+                        nn.LeakyReLU(0.1, inplace=True))
+                elif block['activation'] == 'relu':
+                    model = nn.Sequential(
+                        nn.Linear(prev_filters, filters),
+                        nn.ReLU(inplace=True))
+                prev_filters = filters
+                out_filters.append(prev_filters)
+                out_strides.append(prev_stride)
+                models.append(model)
+            elif block['type'] == 'region':
+                loss = RegionLoss()
+                anchors = block['anchors'].split(',')
+                loss.anchors = [float(i) for i in anchors]
+                loss.num_classes = int(block['classes'])
+                loss.num_anchors = int(block['num'])
+                loss.anchor_step = len(loss.anchors) // loss.num_anchors
+                loss.object_scale = float(block['object_scale'])
+                loss.noobject_scale = float(block['noobject_scale'])
+                loss.class_scale = float(block['class_scale'])
+                loss.coord_scale = float(block['coord_scale'])
+                out_filters.append(prev_filters)
+                out_strides.append(prev_stride)
+                models.append(loss)
+            elif block['type'] == 'yolo':
+                yolo_layer = YoloLayer()
+                anchors = block['anchors'].split(',')
+                anchor_mask = block['mask'].split(',')
+                yolo_layer.anchor_mask = [int(i) for i in anchor_mask]
+                yolo_layer.anchors = [float(i) for i in anchors]
+                yolo_layer.num_classes = int(block['classes'])
+                self.num_classes = yolo_layer.num_classes
+                yolo_layer.num_anchors = int(block['num'])
+                yolo_layer.anchor_step = len(yolo_layer.anchors) // yolo_layer.num_anchors
+                yolo_layer.stride = prev_stride
+                yolo_layer.scale_x_y = float(block['scale_x_y'])
+                yolo_layer.new_coords = int(block.get('new_coords', 0))
+                # yolo_layer.object_scale = float(block['object_scale'])
+                # yolo_layer.noobject_scale = float(block['noobject_scale'])
+                # yolo_layer.class_scale = float(block['class_scale'])
+                # yolo_layer.coord_scale = float(block['coord_scale'])
+                out_filters.append(prev_filters)
+                out_strides.append(prev_stride)
+                models.append(yolo_layer)
+            else:
+                print('unknown type %s' % (block['type']))
+
+        return models
+
+    def load_weights(self, weightfile):
+        fp = open(weightfile, 'rb')
+        header = np.fromfile(fp, count=5, dtype=np.int32)
+        self.header = torch.from_numpy(header)
+        self.seen = self.header[3]
+        buf = np.fromfile(fp, dtype=np.float32)
+        fp.close()
+
+        start = 0
+        ind = -2
+        for block in self.blocks:
+            if start >= buf.size:
+                break
+            ind = ind + 1
+            if block['type'] == 'net':
+                continue
+            elif block['type'] == 'convolutional':
+                model = self.models[ind]
+                batch_normalize = int(block['batch_normalize'])
+                if batch_normalize:
+                    start = load_conv_bn(buf, start, model[0], model[1])
+                else:
+                    start = load_conv(buf, start, model[0])
+            elif block['type'] == 'connected':
+                model = self.models[ind]
+                if block['activation'] != 'linear':
+                    start = load_fc(buf, start, model[0])
+                else:
+                    start = load_fc(buf, start, model)
+            elif block['type'] == 'maxpool':
+                pass
+            elif block['type'] == 'reorg':
+                pass
+            elif block['type'] == 'upsample':
+                pass
+            elif block['type'] == 'route':
+                pass
+            elif block['type'] == 'shortcut':
+                pass
+            elif block['type'] == 'sam':
+                pass
+            elif block['type'] == 'region':
+                pass
+            elif block['type'] == 'yolo':
+                pass
+            elif block['type'] == 'avgpool':
+                pass
+            elif block['type'] == 'softmax':
+                pass
+            elif block['type'] == 'cost':
+                pass
+            else:
+                print('unknown type %s' % (block['type']))
+
+    # def save_weights(self, outfile, cutoff=0):
+    #     if cutoff <= 0:
+    #         cutoff = len(self.blocks) - 1
+    #
+    #     fp = open(outfile, 'wb')
+    #     self.header[3] = self.seen
+    #     header = self.header
+    #     header.numpy().tofile(fp)
+    #
+    #     ind = -1
+    #     for blockId in range(1, cutoff + 1):
+    #         ind = ind + 1
+    #         block = self.blocks[blockId]
+    #         if block['type'] == 'convolutional':
+    #             model = self.models[ind]
+    #             batch_normalize = int(block['batch_normalize'])
+    #             if batch_normalize:
+    #                 save_conv_bn(fp, model[0], model[1])
+    #             else:
+    #                 save_conv(fp, model[0])
+    #         elif block['type'] == 'connected':
+    #             model = self.models[ind]
+    #             if block['activation'] != 'linear':
+    #                 save_fc(fc, model)
+    #             else:
+    #                 save_fc(fc, model[0])
+    #         elif block['type'] == 'maxpool':
+    #             pass
+    #         elif block['type'] == 'reorg':
+    #             pass
+    #         elif block['type'] == 'upsample':
+    #             pass
+    #         elif block['type'] == 'route':
+    #             pass
+    #         elif block['type'] == 'shortcut':
+    #             pass
+    #         elif block['type'] == 'sam':
+    #             pass
+    #         elif block['type'] == 'region':
+    #             pass
+    #         elif block['type'] == 'yolo':
+    #             pass
+    #         elif block['type'] == 'avgpool':
+    #             pass
+    #         elif block['type'] == 'softmax':
+    #             pass
+    #         elif block['type'] == 'cost':
+    #             pass
+    #         else:
+    #             print('unknown type %s' % (block['type']))
+    #     fp.close()

yolov4/demo.py

@@ -0,0 +1,164 @@
+"""
+Demo script for YOLOv4, YOLOv4-tiny and YOLOv4x-mish ONNX models using OpenCV DNN.
+
+All models share the same output format:
+  - boxes: [batch, N, 1, 4]  -> [x1, y1, x2, y2] normalized to [0, 1]
+  - confs: [batch, N, num_classes]
+
+Default input sizes (auto-detected from filename):
+  - yolov4.onnx           : 608x608
+  - yolov4-tiny.onnx      : 416x416
+  - yolov4x-mish.onnx     : 640x640
+
+Usage:
+  python demo.py --model yolov4.onnx --image example_outputs/input.jpg
+  python demo.py --model yolov4-tiny.onnx --image example_outputs/input.jpg
+  python demo.py --model yolov4x-mish.onnx --image example_outputs/input.jpg
+"""
+
+import argparse
+import cv2
+import numpy as np
+
+
+# COCO class names (80 classes) - both yolov4 and yolov4x-mish are trained on COCO
+COCO_CLASSES = [
+    "person", "bicycle", "car", "motorcycle", "airplane", "bus", "train", "truck",
+    "boat", "traffic light", "fire hydrant", "stop sign", "parking meter", "bench",
+    "bird", "cat", "dog", "horse", "sheep", "cow", "elephant", "bear", "zebra",
+    "giraffe", "backpack", "umbrella", "handbag", "tie", "suitcase", "frisbee",
+    "skis", "snowboard", "sports ball", "kite", "baseball bat", "baseball glove",
+    "skateboard", "surfboard", "tennis racket", "bottle", "wine glass", "cup",
+    "fork", "knife", "spoon", "bowl", "banana", "apple", "sandwich", "orange",
+    "broccoli", "carrot", "hot dog", "pizza", "donut", "cake", "chair", "couch",
+    "potted plant", "bed", "dining table", "toilet", "tv", "laptop", "mouse",
+    "remote", "keyboard", "cell phone", "microwave", "oven", "toaster", "sink",
+    "refrigerator", "book", "clock", "vase", "scissors", "teddy bear", "hair drier",
+    "toothbrush",
+]
+
+
+def get_model_input_size(net, default=416):
+    """Read the model's expected input size by inspecting layer shapes."""
+    # Use a probe forward to deduce input size; fall back to default if unavailable.
+    try:
+        in_shape = net.getLayer(0).blobs
+        if in_shape:
+            shape = in_shape[0].shape
+            return shape[3], shape[2]  # (W, H)
+    except Exception:
+        pass
+    return default, default
+
+
+def postprocess(outputs, conf_threshold, nms_threshold):
+    """Parse [boxes, confs] outputs and run NMS."""
+    # outs[0]: boxes [1, N, 1, 4] -> [x1, y1, x2, y2]
+    # outs[1]: confs [1, N, num_classes]
+    boxes_raw = outputs[0].reshape(-1, 4)
+    confs_raw = outputs[1].reshape(boxes_raw.shape[0], -1)
+
+    class_ids = []
+    confidences = []
+    boxes_xywh = []  # for cv2.dnn.NMSBoxes
+
+    for j in range(boxes_raw.shape[0]):
+        cls_id = int(np.argmax(confs_raw[j]))
+        score = float(confs_raw[j][cls_id])
+        if score >= conf_threshold:
+            x1, y1, x2, y2 = boxes_raw[j]
+            class_ids.append(cls_id)
+            confidences.append(score)
+            boxes_xywh.append([float(x1), float(y1), float(x2 - x1), float(y2 - y1)])
+
+    if not boxes_xywh:
+        return []
+
+    indices = cv2.dnn.NMSBoxes(boxes_xywh, confidences, conf_threshold, nms_threshold)
+    if len(indices) == 0:
+        return []
+    indices = np.array(indices).flatten()
+
+    detections = []
+    for i in indices:
+        x, y, w, h = boxes_xywh[i]
+        detections.append((class_ids[i], confidences[i], [x, y, x + w, y + h]))
+    return detections
+
+
+def draw_detections(image, detections, output_path):
+    """Draw bounding boxes and labels on the image."""
+    out = image.copy()
+    h, w = out.shape[:2]
+    for cls_id, score, (x1, y1, x2, y2) in detections:
+        px1, py1 = int(x1 * w), int(y1 * h)
+        px2, py2 = int(x2 * w), int(y2 * h)
+        label = f"{COCO_CLASSES[cls_id]} {score:.2f}"
+        cv2.rectangle(out, (px1, py1), (px2, py2), (0, 0, 255), 2)
+        (tw, th), _ = cv2.getTextSize(label, cv2.FONT_HERSHEY_SIMPLEX, 0.6, 2)
+        cv2.rectangle(out, (px1, py1 - th - 6), (px1 + tw + 4, py1), (0, 0, 255), -1)
+        cv2.putText(out, label, (px1 + 2, py1 - 4), cv2.FONT_HERSHEY_SIMPLEX,
+                    0.6, (255, 255, 255), 2, cv2.LINE_AA)
+    cv2.imwrite(output_path, out)
+    print(f"Saved annotated image: {output_path}")
+
+
+def main():
+    parser = argparse.ArgumentParser(description="YOLOv4 / YOLOv4x-mish ONNX demo (OpenCV DNN)")
+    parser.add_argument("--model", required=True, help="Path to ONNX model")
+    parser.add_argument("--image", default="example_outputs/input.jpg", help="Path to input image")
+    parser.add_argument("--output", default="output.jpg", help="Path for annotated output")
+    parser.add_argument("--input-size", type=int, default=0,
+                        help="Model input size (W=H). Default: auto-detect from filename or 416")
+    parser.add_argument("--conf", type=float, default=0.4, help="Confidence threshold")
+    parser.add_argument("--nms", type=float, default=0.5, help="NMS IoU threshold")
+    args = parser.parse_args()
+
+    # Determine input size: from arg, or by model name
+    if args.input_size > 0:
+        input_size = args.input_size
+    elif "mish" in args.model.lower():
+        input_size = 640
+    elif "tiny" in args.model.lower():
+        input_size = 416
+    else:
+        input_size = 608
+    print(f"Using input size: {input_size}x{input_size}")
+
+    # Load image
+    img = cv2.imread(args.image)
+    if img is None:
+        raise FileNotFoundError(f"Cannot read image: {args.image}")
+    print(f"Input image: {args.image} ({img.shape[1]}x{img.shape[0]})")
+
+    # Build blob: 1/255 scale, swap BGR -> RGB, no crop
+    blob = cv2.dnn.blobFromImage(img, 1.0 / 255.0,
+                                  (input_size, input_size),
+                                  swapRB=True, crop=False)
+    print(f"Blob shape: {blob.shape}")
+
+    # Load network with OpenCV DNN
+    net = cv2.dnn.readNetFromONNX(args.model)
+    net.setPreferableBackend(cv2.dnn.DNN_BACKEND_OPENCV)
+    net.setPreferableTarget(cv2.dnn.DNN_TARGET_CPU)
+
+    # Inference
+    net.setInput(blob)
+    out_names = net.getUnconnectedOutLayersNames()
+    print(f"Output layer names: {out_names}")
+    outputs = net.forward(out_names)
+    for name, o in zip(out_names, outputs):
+        print(f"Output '{name}' shape: {o.shape}")
+
+    # Postprocess
+    detections = postprocess(outputs, args.conf, args.nms)
+    print(f"\nDetections (conf >= {args.conf}, after NMS): {len(detections)}")
+    for cls_id, score, (x1, y1, x2, y2) in detections:
+        print(f"  {COCO_CLASSES[cls_id]:15s} score={score:.4f}  "
+              f"bbox=[{x1:.4f}, {y1:.4f}, {x2:.4f}, {y2:.4f}]")
+
+    draw_detections(img, detections, args.output)
+
+
+if __name__ == "__main__":
+    main()

yolov4/yolo_layer.py

@@ -0,0 +1,327 @@
+import torch.nn as nn
+import torch.nn.functional as F
+from tool.torch_utils import *
+
+def yolo_forward(output, conf_thresh, num_classes, anchors, num_anchors, scale_x_y, only_objectness=1,
+                              validation=False):
+    # Output would be invalid if it does not satisfy this assert
+    # assert (output.size(1) == (5 + num_classes) * num_anchors)
+
+    # print(output.size())
+
+    # Slice the second dimension (channel) of output into:
+    # [ 2, 2, 1, num_classes, 2, 2, 1, num_classes, 2, 2, 1, num_classes ]
+    # And then into
+    # bxy = [ 6 ] bwh = [ 6 ] det_conf = [ 3 ] cls_conf = [ num_classes * 3 ]
+    batch = output.size(0)
+    H = output.size(2)
+    W = output.size(3)
+
+    bxy_list = []
+    bwh_list = []
+    det_confs_list = []
+    cls_confs_list = []
+
+    for i in range(num_anchors):
+        begin = i * (5 + num_classes)
+        end = (i + 1) * (5 + num_classes)
+        
+        bxy_list.append(output[:, begin : begin + 2])
+        bwh_list.append(output[:, begin + 2 : begin + 4])
+        det_confs_list.append(output[:, begin + 4 : begin + 5])
+        cls_confs_list.append(output[:, begin + 5 : end])
+
+    # Shape: [batch, num_anchors * 2, H, W]
+    bxy = torch.cat(bxy_list, dim=1)
+    # Shape: [batch, num_anchors * 2, H, W]
+    bwh = torch.cat(bwh_list, dim=1)
+
+    # Shape: [batch, num_anchors, H, W]
+    det_confs = torch.cat(det_confs_list, dim=1)
+    # Shape: [batch, num_anchors * H * W]
+    det_confs = det_confs.view(batch, num_anchors * H * W)
+
+    # Shape: [batch, num_anchors * num_classes, H, W]
+    cls_confs = torch.cat(cls_confs_list, dim=1)
+    # Shape: [batch, num_anchors, num_classes, H * W]
+    cls_confs = cls_confs.view(batch, num_anchors, num_classes, H * W)
+    # Shape: [batch, num_anchors, num_classes, H * W] --> [batch, num_anchors * H * W, num_classes] 
+    cls_confs = cls_confs.permute(0, 1, 3, 2).reshape(batch, num_anchors * H * W, num_classes)
+
+    # Apply sigmoid(), exp() and softmax() to slices
+    #
+    bxy = torch.sigmoid(bxy) * scale_x_y - 0.5 * (scale_x_y - 1)
+    bwh = torch.exp(bwh)
+    det_confs = torch.sigmoid(det_confs)
+    cls_confs = torch.sigmoid(cls_confs)
+
+    # Prepare C-x, C-y, P-w, P-h (None of them are torch related)
+    grid_x = np.expand_dims(np.expand_dims(np.expand_dims(np.linspace(0, W - 1, W), axis=0).repeat(H, 0), axis=0), axis=0)
+    grid_y = np.expand_dims(np.expand_dims(np.expand_dims(np.linspace(0, H - 1, H), axis=1).repeat(W, 1), axis=0), axis=0)
+    # grid_x = torch.linspace(0, W - 1, W).reshape(1, 1, 1, W).repeat(1, 1, H, 1)
+    # grid_y = torch.linspace(0, H - 1, H).reshape(1, 1, H, 1).repeat(1, 1, 1, W)
+
+    anchor_w = []
+    anchor_h = []
+    for i in range(num_anchors):
+        anchor_w.append(anchors[i * 2])
+        anchor_h.append(anchors[i * 2 + 1])
+
+    device = None
+    cuda_check = output.is_cuda
+    if cuda_check:
+        device = output.get_device()
+
+    bx_list = []
+    by_list = []
+    bw_list = []
+    bh_list = []
+
+    # Apply C-x, C-y, P-w, P-h
+    for i in range(num_anchors):
+        ii = i * 2
+        # Shape: [batch, 1, H, W]
+        bx = bxy[:, ii : ii + 1] + torch.tensor(grid_x, device=device, dtype=torch.float32) # grid_x.to(device=device, dtype=torch.float32)
+        # Shape: [batch, 1, H, W]
+        by = bxy[:, ii + 1 : ii + 2] + torch.tensor(grid_y, device=device, dtype=torch.float32) # grid_y.to(device=device, dtype=torch.float32)
+        # Shape: [batch, 1, H, W]
+        bw = bwh[:, ii : ii + 1] * anchor_w[i]
+        # Shape: [batch, 1, H, W]
+        bh = bwh[:, ii + 1 : ii + 2] * anchor_h[i]
+
+        bx_list.append(bx)
+        by_list.append(by)
+        bw_list.append(bw)
+        bh_list.append(bh)
+
+
+    ########################################
+    #   Figure out bboxes from slices     #
+    ########################################
+    
+    # Shape: [batch, num_anchors, H, W]
+    bx = torch.cat(bx_list, dim=1)
+    # Shape: [batch, num_anchors, H, W]
+    by = torch.cat(by_list, dim=1)
+    # Shape: [batch, num_anchors, H, W]
+    bw = torch.cat(bw_list, dim=1)
+    # Shape: [batch, num_anchors, H, W]
+    bh = torch.cat(bh_list, dim=1)
+
+    # Shape: [batch, 2 * num_anchors, H, W]
+    bx_bw = torch.cat((bx, bw), dim=1)
+    # Shape: [batch, 2 * num_anchors, H, W]
+    by_bh = torch.cat((by, bh), dim=1)
+
+    # normalize coordinates to [0, 1]
+    bx_bw /= W
+    by_bh /= H
+
+    # Shape: [batch, num_anchors * H * W, 1]
+    bx = bx_bw[:, :num_anchors].view(batch, num_anchors * H * W, 1)
+    by = by_bh[:, :num_anchors].view(batch, num_anchors * H * W, 1)
+    bw = bx_bw[:, num_anchors:].view(batch, num_anchors * H * W, 1)
+    bh = by_bh[:, num_anchors:].view(batch, num_anchors * H * W, 1)
+
+    bx1 = bx - bw * 0.5
+    by1 = by - bh * 0.5
+    bx2 = bx1 + bw
+    by2 = by1 + bh
+
+    # Shape: [batch, num_anchors * h * w, 4] -> [batch, num_anchors * h * w, 1, 4]
+    boxes = torch.cat((bx1, by1, bx2, by2), dim=2).view(batch, num_anchors * H * W, 1, 4)
+    # boxes = boxes.repeat(1, 1, num_classes, 1)
+
+    # boxes:     [batch, num_anchors * H * W, 1, 4]
+    # cls_confs: [batch, num_anchors * H * W, num_classes]
+    # det_confs: [batch, num_anchors * H * W]
+
+    det_confs = det_confs.view(batch, num_anchors * H * W, 1)
+    confs = cls_confs * det_confs
+
+    # boxes: [batch, num_anchors * H * W, 1, 4]
+    # confs: [batch, num_anchors * H * W, num_classes]
+
+    return  boxes, confs
+
+
+def yolo_forward_dynamic(output, conf_thresh, num_classes, anchors, num_anchors, scale_x_y, only_objectness=1,
+                              validation=False, new_coords=0):
+    # Output would be invalid if it does not satisfy this assert
+    # assert (output.size(1) == (5 + num_classes) * num_anchors)
+
+    # print(output.size())
+
+    # Slice the second dimension (channel) of output into:
+    # [ 2, 2, 1, num_classes, 2, 2, 1, num_classes, 2, 2, 1, num_classes ]
+    # And then into
+    # bxy = [ 6 ] bwh = [ 6 ] det_conf = [ 3 ] cls_conf = [ num_classes * 3 ]
+    # batch = output.size(0)
+    # H = output.size(2)
+    # W = output.size(3)
+
+    bxy_list = []
+    bwh_list = []
+    det_confs_list = []
+    cls_confs_list = []
+
+    for i in range(num_anchors):
+        begin = i * (5 + num_classes)
+        end = (i + 1) * (5 + num_classes)
+        
+        bxy_list.append(output[:, begin : begin + 2])
+        bwh_list.append(output[:, begin + 2 : begin + 4])
+        det_confs_list.append(output[:, begin + 4 : begin + 5])
+        cls_confs_list.append(output[:, begin + 5 : end])
+
+    # Shape: [batch, num_anchors * 2, H, W]
+    bxy = torch.cat(bxy_list, dim=1)
+    # Shape: [batch, num_anchors * 2, H, W]
+    bwh = torch.cat(bwh_list, dim=1)
+
+    # Shape: [batch, num_anchors, H, W]
+    det_confs = torch.cat(det_confs_list, dim=1)
+    # Shape: [batch, num_anchors * H * W]
+    det_confs = det_confs.view(output.size(0), num_anchors * output.size(2) * output.size(3))
+
+    # Shape: [batch, num_anchors * num_classes, H, W]
+    cls_confs = torch.cat(cls_confs_list, dim=1)
+    # Shape: [batch, num_anchors, num_classes, H * W]
+    cls_confs = cls_confs.view(output.size(0), num_anchors, num_classes, output.size(2) * output.size(3))
+    # Shape: [batch, num_anchors, num_classes, H * W] --> [batch, num_anchors * H * W, num_classes] 
+    cls_confs = cls_confs.permute(0, 1, 3, 2).reshape(output.size(0), num_anchors * output.size(2) * output.size(3), num_classes)
+
+    # Apply activations based on new_coords flag
+    if new_coords:
+        # new_coords=1: no sigmoid on xy/conf/cls, squared width/height instead of exp
+        bxy = bxy * scale_x_y - 0.5 * (scale_x_y - 1)
+        bwh = (bwh * 2) ** 2
+        # det_confs and cls_confs are used as-is (no sigmoid)
+    else:
+        # Standard YOLO: sigmoid on xy/conf/cls, exp on wh
+        bxy = torch.sigmoid(bxy) * scale_x_y - 0.5 * (scale_x_y - 1)
+        bwh = torch.exp(bwh)
+        det_confs = torch.sigmoid(det_confs)
+        cls_confs = torch.sigmoid(cls_confs)
+
+    # Prepare C-x, C-y, P-w, P-h (None of them are torch related)
+    grid_x = np.expand_dims(np.expand_dims(np.expand_dims(np.linspace(0, output.size(3) - 1, output.size(3)), axis=0).repeat(output.size(2), 0), axis=0), axis=0)
+    grid_y = np.expand_dims(np.expand_dims(np.expand_dims(np.linspace(0, output.size(2) - 1, output.size(2)), axis=1).repeat(output.size(3), 1), axis=0), axis=0)
+
+    anchor_w = []
+    anchor_h = []
+    for i in range(num_anchors):
+        anchor_w.append(anchors[i * 2])
+        anchor_h.append(anchors[i * 2 + 1])
+
+    device = None
+    cuda_check = output.is_cuda
+    if cuda_check:
+        device = output.get_device()
+
+    bx_list = []
+    by_list = []
+    bw_list = []
+    bh_list = []
+
+    # Apply C-x, C-y, P-w, P-h
+    for i in range(num_anchors):
+        ii = i * 2
+        # Shape: [batch, 1, H, W]
+        bx = bxy[:, ii : ii + 1] + torch.tensor(grid_x, device=device, dtype=torch.float32)
+        # Shape: [batch, 1, H, W]
+        by = bxy[:, ii + 1 : ii + 2] + torch.tensor(grid_y, device=device, dtype=torch.float32)
+        # Shape: [batch, 1, H, W]
+        bw = bwh[:, ii : ii + 1] * anchor_w[i]
+        # Shape: [batch, 1, H, W]
+        bh = bwh[:, ii + 1 : ii + 2] * anchor_h[i]
+
+        bx_list.append(bx)
+        by_list.append(by)
+        bw_list.append(bw)
+        bh_list.append(bh)
+
+
+    ########################################
+    #   Figure out bboxes from slices     #
+    ########################################
+    
+    # Shape: [batch, num_anchors, H, W]
+    bx = torch.cat(bx_list, dim=1)
+    # Shape: [batch, num_anchors, H, W]
+    by = torch.cat(by_list, dim=1)
+    # Shape: [batch, num_anchors, H, W]
+    bw = torch.cat(bw_list, dim=1)
+    # Shape: [batch, num_anchors, H, W]
+    bh = torch.cat(bh_list, dim=1)
+
+    # Shape: [batch, 2 * num_anchors, H, W]
+    bx_bw = torch.cat((bx, bw), dim=1)
+    # Shape: [batch, 2 * num_anchors, H, W]
+    by_bh = torch.cat((by, bh), dim=1)
+
+    # normalize coordinates to [0, 1]
+    bx_bw /= output.size(3)
+    by_bh /= output.size(2)
+
+    # Shape: [batch, num_anchors * H * W, 1]
+    bx = bx_bw[:, :num_anchors].view(output.size(0), num_anchors * output.size(2) * output.size(3), 1)
+    by = by_bh[:, :num_anchors].view(output.size(0), num_anchors * output.size(2) * output.size(3), 1)
+    bw = bx_bw[:, num_anchors:].view(output.size(0), num_anchors * output.size(2) * output.size(3), 1)
+    bh = by_bh[:, num_anchors:].view(output.size(0), num_anchors * output.size(2) * output.size(3), 1)
+
+    bx1 = bx - bw * 0.5
+    by1 = by - bh * 0.5
+    bx2 = bx1 + bw
+    by2 = by1 + bh
+
+    # Shape: [batch, num_anchors * h * w, 4] -> [batch, num_anchors * h * w, 1, 4]
+    boxes = torch.cat((bx1, by1, bx2, by2), dim=2).view(output.size(0), num_anchors * output.size(2) * output.size(3), 1, 4)
+    # boxes = boxes.repeat(1, 1, num_classes, 1)
+
+    # boxes:     [batch, num_anchors * H * W, 1, 4]
+    # cls_confs: [batch, num_anchors * H * W, num_classes]
+    # det_confs: [batch, num_anchors * H * W]
+
+    det_confs = det_confs.view(output.size(0), num_anchors * output.size(2) * output.size(3), 1)
+    confs = cls_confs * det_confs
+
+    # boxes: [batch, num_anchors * H * W, 1, 4]
+    # confs: [batch, num_anchors * H * W, num_classes]
+
+    return  boxes, confs
+
+class YoloLayer(nn.Module):
+    ''' Yolo layer
+    model_out: while inference,is post-processing inside or outside the model
+        true:outside
+    '''
+    def __init__(self, anchor_mask=[], num_classes=0, anchors=[], num_anchors=1, stride=32, model_out=False):
+        super(YoloLayer, self).__init__()
+        self.anchor_mask = anchor_mask
+        self.num_classes = num_classes
+        self.anchors = anchors
+        self.num_anchors = num_anchors
+        self.anchor_step = len(anchors) // num_anchors
+        self.coord_scale = 1
+        self.noobject_scale = 1
+        self.object_scale = 5
+        self.class_scale = 1
+        self.thresh = 0.6
+        self.stride = stride
+        self.seen = 0
+        self.scale_x_y = 1
+        self.new_coords = 0
+
+        self.model_out = model_out
+
+    def forward(self, output, target=None):
+        if self.training:
+            return output
+        masked_anchors = []
+        for m in self.anchor_mask:
+            masked_anchors += self.anchors[m * self.anchor_step:(m + 1) * self.anchor_step]
+        masked_anchors = [anchor / self.stride for anchor in masked_anchors]
+
+        return yolo_forward_dynamic(output, self.thresh, self.num_classes, masked_anchors, len(self.anchor_mask),scale_x_y=self.scale_x_y, new_coords=self.new_coords)
+

yolov4/yolov4-tiny.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:24c558aaeca05d96337f48b94ac0d8d61b94f7a2591eab9a2bab87cabfd71e07
+size 24316593

yolov4/yolov4.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fb56ae63aad4ba2c320cb9275f62a9a864f8cae680135821c1810fa560a00512
+size 257676998

yolov4/yolov4x-mish.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d14a605672e45d2c36618705c7e27124ed4e4189ad8db9ec0bd71d3ab0a860e
+size 399210397