bug

Question

bug

Dontfall opened this issue 4 years ago · 37 comments

作者，您好！
我成功的在yolo中使我解决了上一个问题 from n params module 0 -1 1 1 -1 1 2 -1 1 3 -1 1 4 5 6 7 8 9 10 11 -1 1 12 [-1, 6] 1 13 14 15 -1 1 16 [-1, 4] 1 17 18 19 [-1, 14] 1 20 21 22 [-1, 10] 1 23 24 [17, 20, 23] 1 Model Summary: 用了det/mmdet/models/utils/involution_naive.py。但是，在使用involution_cuda.py 的过程中碰到了麻烦。
https://github.com/d-li14/involution/issues/15，但是遇到了新的问题，问题如下：
arguments
3520 models.common.Focus [3, 32, 3]
1154 models.experimental.involution [32, 7, 2]
16768 models.common.C3 [32, 64, 1]
73984 models.common.Conv [64, 128, 3, 2]
-1 1 156928 models.common.C3 [128, 128, 3]
-1 1 295424 models.common.Conv [128, 256, 3, 2]
-1 1 625152 models.common.C3 [256, 256, 3]
-1 1 1180672 models.common.Conv [256, 512, 3, 2]
-1 1 656896 models.common.SPP [512, 512, [5, 9, 13]]
-1 1 1182720 models.common.C3 [512, 512, 1, False]
-1 1 131584 models.common.Conv [512, 256, 1, 1]
0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
0 models.common.Concat [1]
-1 1 361984 models.common.C3 [512, 256, 1, False]
-1 1 33024 models.common.Conv [256, 128, 1, 1]
0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
0 models.common.Concat [1]
-1 1 90880 models.common.C3 [256, 128, 1, False]
-1 1 147712 models.common.Conv [128, 128, 3, 2]
0 models.common.Concat [1]
-1 1 296448 models.common.C3 [256, 256, 1, False]
-1 1 590336 models.common.Conv [256, 256, 3, 2]
0 models.common.Concat [1]
-1 1 1182720 models.common.C3 [512, 512, 1, False]
229245 models.yolo.Detect [80, [[10, 13, 16, 30, 33, 23], [30, 61, 62, 45, 59, 119], [116, 90, 156, 198, 373, 326]], [128, 256, 512]]
288 layers, 7257151 parameters, 7257151 gradients, 16.1 GFLOPS

Scaled weight_decay = 0.0005
Optimizer groups: 63 .bias, 63 conv.weight, 59 other
train: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|████| 128/128 [00:00<?, ?it/sPlotting labels... coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|██████| 128/128 [00:00<?, ?it/s]
val: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|██████| 128/128 [00:00<?, ?it/s]
train: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|████| 128/128 [00:00<?, ?it/s]
val: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|██████| 128/128 [00:00<?, ?it/s]
val: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|██████| 128/128 [00:00<?, ?it/s]

autoanchor: Analyzing anchors... anchors/target = 4.26, Best Possible Recall (BPR) = 0.9946
Image sizes 640 train, 640 test
Using 2 dataloader workers
Logging results to runs/train/exp22
Starting training for 300 epochs...

 Epoch   gpu_mem       box       obj       cls     total   targets  img_size

0%| | 0/64 [00:05<?, ?it/s]
Traceback (most recent call last):
File "train.py", line 532, in
train(hyp, opt, device, tb_writer, wandb)
File "train.py", line 297, in train
pred = model(imgs) # forward
File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/parallel/data_parallel.py", line 161, in forward
outputs = self.parallel_apply(replicas, inputs, kwargs)
File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/parallel/data_parallel.py", line 171, in parallel_apply
return parallel_apply(replicas, inputs, kwargs, self.device_ids[:len(replicas)])
File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/parallel/parallel_apply.py", line 86, in parallel_apply
output.reraise()
File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/_utils.py", line 428, in reraise
raise self.exc_type(msg)
TypeError: init() missing 3 required positional arguments: 'source', 'name', and 'options'

希望能得到您解决此问题的建议，谢谢！

Answer 1 · 2021-04-04T12:16:49.000Z

@aidarikako

Answer 2 · 2021-04-04T14:20:27.000Z

我的环境是pytorch1.7.0-gpu，mmcv-full-1.2.7，cuda 10.2

Answer 3 · 2021-04-07T02:46:50.000Z

作者，您好！
我成功的在yolo中使用了det/mmdet/中使用了/utils/inconcription_nayve.py。但是，在使用对合cuda.py的过程中碰到了麻烦.
我解决了上一个问题 Https://github.com/d-li14/involution/issues/15，但是遇到了新的问题，问题如下：
从n个仿射模自变量
0-1，1,3520型号.公共.焦点[3，32，3]
1-1 1154模型.试验.对合[32，7，2]
2-11 16768型号.公共.C3[32，64，1]
3-11 73984型号.通用.Conv[64,128，3，2]
4-11 156928型号.公共.C3[128,128，3]
5-1 1 295424型号通用.Conv[128,256，3，2]
6-1 1 625152型号.公共.C3[256,256，3]
7-1 1 1180672型号.通用.Conv[256,512，3，2]
8-11 656896型号.通用.spp[512,512，[5，9，13]
9-1 1 1182720型号.Common.C3[512,512，1，false]
10-1 1 131584型号.通用.Conv[512,256，1，1]
11-110 Torch.nn.Modes.upsampl.Upsampl.UpSample[无，2，‘最近’]
12[-1，6]10型号.通用.Concat[1]
13-11 361984型号.公共.C3[512,256，1，假]
14-1 1 33024型号.通用.Conv[256,128，1，1]
15-10 Torch.nn.Modes.upsampl.Upsampl.UpSample[无，2，‘最近’]
16[-1，4]10型号.通用.Concat[1]
17-1 1 90880型号.公共.C3[256,128，1，假]
18-1 1 147712型号.通用.Conv[128,128，3，2]
19[-1，14]10种型号.通用.Concat[1]
20-1 1 296448型号.公共.C3[256,256，1，假]
21-1 1 590336型号.通用.Conv[256,256，3，2]
22[-1，10]10型号.通用.Concat[1]
23-1 1 1182720型号.Common.C3[512,512，1，false]
24[17，20，23]1 229245型号s.yolo.detect[80，[10，13，16，30，33，23]，[30，61，62，45，59,119]，[116，90,156,198,373,326]，[128,256,512]]
模型摘要：288个层，7257151个参数，7257151个梯度，16.1个GFLOPS

标度重量衰减=0.0005
优化组：63.偏差，63卷积.权重，59其他
列车：扫描‘./可可128/标签/火车2017.cache’的图像和标签.找到，0缺失，2空，0损坏：100%████128/128[00：00<？，it/sPlotting标签.可可128/标签/火车2017.cache‘的图像和标签.找到，0缺失，2空，0损坏：100%█128/128[00：00<？，it/s]
Val：扫描图像和标签的‘../coo 128/标签/列车2017.cache’.找到，0缺失，2空，0损坏：100%█128/128[00：00<？，it/s]
列车：扫描‘./可可128/标签/火车2017.cache’的图像和标签.找到，0缺失，2空，0损坏：100%████128/128[00：00<？，it/s]
Val：扫描图像和标签的‘../coo 128/标签/列车2017.cache’.找到，0缺失，2空，0损坏：100%█128/128[00：00<？，it/s]
Val：扫描图像和标签的‘../coo 128/标签/列车2017.cache’.找到，0缺失，2空，0损坏：100%█128/128[00：00<？，it/s]

分析锚..。锚/目标=4.26，最佳可能召回(BPR)=0.9946
图像尺寸640列车，640测试
使用2个数据处理程序工作人员
记录运行/列车/费用的结果22
开始训练300次.
GPU_mem盒OBJ CLS总目标IMG_size
0%0/64[00：05<？，it/s]
追溯(最近一次调用)：
文件“tra.py”，第532行，在
列车(Hyp，OPT，设备，TB_WANDER，WANDEB)
文件“tra.py”，第297行，在火车上
Pred=模型(IMGS)#前进
文件“/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/modules/module.py”，第727行，in_Call_impl
结果=自我.前进(*输入，**kwargs)
文件“/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/parallel/data_parallel.py”，第161行，向前
输出=自。并行_应用(副本、输入、kwargs)
文件“/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/parallel/data_parallel.py”，第171行，并行应用
返回ParallyApply(副本、输入、kwargs、Self.Device_ID[：Len(副本)])
文件“/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/parallel/parallel_apply.py”，第86行，并行应用
Output.reraise()
文件“/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/_utils.py”，第428行，重新排列
提升.exc_type(Msg)
TypeError：依尼特()缺少3个所需的位置参数：“源”、“名称”和“选项”

希望能得到您解决此问题的建议，谢谢！

我也是加到yolo，但是我loss显示为nan，请问我该怎么解决

Answer 4 · 2021-04-20T12:22:24.000Z

作者，您好！
我成功的在yolo中使用了det/mmdet/models/utils/involution_naive.py。但是，在使用involution_cuda.py 的过程中碰到了麻烦。
我解决了上一个问题 https://github.com/d-li14/involution/issues/15，但是遇到了新的问题，问题如下：
from n params module arguments
0 -1 1 3520 models.common.Focus [3, 32, 3]
1 -1 1 1154 models.experimental.involution [32, 7, 2]
2 -1 1 16768 models.common.C3 [32, 64, 1]
3 -1 1 73984 models.common.Conv [64, 128, 3, 2]
4 -1 1 156928 models.common.C3 [128, 128, 3]
5 -1 1 295424 models.common.Conv [128, 256, 3, 2]
6 -1 1 625152 models.common.C3 [256, 256, 3]
7 -1 1 1180672 models.common.Conv [256, 512, 3, 2]
8 -1 1 656896 models.common.SPP [512, 512, [5, 9, 13]]
9 -1 1 1182720 models.common.C3 [512, 512, 1, False]
10 -1 1 131584 models.common.Conv [512, 256, 1, 1]
11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
12 [-1, 6] 1 0 models.common.Concat [1]
13 -1 1 361984 models.common.C3 [512, 256, 1, False]
14 -1 1 33024 models.common.Conv [256, 128, 1, 1]
15 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
16 [-1, 4] 1 0 models.common.Concat [1]
17 -1 1 90880 models.common.C3 [256, 128, 1, False]
18 -1 1 147712 models.common.Conv [128, 128, 3, 2]
19 [-1, 14] 1 0 models.common.Concat [1]
20 -1 1 296448 models.common.C3 [256, 256, 1, False]
21 -1 1 590336 models.common.Conv [256, 256, 3, 2]
22 [-1, 10] 1 0 models.common.Concat [1]
23 -1 1 1182720 models.common.C3 [512, 512, 1, False]
24 [17, 20, 23] 1 229245 models.yolo.Detect [80, [[10, 13, 16, 30, 33, 23], [30, 61, 62, 45, 59, 119], [116, 90, 156, 198, 373, 326]], [128, 256, 512]]
Model Summary: 288 layers, 7257151 parameters, 7257151 gradients, 16.1 GFLOPS

Scaled weight_decay = 0.0005
Optimizer groups: 63 .bias, 63 conv.weight, 59 other
train: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|████| 128/128 [00:00<?, ?it/sPlotting labels... coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|██████| 128/128 [00:00<?, ?it/s]
val: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|██████| 128/128 [00:00<?, ?it/s]
train: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|████| 128/128 [00:00<?, ?it/s]
val: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|██████| 128/128 [00:00<?, ?it/s]
val: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|██████| 128/128 [00:00<?, ?it/s]

autoanchor: Analyzing anchors... anchors/target = 4.26, Best Possible Recall (BPR) = 0.9946
Image sizes 640 train, 640 test
Using 2 dataloader workers
Logging results to runs/train/exp22
Starting training for 300 epochs...
 Epoch   gpu_mem       box       obj       cls     total   targets  img_size
0%| | 0/64 [00:05<?, ?it/s]
Traceback (most recent call last):
File "train.py", line 532, in
train(hyp, opt, device, tb_writer, wandb)
File "train.py", line 297, in train
pred = model(imgs) # forward
File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/parallel/data_parallel.py", line 161, in forward
outputs = self.parallel_apply(replicas, inputs, kwargs)
File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/parallel/data_parallel.py", line 171, in parallel_apply
return parallel_apply(replicas, inputs, kwargs, self.device_ids[:len(replicas)])
File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/parallel/parallel_apply.py", line 86, in parallel_apply
output.reraise()
File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/_utils.py", line 428, in reraise
raise self.exc_type(msg)
TypeError: init() missing 3 required positional arguments: 'source', 'name', and 'options'

希望能得到您解决此问题的建议，谢谢！
@Dontfall 您好！我用involution替换YOLO中的conv模块，用了involution_naive.py，一直报错，想问一下还需要involution项目的什么文件吗？
File "D:\win10 pytorch17\yolov5-master1\utils\involution.py", line 21, in init
self.conv1 = ConvModule(
TypeError: init() got an unexpected keyword argument 'k'

Answer 5 · 2021-04-20T12:25:37.000Z

只需要involution_naive.py 发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Tue,Apr 20,2021 8:22 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30) 作者，您好！我成功的在yolo中使用了det/mmdet/models/utils/involution_naive.py。但是，在使用involution_cuda.py 的过程中碰到了麻烦。我解决了上一个问题 #15，但是遇到了新的问题，问题如下： from n params module arguments 0 -1 1 3520 models.common.Focus [3, 32, 3] 1 -1 1 1154 models.experimental.involution [32, 7, 2] 2 -1 1 16768 models.common.C3 [32, 64, 1] 3 -1 1 73984 models.common.Conv [64, 128, 3, 2] 4 -1 1 156928 models.common.C3 [128, 128, 3] 5 -1 1 295424 models.common.Conv [128, 256, 3, 2] 6 -1 1 625152 models.common.C3 [256, 256, 3] 7 -1 1 1180672 models.common.Conv [256, 512, 3, 2] 8 -1 1 656896 models.common.SPP [512, 512, [5, 9, 13]] 9 -1 1 1182720 models.common.C3 [512, 512, 1, False] 10 -1 1 131584 models.common.Conv [512, 256, 1, 1] 11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 12 [-1, 6] 1 0 models.common.Concat [1] 13 -1 1 361984 models.common.C3 [512, 256, 1, False] 14 -1 1 33024 models.common.Conv [256, 128, 1, 1] 15 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 16 [-1, 4] 1 0 models.common.Concat [1] 17 -1 1 90880 models.common.C3 [256, 128, 1, False] 18 -1 1 147712 models.common.Conv [128, 128, 3, 2] 19 [-1, 14] 1 0 models.common.Concat [1] 20 -1 1 296448 models.common.C3 [256, 256, 1, False] 21 -1 1 590336 models.common.Conv [256, 256, 3, 2] 22 [-1, 10] 1 0 models.common.Concat [1] 23 -1 1 1182720 models.common.C3 [512, 512, 1, False] 24 [17, 20, 23] 1 229245 models.yolo.Detect [80, [[10, 13, 16, 30, 33, 23], [30, 61, 62, 45, 59, 119], [116, 90, 156, 198, 373, 326]], [128, 256, 512]] Model Summary: 288 layers, 7257151 parameters, 7257151 gradients, 16.1 GFLOPS Scaled weight_decay = 0.0005 Optimizer groups: 63 .bias, 63 conv.weight, 59 other train: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|████| 128/128 [00:00<?, ?it/sPlotting labels... coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|██████| 128/128 [00:00<?, ?it/s] val: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|██████| 128/128 [00:00<?, ?it/s] train: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|████| 128/128 [00:00<?, ?it/s] val: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|██████| 128/128 [00:00<?, ?it/s] val: Scanning '../coco128/labels/train2017.cache' for images and labels... 128 found, 0 missing, 2 empty, 0 corrupted: 100%|██████| 128/128 [00:00<?, ?it/s] autoanchor: Analyzing anchors... anchors/target = 4.26, Best Possible Recall (BPR) = 0.9946 Image sizes 640 train, 640 test Using 2 dataloader workers Logging results to runs/train/exp22 Starting training for 300 epochs... Epoch gpu_mem box obj cls total targets img_size 0%| | 0/64 [00:05<?, ?it/s] Traceback (most recent call last): File "train.py", line 532, in train(hyp, opt, device, tb_writer, wandb) File "train.py", line 297, in train pred = model(imgs) # forward File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/modules/module.py", line 727, in _call_impl result = self.forward(*input, **kwargs) File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/parallel/data_parallel.py", line 161, in forward outputs = self.parallel_apply(replicas, inputs, kwargs) File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/parallel/data_parallel.py", line 171, in parallel_apply return parallel_apply(replicas, inputs, kwargs, self.device_ids[:len(replicas)]) File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/nn/parallel/parallel_apply.py", line 86, in parallel_apply output.reraise() File "/home/dhh135246/anaconda3/envs/pytorch1.7.0-gpu/lib/python3.8/site-packages/torch/_utils.py", line 428, in reraise raise self.exc_type(msg) TypeError: init() missing 3 required positional arguments: 'source', 'name', and 'options' 希望能得到您解决此问题的建议，谢谢！ @Dontfall 您好！我用involution替换YOLO中的conv模块，用了involution_naive.py，一直报错，想问一下还需要involution项目的什么文件吗？ File "D:\win10 pytorch17\yolov5-master1\utils\involution.py", line 21, in init self.conv1 = ConvModule( TypeError: init() got an unexpected keyword argument 'k' — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 6 · 2021-04-21T03:31:39.000Z

@Dontfall 嗯，谢谢，但是我的involution_naive。py中输出的通道为54，但是对接YOLO中对应的通道是32 ，出现不一致情况，请问一般怎么解决

Answer 7 · 2021-04-21T03:32:45.000Z

调参数发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Wed,Apr 21,2021 11:31 AM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30) @Dontfall 嗯，谢谢，但是我的involution_naive。py中输出的通道为54，但是对接YOLO中对应的通道是32 ，出现不一致情况，请问一般怎么解决 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 8 · 2021-04-21T12:39:06.000Z

@Dontfall 如果我用involution_cuda.py也是只用这个文件就可以是吗

Answer 9 · 2021-04-21T12:43:08.000Z

是的，只要用这个文件就行发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Wed,Apr 21,2021 8:39 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30) @Dontfall 如果我用involution_cuda.py也是只用这个文件就可以是吗 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 10 · 2021-04-21T13:17:43.000Z

@Dontfall 您好，我这边一直出问题，不知道怎么处理，可以请教一下不
out = self.unfold(x).view(b, self.g, self.gc, self.k ** 2, h, w)
RuntimeError: shape '[1, 2, 24, 9, 128, 128]' is invalid for input of size 1769472

Answer 11 · 2021-04-21T13:36:16.000Z

你需要安装mmcv库发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Wed,Apr 21,2021 9:18 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30) @Dontfall 您好，我这边一直出问题，不知道怎么处理，可以请教一下不 out = self.unfold(x).view(b, self.g, self.gc, self.k ** 2, h, w) RuntimeError: shape '[1, 2, 24, 9, 128, 128]' is invalid for input of size 1769472 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 12 · 2021-04-21T13:56:08.000Z

@Dontfall mmcv库我安装了

Answer 13 · 2021-04-21T13:57:27.000Z

@Dontfall 安装的mmcv1.3.1的

Answer 14 · 2021-04-21T14:15:30.000Z

请参考 ultralytics/yolov5#2435

…

------------------ 原始邮件 ------------------ 发件人: "d-li14/involution" ***@***.***>; 发送时间: 2021年4月21日(星期三) 晚上9:18 ***@***.***>; ***@***.******@***.***>; 主题: Re: [d-li14/involution] bug (#30) @Dontfall 您好，我这边一直出问题，不知道怎么处理，可以请教一下不 out = self.unfold(x).view(b, self.g, self.gc, self.k ** 2, h, w) RuntimeError: shape '[1, 2, 24, 9, 128, 128]' is invalid for input of size 1769472 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 15 · 2021-04-22T02:41:24.000Z

@Dontfall 我看了一下不太明白下面这个是放在哪里
import torch
from models.experimental import Involution
m = Involution(ch=128, k=1, s=1) # ch_in, kernel, stride
x = torch.zeros(16, 128, 20, 20) # input
y = m(x) # forward

Answer 16 · 2021-04-22T02:44:52.000Z

https://github.com/ultralytics/yolov5/tree/involution

…

------------------ 原始邮件 ------------------ 发件人: "d-li14/involution" ***@***.***>; 发送时间: 2021年4月22日(星期四) 上午10:41 ***@***.***>; ***@***.******@***.***>; 主题: Re: [d-li14/involution] bug (#30) @Dontfall 我看了一下不太明白下面这个是放在哪里 import torch from models.experimental import Involution m = Involution(ch=128, k=1, s=1) # ch_in, kernel, stride x = torch.zeros(16, 128, 20, 20) # input y = m(x) # forward — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 17 · 2021-04-22T03:27:09.000Z

@Dontfall 嗯嗯，我看到这个项目了，我是想问一下，这里面experiental.py文件中创建了Involution类，但如果想用involution替换掉对应模块，把下面的替换进去是吗

m = Involution(ch=128, k=1, s=1) # ch_in, kernel, stride

x = torch.zeros(16, 128, 20, 20) # input

y = m(x) # forward

Answer 18 · 2021-04-22T03:29:30.000Z

是替换yolov5中的.yaml文件中的conv、focus等发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Thu,Apr 22,2021 11:27 AM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30)

Answer 19 · 2021-04-22T04:02:51.000Z

@Dontfall 嗯嗯，谢谢您，我在.yaml文件中修改了对应的部分，对应的在common.py文件中调用Involution类，但是报错
Traceback (most recent call last):
File "D:/win10 pytorch17/yolov5-involution/train.py", line 531, in
train(hyp, opt, device, tb_writer, wandb)
File "D:/win10 pytorch17/yolov5-involution/train.py", line 78, in train
model = Model(opt.cfg or ckpt['model'].yaml, ch=3, nc=nc, anchors=hyp.get('anchors')).to(device) # create
File "D:\win10 pytorch17\yolov5-involution\models\yolo.py", line 83, in init
self.model, self.save = parse_model(deepcopy(self.yaml), ch=[ch]) # model, savelist
File "D:\win10 pytorch17\yolov5-involution\models\yolo.py", line 240, in parse_model
m_ = nn.Sequential(*[m(*args) for _ in range(n)]) if n > 1 else m(*args) # module
File "D:\win10 pytorch17\yolov5-involution\models\common.py", line 118, in init
self.conv = Involution(c1 * 4, k, s)
NameError: name 'Involution' is not defined

Answer 20 · 2021-04-22T04:04:13.000Z

Involution大小写要对应发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Thu,Apr 22,2021 0:03 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30)

Answer 21 · 2021-04-22T04:04:48.000Z

@Dontfall 是对应的

Answer 22 · 2021-04-22T04:05:54.000Z

你仔细检查吧，没有对应，我也碰到过发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Thu,Apr 22,2021 0:05 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30)

Answer 23 · 2021-04-22T06:02:49.000Z

@Dontfall 您好，我检查了几遍没有存在不对应的问题，我觉得应该是我代码哪里的问题：假设我用involution替换focus模块中的Conv模块，做如下修改对吗？
1..yaml文件中：
backbone:

[from, number, module, args]

[[-1, 1, Focus, [64, 3]], # 0-P1/2
#[-1, 1, Conv, [128, 3, 2]], # 1-P2/4
[-1, 1, Conv, [128, 1, 1]],

common.py文件中
lass Focus(nn.Module):
# Focus wh information into c-space
def init(self, c1, c2, k=1, s=1, p=None, g=1, act=True): # ch_in, ch_out, kernel, stride, padding, groups
super(Focus, self).init()
# self.conv = Involution(c1, k, s)
#self.conv = Conv(c1 * 4, c2, k, s, p, g, act)
self.conv = Involution(c1 * 4, k, s)

除此之外还需要修改哪里吗？

Answer 24 · 2021-04-22T06:53:04.000Z

我用他的代码跑没有问题，只是根据报错调整了Conv的参数发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Thu,Apr 22,2021 2:03 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30) @Dontfall 您好，我检查了几遍没有存在不对应的问题，我觉得应该是我代码哪里的问题：假设我用involution替换focus模块中的Conv模块，做如下修改对吗？ 1..yaml文件中： backbone: [from, number, module, args] [[-1, 1, Focus, [64, 3]], # 0-P1/2 #[-1, 1, Conv, [128, 3, 2]], # 1-P2/4 [-1, 1, Conv, [128, 1, 1]], common.py文件中 lass Focus(nn.Module): # Focus wh information into c-space def init(self, c1, c2, k=1, s=1, p=None, g=1, act=True): # ch_in, ch_out, kernel, stride, padding, groups super(Focus, self).init() # self.conv = Involution(c1, k, s) #self.conv = Conv(c1 * 4, c2, k, s, p, g, act) self.conv = Involution(c1 * 4, k, s) 除此之外还需要修改哪里吗？ — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 25 · 2021-04-22T07:12:51.000Z

@Dontfall 好吧，可以帮忙看一下我是不是下面代码用的有问题 **1..yaml文件中：** backbone: [[-1, 1, Focus, [64, 3]], # 0-P1/2 #m = Involution(ch=128, k=1, s=1) # ch_in, kernel, stride替换 [-1, 1, Conv, [128, 1, 1]], [-1, 3, C3, [128]], **2.common.py文件中** class Focus(nn.Module): def init(self, c1, c2, k=1, s=1, p=None, g=1, act=True): super(Focus, self).init() #Involution替换Conv #self.conv = Conv(c1 * 4, c2, k, s, p, g, act) self.conv = Involution(c1 * 4, k, s) 这些有问题吗在 2021-04-22 14:53:22，"Dontfall" ***@***.***> 写道：我用他的代码跑没有问题，只是根据报错调整了Conv的参数发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Thu,Apr 22,2021 2:03 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30) @Dontfall 您好，我检查了几遍没有存在不对应的问题，我觉得应该是我代码哪里的问题：假设我用involution替换focus模块中的Conv模块，做如下修改对吗？ 1..yaml文件中： backbone: [from, number, module, args] [[-1, 1, Focus, [64, 3]], # 0-P1/2 #[-1, 1, Conv, [128, 3, 2]], # 1-P2/4 [-1, 1, Conv, [128, 1, 1]], common.py文件中 lass Focus(nn.Module): # Focus wh information into c-space def init(self, c1, c2, k=1, s=1, p=None, g=1, act=True): # ch_in, ch_out, kernel, stride, padding, groups super(Focus, self).init() # self.conv = Involution(c1, k, s) #self.conv = Conv(c1 * 4, c2, k, s, p, g, act) self.conv = Involution(c1 * 4, k, s) 除此之外还需要修改哪里吗？ — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 26 · 2021-04-22T09:56:20.000Z

您好，再次打扰一下，想问一下下面代码对应可以解释一下吗？对应放到哪里？十分谢谢 importtorchfrommodels.experimentalimportInvolutionm=Involution(ch=128, k=1, s=1) # ch_in, kernel, stridex=torch.zeros(16, 128, 20, 20) # inputy=m(x) # forward !gitcheckoutinvolution# Profilefromutils.torch_utilsimportprofilefrommodels.experimentalimportConv, Involutionm1=Conv(c1=128, c2=128, k=3) m2=torch.nn.Conv2d(128, 128, kernel_size=3) m3=Involution(ch=128, k=3, s=1) profile(x=torch.randn(16, 128, 80, 80), ops=[m1, m2, m3], n=100) 在 2021-04-22 14:53:22，"Dontfall" ***@***.***> 写道：我用他的代码跑没有问题，只是根据报错调整了Conv的参数发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Thu,Apr 22,2021 2:03 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30) @Dontfall 您好，我检查了几遍没有存在不对应的问题，我觉得应该是我代码哪里的问题：假设我用involution替换focus模块中的Conv模块，做如下修改对吗？ 1..yaml文件中： backbone: [from, number, module, args] [[-1, 1, Focus, [64, 3]], # 0-P1/2 #[-1, 1, Conv, [128, 3, 2]], # 1-P2/4 [-1, 1, Conv, [128, 1, 1]], common.py文件中 lass Focus(nn.Module): # Focus wh information into c-space def init(self, c1, c2, k=1, s=1, p=None, g=1, act=True): # ch_in, ch_out, kernel, stride, padding, groups super(Focus, self).init() # self.conv = Involution(c1, k, s) #self.conv = Conv(c1 * 4, c2, k, s, p, g, act) self.conv = Involution(c1 * 4, k, s) 除此之外还需要修改哪里吗？ — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 27 · 2021-04-22T10:01:18.000Z

你可以先用involution代替第一层，也就是focus后的conv，看看能不能成功发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Thu,Apr 22,2021 3:13 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30) @Dontfall 好吧，可以帮忙看一下我是不是下面代码用的有问题 **1..yaml文件中：** backbone: [[-1, 1, Focus, [64, 3]], # 0-P1/2 #m = Involution(ch=128, k=1, s=1) # ch_in, kernel, stride替换 [-1, 1, Conv, [128, 1, 1]], [-1, 3, C3, [128]], **2.common.py文件中** class Focus(nn.Module): def init(self, c1, c2, k=1, s=1, p=None, g=1, act=True): super(Focus, self).init() #Involution替换Conv #self.conv = Conv(c1 * 4, c2, k, s, p, g, act) self.conv = Involution(c1 * 4, k, s) 这些有问题吗在 2021-04-22 14:53:22，"Dontfall" ***@***.***> 写道：我用他的代码跑没有问题，只是根据报错调整了Conv的参数发自我的iPhone

------------------ Original ------------------ From: QY1994-0919 ***@***.***&gt; Date: Thu,Apr 22,2021 2:03 PM To: d-li14/involution ***@***.***&gt; Cc: Dontfall ***@***.***&gt;, Mention ***@***.***&gt; Subject: Re: [d-li14/involution] bug (#30) @Dontfall 您好，我检查了几遍没有存在不对应的问题，我觉得应该是我代码哪里的问题：假设我用involution替换focus模块中的Conv模块，做如下修改对吗？ 1..yaml文件中： backbone: [from, number, module, args] [[-1, 1, Focus, [64, 3]], # 0-P1/2 #[-1, 1, Conv, [128, 3, 2]], # 1-P2/4 [-1, 1, Conv, [128, 1, 1]], common.py文件中 lass Focus(nn.Module): # Focus wh information into c-space def init(self, c1, c2, k=1, s=1, p=None, g=1, act=True): # ch_in, ch_out, kernel, stride, padding, groups super(Focus, self).init() # self.conv = Involution(c1, k, s) #self.conv = Conv(c1 * 4, c2, k, s, p, g, act) self.conv = Involution(c1 * 4, k, s) 除此之外还需要修改哪里吗？ — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 28 · 2021-04-22T10:02:48.000Z

这个代码只是作者测试involution用的，不必看发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Thu,Apr 22,2021 5:56 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30) 您好，再次打扰一下，想问一下下面代码对应可以解释一下吗？对应放到哪里？十分谢谢 importtorchfrommodels.experimentalimportInvolutionm=Involution(ch=128, k=1, s=1) # ch_in, kernel, stridex=torch.zeros(16, 128, 20, 20) # inputy=m(x) # forward !gitcheckoutinvolution# Profilefromutils.torch_utilsimportprofilefrommodels.experimentalimportConv, Involutionm1=Conv(c1=128, c2=128, k=3) m2=torch.nn.Conv2d(128, 128, kernel_size=3) m3=Involution(ch=128, k=3, s=1) profile(x=torch.randn(16, 128, 80, 80), ops=[m1, m2, m3], n=100) 在 2021-04-22 14:53:22，"Dontfall" ***@***.***> 写道：我用他的代码跑没有问题，只是根据报错调整了Conv的参数发自我的iPhone

------------------ Original ------------------ From: QY1994-0919 ***@***.***&gt; Date: Thu,Apr 22,2021 2:03 PM To: d-li14/involution ***@***.***&gt; Cc: Dontfall ***@***.***&gt;, Mention ***@***.***&gt; Subject: Re: [d-li14/involution] bug (#30) @Dontfall 您好，我检查了几遍没有存在不对应的问题，我觉得应该是我代码哪里的问题：假设我用involution替换focus模块中的Conv模块，做如下修改对吗？ 1..yaml文件中： backbone: [from, number, module, args] [[-1, 1, Focus, [64, 3]], # 0-P1/2 #[-1, 1, Conv, [128, 3, 2]], # 1-P2/4 [-1, 1, Conv, [128, 1, 1]], common.py文件中 lass Focus(nn.Module): # Focus wh information into c-space def init(self, c1, c2, k=1, s=1, p=None, g=1, act=True): # ch_in, ch_out, kernel, stride, padding, groups super(Focus, self).init() # self.conv = Involution(c1, k, s) #self.conv = Conv(c1 * 4, c2, k, s, p, g, act) self.conv = Involution(c1 * 4, k, s) 除此之外还需要修改哪里吗？ — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 29 · 2021-04-22T10:09:04.000Z

下载这个代码库，只需要改yaml文件，具体怎么改，看报错，改个参数就行了发自我的iPhone

…

------------------ Original ------------------ From: 谢玮 ***@***.***> Date: Thu,Apr 22,2021 10:44 AM To: d-li14/involution ***@***.***> Subject: Re: [d-li14/involution] bug (#30) https://github.com/ultralytics/yolov5/tree/involution

------------------ 原始邮件 ------------------ 发件人: "d-li14/involution" ***@***.***>; 发送时间: 2021年4月22日(星期四) 上午10:41 ***@***.***>; ***@***.******@***.***>; 主题: Re: [d-li14/involution] bug (#30) @Dontfall 我看了一下不太明白下面这个是放在哪里 import torch from models.experimental import Involution m = Involution(ch=128, k=1, s=1) # ch_in, kernel, stride x = torch.zeros(16, 128, 20, 20) # input y = m(x) # forward — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 30 · 2021-04-22T10:59:02.000Z

太感谢您，我现在就是先修改的focus下面的conv层，另外就是common.py里面focus模块中的conv层是不需要做任何改变的是吧在 2021-04-22 18:09:25，"Dontfall" ***@***.***> 写道：下载这个代码库，只需要改yaml文件，具体怎么改，看报错，改个参数就行了发自我的iPhone

…

------------------ Original ------------------ From: 谢玮 ***@***.***> Date: Thu,Apr 22,2021 10:44 AM To: d-li14/involution ***@***.***> Subject: Re: [d-li14/involution] bug (#30) https://github.com/ultralytics/yolov5/tree/involution

------------------ 原始邮件 ------------------ 发件人: "d-li14/involution" ***@***.***>; 发送时间: 2021年4月22日(星期四) 上午10:41 ***@***.***>; ***@***.******@***.***>; 主题: Re: [d-li14/involution] bug (#30) @Dontfall 我看了一下不太明白下面这个是放在哪里 import torch from models.experimental import Involution m = Involution(ch=128, k=1, s=1) # ch_in, kernel, stride x = torch.zeros(16, 128, 20, 20) # input y = m(x) # forward — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 31 · 2021-04-22T11:00:35.000Z

只需要改配置yaml，其他的不用发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Thu,Apr 22,2021 6:59 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30) 太感谢您，我现在就是先修改的focus下面的conv层，另外就是common.py里面focus模块中的conv层是不需要做任何改变的是吧在 2021-04-22 18:09:25，"Dontfall" ***@***.***> 写道：下载这个代码库，只需要改yaml文件，具体怎么改，看报错，改个参数就行了发自我的iPhone

------------------ Original ------------------ From: 谢玮 ***@***.***&gt; Date: Thu,Apr 22,2021 10:44 AM To: d-li14/involution ***@***.***&gt; Subject: Re: [d-li14/involution] bug (#30) https://github.com/ultralytics/yolov5/tree/involution

------------------ 原始邮件 ------------------ 发件人: "d-li14/involution" ***@***.***&gt;; 发送时间:&nbsp;2021年4月22日(星期四) 上午10:41 ***@***.***&gt;; ***@***.******@***.***&gt;; 主题:&nbsp;Re: [d-li14/involution] bug (#30) @Dontfall 我看了一下不太明白下面这个是放在哪里 import torch from models.experimental import Involution m = Involution(ch=128, k=1, s=1) # ch_in, kernel, stride x = torch.zeros(16, 128, 20, 20) # input y = m(x) # forward — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 32 · 2021-04-22T11:19:04.000Z

哦哦，太谢谢您，还有我修改yolov5s.yaml中的focus下的conv层，这样没错吧[-1, 1, Involution, [128, 1, 1]], 在 2021-04-22 19:00:55，"Dontfall" ***@***.***> 写道：只需要改配置yaml，其他的不用发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Thu,Apr 22,2021 6:59 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30) 太感谢您，我现在就是先修改的focus下面的conv层，另外就是common.py里面focus模块中的conv层是不需要做任何改变的是吧在 2021-04-22 18:09:25，"Dontfall" ***@***.***> 写道：下载这个代码库，只需要改yaml文件，具体怎么改，看报错，改个参数就行了发自我的iPhone

------------------ Original ------------------ From: 谢玮 ***@***.***&gt; Date: Thu,Apr 22,2021 10:44 AM To: d-li14/involution ***@***.***&gt; Subject: Re: [d-li14/involution] bug (#30) https://github.com/ultralytics/yolov5/tree/involution

------------------ 原始邮件 ------------------ 发件人: "d-li14/involution" ***@***.***&gt;; 发送时间:&nbsp;2021年4月22日(星期四) 上午10:41 ***@***.***&gt;; ***@***.******@***.***&gt;; 主题:&nbsp;Re: [d-li14/involution] bug (#30) @Dontfall 我看了一下不太明白下面这个是放在哪里 import torch from models.experimental import Involution m = Involution(ch=128, k=1, s=1) # ch_in, kernel, stride x = torch.zeros(16, 128, 20, 20) # input y = m(x) # forward — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 33 · 2021-04-22T11:20:44.000Z

我的报错是需要改成32，其他随意发自我的iPhone

…

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Thu,Apr 22,2021 7:19 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30)

Answer 34 · 2021-04-29T09:31:18.000Z

请问，你的问题解决了吗？发自我的iPhone

…

------------------ Original ------------------ From: 谢玮 ***@***.***> Date: Thu,Apr 22,2021 7:00 PM To: reply+AGY3LYYI24DFZNTDJNKPFLN6RU4AVEVBNHHDFJ5D2E ***@***.***> Subject: Re: [d-li14/involution] bug (#30) 只需要改配置yaml，其他的不用发自我的iPhone

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Thu,Apr 22,2021 6:59 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30) 太感谢您，我现在就是先修改的focus下面的conv层，另外就是common.py里面focus模块中的conv层是不需要做任何改变的是吧在 2021-04-22 18:09:25，"Dontfall" ***@***.***> 写道：下载这个代码库，只需要改yaml文件，具体怎么改，看报错，改个参数就行了发自我的iPhone

------------------ Original ------------------ From: 谢玮 ***@***.***&gt; Date: Thu,Apr 22,2021 10:44 AM To: d-li14/involution ***@***.***&gt; Subject: Re: [d-li14/involution] bug (#30) https://github.com/ultralytics/yolov5/tree/involution

------------------ 原始邮件 ------------------ 发件人: "d-li14/involution" ***@***.***&gt;; 发送时间:&nbsp;2021年4月22日(星期四) 上午10:41 ***@***.***&gt;; ***@***.******@***.***&gt;; 主题:&nbsp;Re: [d-li14/involution] bug (#30) @Dontfall 我看了一下不太明白下面这个是放在哪里 import torch from models.experimental import Involution m = Involution(ch=128, k=1, s=1) # ch_in, kernel, stride x = torch.zeros(16, 128, 20, 20) # input y = m(x) # forward — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 35 · 2021-04-29T09:55:37.000Z

我可以把我调好的代码发给你，做参考发自我的iPhone

…

------------------ Original ------------------ From: 谢玮 ***@***.***> Date: Thu,Apr 29,2021 5:31 PM To: reply+AGY3LYYI24DFZNTDJNKPFLN6RU4AVEVBNHHDFJ5D2E ***@***.***> Subject: Re: [d-li14/involution] bug (#30) 请问，你的问题解决了吗？发自我的iPhone

------------------ Original ------------------ From: 谢玮 ***@***.***> Date: Thu,Apr 22,2021 7:00 PM To: reply+AGY3LYYI24DFZNTDJNKPFLN6RU4AVEVBNHHDFJ5D2E ***@***.***> Subject: Re: [d-li14/involution] bug (#30) 只需要改配置yaml，其他的不用发自我的iPhone

------------------ Original ------------------ From: QY1994-0919 ***@***.***> Date: Thu,Apr 22,2021 6:59 PM To: d-li14/involution ***@***.***> Cc: Dontfall ***@***.***>, Mention ***@***.***> Subject: Re: [d-li14/involution] bug (#30) 太感谢您，我现在就是先修改的focus下面的conv层，另外就是common.py里面focus模块中的conv层是不需要做任何改变的是吧在 2021-04-22 18:09:25，"Dontfall" ***@***.***> 写道：下载这个代码库，只需要改yaml文件，具体怎么改，看报错，改个参数就行了发自我的iPhone

------------------ Original ------------------ From: 谢玮 ***@***.***&gt; Date: Thu,Apr 22,2021 10:44 AM To: d-li14/involution ***@***.***&gt; Subject: Re: [d-li14/involution] bug (#30) https://github.com/ultralytics/yolov5/tree/involution

------------------ 原始邮件 ------------------ 发件人: "d-li14/involution" ***@***.***&gt;; 发送时间:&nbsp;2021年4月22日(星期四) 上午10:41 ***@***.***&gt;; ***@***.******@***.***&gt;; 主题:&nbsp;Re: [d-li14/involution] bug (#30) @Dontfall 我看了一下不太明白下面这个是放在哪里 import torch from models.experimental import Involution m = Involution(ch=128, k=1, s=1) # ch_in, kernel, stride x = torch.zeros(16, 128, 20, 20) # input y = m(x) # forward — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 36 · 2022-01-18T03:00:08.000Z

您好，请问yolo.py里要怎么改呢，跑yolov5一直没跑通

Answer 37 · 2022-01-18T03:03:51.000Z

@luyunfighting 你好，我之前跑通了cpu的版本，cuda的版本没有跑通，所以放弃了，祝你成功！