yolov7-w6 onnx with 8 output

[alicera]

Do you know the '630', '633', '636', '639' ? What is it?

python export.py yolov7-w6.pt --simplify

model : yolov7-w6

Inputs:  ['images']
Outputs:  ['output', '667', '681', '695', '630', '633', '636', '639'] 
-------------------------------------------
input name images
input shape [1, 3, 1280, 1280]
input type tensor(float) 

output name: output
output shape: [1, 3, 160, 160, 85]
output type: tensor(float) 

output name: 667
output shape: [1, 3, 80, 80, 85]
output type: tensor(float) 

output name: 681
output shape: [1, 3, 40, 40, 85]
output type: tensor(float) 

output name: 695
output shape: [1, 3, 20, 20, 85]
output type: tensor(float) 

output name: 630
output shape: [1, 320, 160, 160]
output type: tensor(float) 

output name: 633
output shape: [1, 640, 80, 80]
output type: tensor(float) 

output name: 636
output shape: [1, 960, 40, 40]
output type: tensor(float) 

output name: 639
output shape: [1, 1280, 20, 20]
output type: tensor(float) 

[SinChen]

you should run reparameterization before export onnx model

[alicera]

I train my data with yolov7-w6 It doesn't nc. How can I solve it ?

/opt/conda/lib/python3.8/site-packages/scipy/__init__.py:138: UserWarning: A NumPy version >=1.16.5 and <1.23.0 is required for this version of SciPy (detected version 1.23.4)
  warnings.warn(f"A NumPy version >={np_minversion} and <{np_maxversion} is required for this version of "
Traceback (most recent call last):
  File "reparam_yolov7-w6.py", line 24, in <module>
    model.nc = ckpt['model'].nc
  File "/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1131, in __getattr__
    raise AttributeError("'{}' object has no attribute '{}'".format(
AttributeError: 'Model' object has no attribute 'nc'

[alicera]

Here is the reparam.py

# import
from copy import deepcopy
from models.yolo import Model
import torch
from utils.torch_utils import select_device, is_parallel
import yaml

device = select_device('0', batch_size=1)
# model trained by cfg/training/*.yaml
ckpt = torch.load('./last.pt', map_location=device)
# reparameterized model in cfg/deploy/*.yaml
model = Model('cfg/deploy/yolov7-w6.yaml', ch=3, nc=3).to(device)

with open('cfg/deploy/yolov7-w6.yaml') as f:
    yml = yaml.load(f, Loader=yaml.SafeLoader)
anchors = len(yml['anchors'][0]) // 2

# copy intersect weights
state_dict = ckpt['model'].float().state_dict()
exclude = []
intersect_state_dict = {k: v for k, v in state_dict.items() if k in model.state_dict() and not any(x in k for x in exclude) and v.shape == model.state_dict()[k].shape}
model.load_state_dict(intersect_state_dict, strict=False)
model.names = ckpt['model'].names
model.nc = ckpt['model'].nc

idx = 118
idx2 = 122

# copy weights of lead head
model.state_dict()['model.{}.m.0.weight'.format(idx)].data -= model.state_dict()['model.{}.m.0.weight'.format(idx)].data
model.state_dict()['model.{}.m.1.weight'.format(idx)].data -= model.state_dict()['model.{}.m.1.weight'.format(idx)].data
model.state_dict()['model.{}.m.2.weight'.format(idx)].data -= model.state_dict()['model.{}.m.2.weight'.format(idx)].data
model.state_dict()['model.{}.m.3.weight'.format(idx)].data -= model.state_dict()['model.{}.m.3.weight'.format(idx)].data
model.state_dict()['model.{}.m.0.weight'.format(idx)].data += state_dict['model.{}.m.0.weight'.format(idx2)].data
model.state_dict()['model.{}.m.1.weight'.format(idx)].data += state_dict['model.{}.m.1.weight'.format(idx2)].data
model.state_dict()['model.{}.m.2.weight'.format(idx)].data += state_dict['model.{}.m.2.weight'.format(idx2)].data
model.state_dict()['model.{}.m.3.weight'.format(idx)].data += state_dict['model.{}.m.3.weight'.format(idx2)].data
model.state_dict()['model.{}.m.0.bias'.format(idx)].data -= model.state_dict()['model.{}.m.0.bias'.format(idx)].data
model.state_dict()['model.{}.m.1.bias'.format(idx)].data -= model.state_dict()['model.{}.m.1.bias'.format(idx)].data
model.state_dict()['model.{}.m.2.bias'.format(idx)].data -= model.state_dict()['model.{}.m.2.bias'.format(idx)].data
model.state_dict()['model.{}.m.3.bias'.format(idx)].data -= model.state_dict()['model.{}.m.3.bias'.format(idx)].data
model.state_dict()['model.{}.m.0.bias'.format(idx)].data += state_dict['model.{}.m.0.bias'.format(idx2)].data
model.state_dict()['model.{}.m.1.bias'.format(idx)].data += state_dict['model.{}.m.1.bias'.format(idx2)].data
model.state_dict()['model.{}.m.2.bias'.format(idx)].data += state_dict['model.{}.m.2.bias'.format(idx2)].data
model.state_dict()['model.{}.m.3.bias'.format(idx)].data += state_dict['model.{}.m.3.bias'.format(idx2)].data

# reparametrized YOLOR
for i in range((model.nc+5)*anchors):
    model.state_dict()['model.{}.m.0.weight'.format(idx)].data[i, :, :, :] *= state_dict['model.{}.im.0.implicit'.format(idx2)].data[:, i, : :].squeeze()
    model.state_dict()['model.{}.m.1.weight'.format(idx)].data[i, :, :, :] *= state_dict['model.{}.im.1.implicit'.format(idx2)].data[:, i, : :].squeeze()
    model.state_dict()['model.{}.m.2.weight'.format(idx)].data[i, :, :, :] *= state_dict['model.{}.im.2.implicit'.format(idx2)].data[:, i, : :].squeeze()
    model.state_dict()['model.{}.m.3.weight'.format(idx)].data[i, :, :, :] *= state_dict['model.{}.im.3.implicit'.format(idx2)].data[:, i, : :].squeeze()
model.state_dict()['model.{}.m.0.bias'.format(idx)].data += state_dict['model.{}.m.0.weight'.format(idx2)].mul(state_dict['model.{}.ia.0.implicit'.format(idx2)]).sum(1).squeeze()
model.state_dict()['model.{}.m.1.bias'.format(idx)].data += state_dict['model.{}.m.1.weight'.format(idx2)].mul(state_dict['model.{}.ia.1.implicit'.format(idx2)]).sum(1).squeeze()
model.state_dict()['model.{}.m.2.bias'.format(idx)].data += state_dict['model.{}.m.2.weight'.format(idx2)].mul(state_dict['model.{}.ia.2.implicit'.format(idx2)]).sum(1).squeeze()
model.state_dict()['model.{}.m.3.bias'.format(idx)].data += state_dict['model.{}.m.3.weight'.format(idx2)].mul(state_dict['model.{}.ia.3.implicit'.format(idx2)]).sum(1).squeeze()
model.state_dict()['model.{}.m.0.bias'.format(idx)].data *= state_dict['model.{}.im.0.implicit'.format(idx2)].data.squeeze()
model.state_dict()['model.{}.m.1.bias'.format(idx)].data *= state_dict['model.{}.im.1.implicit'.format(idx2)].data.squeeze()
model.state_dict()['model.{}.m.2.bias'.format(idx)].data *= state_dict['model.{}.im.2.implicit'.format(idx2)].data.squeeze()
model.state_dict()['model.{}.m.3.bias'.format(idx)].data *= state_dict['model.{}.im.3.implicit'.format(idx2)].data.squeeze()

# model to be saved
ckpt = {'model': deepcopy(model.module if is_parallel(model) else model).half(),
        'optimizer': None,
        'training_results': None,
        'epoch': -1}

# save reparameterized model
torch.save(ckpt, 'cfg/deploy/yolov7-w6.pt')

[Monalsingh]

I am facing the same issue. I am trying to convert yolov7_pose to onnx and its has multiple outputs. How to process these predictions? Please help.