yolov3
yolov3 copied to clipboard
erikguo
How to do fine-tuning?
I am using latest python 3 and pytorch version under latest anaconda.
As we know, there are four steps to train custom model
- normal training
- sparsity training
- channel pruning
- fine tuning
by step 3, I have got pruned model. And the result is a .pt file. In original paper, the author describe that the last fine tuning step is done by original darknet, not ultralytics's yolov3. So I am trying to run using darknet. Then how to do fine-tuning with the .pt file? I have tried to convert using convert_pt_weights.py, it fails to convert because of decoding error like below.
command: python convert_pt_weights.py weights/last.pt ./pruned.weights
result: UnicodeDecodeError: 'utf-8' codec can't decode byte 0x80 in position 0: invalid start byte
Please let me know a way to do last step.
Thanks in advance.
I wrote wrong command so far.
Below one is correct. python convert_pt_weights.py prune_0.5_0.1/prune.cfg prune_0.5_0.1/prune.pt
Then, I have converted weight model with below message. Success: converted 'prune_0.5_0.1/prune.pt' to 'converted.weights'
And I also have got error in models.py, So I have modified like below.
from: model.load_state_dict(torch.load(weights, map_location='cpu')['model']) to: model.load_state_dict(torch.load(weights, map_location='cpu'))
I wrote wrong command so far.
Below one is correct. python convert_pt_weights.py prune_0.5_0.1/prune.cfg prune_0.5_0.1/prune.pt
Then, I have got converted weight model with below message. Success: converted 'prune_0.5_0.1/prune.pt' to 'converted.weights'
And I also have got error in models.py, So I have modified like below.
from: model.load_state_dict(torch.load(weights, map_location='cpu')['model']) to: model.load_state_dict(torch.load(weights, map_location='cpu'))
Dear @dsgh2 ,
I am so sorry to bother you, but could you please do me a favor?
I want to sparse train resnet152_trident.cfg on my custom dataset. However, an error occur when I input the following command: python train.py --data data/obj.data --cfg cfg/resnet152_trident.cfg
I find that all the Resnet-structure can not been sparsity train on this project. Do you know what should I do now? I wpuld be appreciated if you can give me a hand.
I give the content of resnet152_trident.cfg as following, looking forward to hearing from you.
Thank you very much.
[net]
Testing
#batch=1 #subdivisions=1
Training
batch=64 subdivisions=64 width=416 height=416 channels=3 momentum=0.9 decay=0.0005 angle=0 saturation = 1.5 exposure = 1.5 hue=.1
learning_rate=0.001 burn_in=1000 max_batches = 10000
policy=sgdr sgdr_cycle=1000 sgdr_mult=2 steps=4000,6000,8000,9000 #scales=1, 1, 0.1, 0.1
[convolutional] batch_normalize=1 filters=64 size=7 stride=2 pad=1 activation=leaky
[maxpool] size=2 stride=2
[convolutional] batch_normalize=1 filters=64 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=64 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=64 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=64 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=64 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=64 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=3 stride=2 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=128 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
Conv 4
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=2 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
TridentNet - large objects - Start
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] dilation=3 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] dilation=3 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] dilation=3 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] dilation=3 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] dilation=3 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] dilation=3 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] dilation=3 batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
Conv 5
[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky
[convolutional] dilation=3 batch_normalize=1 filters=512 size=3 stride=2 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=2048 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky
[convolutional] dilation=3 batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=2048 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky
[convolutional] dilation=3 batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky
[convolutional] batch_normalize=1 filters=2048 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 size=1 stride=1 pad=1 filters=2048 activation=leaky
[convolutional] size=1 stride=1 pad=1 filters=24 activation=linear
[yolo] mask = 8,9,10,11 anchors = 8,8, 10,13, 16,30, 33,23, 32,32, 30,61, 62,45, 59,119, 80,80, 116,90, 156,198, 373,326 classes=1 num=12 jitter=.3 ignore_thresh = .7 truth_thresh = 1 random=0
TridentNet - large objects - End
TridentNet - medium objects - Start
[route] layers = 165
[convolutional] share_index=166 batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=167 dilation=2 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=168 batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=170 batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=171 dilation=2 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=172 batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=174 batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=175 dilation=2 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=176 batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=178 batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=179 dilation=2 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=180 batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=182 batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=183 dilation=2 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=184 batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=186 batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=187 dilation=2 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=188 dilation=2 batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
Conv 5
[convolutional] share_index=190 batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=191 dilation=2 batch_normalize=1 filters=512 size=3 stride=2 pad=1 activation=leaky
[convolutional] share_index=192 batch_normalize=1 filters=2048 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=194 batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=195 dilation=2 batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=196 batch_normalize=1 filters=2048 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=198 batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=199 dilation=2 batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=200 batch_normalize=1 filters=2048 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 size=1 stride=1 pad=1 filters=1024 activation=leaky
[upsample] stride=2
[route] layers = -1, 49
[convolutional] batch_normalize=1 size=1 stride=1 pad=1 filters=1024 activation=leaky
[convolutional] size=1 stride=1 pad=1 filters=24 activation=linear
[yolo] mask = 4,5,6,7 anchors = 8,8, 10,13, 16,30, 33,23, 32,32, 30,61, 62,45, 64,64, 59,119, 116,90, 156,198, 373,326 classes=1 num=12 jitter=.3 ignore_thresh = .7 truth_thresh = 1 random=0
TridentNet - medium objects - End
TridentNet - small objects - Start
[route] layers = 165
[convolutional] share_index=166 batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=167 dilation=1 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=168 batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=170 batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=171 dilation=1 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=172 batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=174 batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=175 dilation=1 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=176 batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=178 batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=179 dilation=1 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=180 batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=182 batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=183 dilation=1 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=184 batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=186 batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=187 dilation=1 batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=188 dilation=1 batch_normalize=1 filters=1024 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
Conv 5
[convolutional] share_index=190 batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=191 dilation=1 batch_normalize=1 filters=512 size=3 stride=2 pad=1 activation=leaky
[convolutional] share_index=192 batch_normalize=1 filters=2048 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=194 batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=195 dilation=1 batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=196 batch_normalize=1 filters=2048 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] share_index=198 batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky
[convolutional] share_index=199 dilation=1 batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky
[convolutional] share_index=200 batch_normalize=1 filters=2048 size=1 stride=1 pad=1 activation=linear
[shortcut] from=-4 activation=leaky
[convolutional] batch_normalize=1 size=1 stride=1 pad=1 filters=512 activation=leaky
[upsample] stride=4
[route] layers = -1, 17
[convolutional] batch_normalize=1 size=1 stride=1 pad=1 filters=512 activation=leaky
[convolutional] size=1 stride=1 pad=1 filters=24 activation=linear
[yolo] mask = 0,1,2,3 anchors = 8,8, 10,13, 16,30, 33,23, 32,32, 30,61, 62,45, 64,64, 59,119, 116,90, 156,198, 373,326 classes=1 num=12 jitter=.3 ignore_thresh = .7 truth_thresh = 1 random=0
TridentNet - small objects - End
Well, I have never tried without yolov3. I am struggling with this many days. If I got any solution, I notice you.
My pruned yolov3 model fails to detect anything even though fine-tuning is done.
Well, I have never tried without yolov3. I am struggling with this many days. If I got any solution, I notice you.
My pruned yolov3 model fails to detect anything even though fine-tuning is done.
Thank you very much for your reply.
Looking forward to hearing from you, good luck!