evan-kolberg commented 1 year ago

Main issue ↓

The program is running without producing any visual representation (output) of what is happening. The input image is the same as the output and even after specifying to save txt and conf, nothing is stored in the labels directory.

Attempt 1 ↓

Command used: python detect.py --source city-people.jpg

Windows PowerShell
Copyright (C) Microsoft Corporation. All rights reserved.

Install the latest PowerShell for new features and improvements! https://aka.ms/PSWindows

PS C:\Users\ekpro\VS Code Projects\CAPTCHA-slayer\yolov7> & c:/Users/ekpro/venvs✨/yolov7/Scripts/Activate.ps1
(yolov7) PS C:\Users\ekpro\VS Code Projects\CAPTCHA-slayer\yolov7> python detect.py --source city-people.jpg
Namespace(weights='yolov7.pt', source='city-people.jpg', img_size=640, conf_thres=0.25, iou_thres=0.45, device='', view_img=False, save_txt=False, save_conf=False, nosave=False, classes=None, agnostic_nms=False, augment=False, update=False, project='runs/detect', name='exp', exist_ok=False, no_trace=False)
YOLOR  2022-8-27 torch 1.12.1+cu116 CUDA:0 (NVIDIA GeForce GTX 1660 Ti, 6143.6875MB)

Fusing layers...
RepConv.fuse_repvgg_block
RepConv.fuse_repvgg_block
RepConv.fuse_repvgg_block
Model Summary: 306 layers, 36905341 parameters, 6652669 gradients
 Convert model to Traced-model...
 traced_script_module saved! 
 model is traced! 

C:\Users\ekpro\venvs✨\yolov7\lib\site-packages\torch\functional.py:478: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at  ..\aten\src\ATen\native\TensorShape.cpp:2895.)
  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
Done. (89.0ms) Inference, (0.0ms) NMS
 The image with the result is saved in: runs\detect\exp\city-people.jpg
Done. (0.395s)

Input image ↓ city-people

Output image in runs\detect\exp ↓ city-people

As pictured above, no changes were saved onto the output image

Attempt 2 ↓

Command used: python detect.py --weights yolov7.pt --source city-people.jpg --view-img --save-txt --save-conf --conf-thres 0.10

Windows PowerShell
Copyright (C) Microsoft Corporation. All rights reserved.

Install the latest PowerShell for new features and improvements! https://aka.ms/PSWindows

PS C:\Users\ekpro\VS Code Projects\CAPTCHA-slayer\yolov7> & c:/Users/ekpro/venvs✨/yolov7/Scripts/Activate.ps1
(yolov7) PS C:\Users\ekpro\VS Code Projects\CAPTCHA-slayer\yolov7> python detect.py --weights yolov7.pt --source city-people.jpg --view-img --save-txt --save-conf --conf-thres 0.10
Namespace(weights=['yolov7.pt'], source='city-people.jpg', img_size=640, conf_thres=0.1, iou_thres=0.45, device='', view_img=True, save_txt=True, save_conf=True, nosave=False, classes=None, agnostic_nms=False, augment=False, update=False, project='runs/detect', name='exp', exist_ok=False, no_trace=False)
YOLOR  2022-8-27 torch 1.12.1+cu116 CUDA:0 (NVIDIA GeForce GTX 1660 Ti, 6143.6875MB)

Fusing layers... 
RepConv.fuse_repvgg_block
RepConv.fuse_repvgg_block
RepConv.fuse_repvgg_block
Model Summary: 306 layers, 36905341 parameters, 6652669 gradients
 Convert model to Traced-model... 
 traced_script_module saved! 
 model is traced! 

C:\Users\ekpro\venvs✨\yolov7\lib\site-packages\torch\functional.py:478: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at  ..\aten\src\ATen\native\TensorShape.cpp:2895.)
  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
Done. (88.0ms) Inference, (1.0ms) NMS
 The image with the result is saved in: runs\detect\exp2\city-people.jpg
Done. (0.416s)

Nothing was saved in labels and no changes were made to the city-people.jpg

Situation conditions ↓

Git cloned Saturday, August 27
1660 ti GPU with CUDA 11.7 and PyTorch 1.12.1+cu116
Using vscode server on Win11 machine over LAN controlled by a Mac (ssh into pc)
Python version

Running on a Windows 11 machine ↓

Host Name:                 WINDOWS-11-RIG
OS Name:                   Microsoft Windows 11 Pro
OS Version:                10.0.22000 N/A Build 22000
OS Manufacturer:           Microsoft Corporation
OS Configuration:          Standalone Workstation
OS Build Type:             Multiprocessor Free
Registered Owner:          **********************
Registered Organization:   N/A
Product ID:                00330-80000-00000-AA329
Original Install Date:     12/11/2021, 12:31:08 AM
System Boot Time:          8/28/2022, 7:01:47 PM
System Manufacturer:       System manufacturer
System Model:              System Product Name
System Type:               x64-based PC
Processor(s):              1 Processor(s) Installed.
                       [01]: AMD64 Family 23 Model 113 Stepping 0 AuthenticAMD ~3593 Mhz
BIOS Version:              American Megatrends Inc. 2802, 10/21/2020
Windows Directory:         C:\WINDOWS
System Directory:          C:\WINDOWS\system32
Boot Device:               \Device\HarddiskVolume1
System Locale:             en-us;English (United States)
Input Locale:              N/A
Time Zone:                 (UTC-05:00) Eastern Time (US & Canada)
Total Physical Memory:     16,298 MB
Available Physical Memory: 12,533 MB
Virtual Memory: Max Size:  21,162 MB
Virtual Memory: Available: 14,849 MB
Virtual Memory: In Use:    6,313 MB
Page File Location(s):     C:\pagefile.sys
Domain:                    WORKGROUP
Logon Server:              N/A
Hotfix(s):                 4 Hotfix(s) Installed.
Network Card(s):           5 NIC(s) Installed.
                       [01]: Realtek Gaming GbE Family Controller
                             Connection Name: Ethernet
                             Status:          Media disconnected
                       [02]: Intel(R) Wireless-AC 9260 160MHz
                             Connection Name: Wi-Fi
                             DHCP Enabled:    Yes
                             DHCP Server:     192.168.1.1
                             IP address(es)
                       [03]: Bluetooth Device (Personal Area Network)
                             Connection Name: Bluetooth Network Connection
                             Status:          Media disconnected
                       [04]: TAP-ProtonVPN Windows Adapter V9
                             Connection Name: Local Area Connection
                             Status:          Media disconnected
                       [05]: Wintun Userspace Tunnel
                             Connection Name: ProtonVPN TUN
                             Status:          Media disconnected
Hyper-V Requirements:      A hypervisor has been detected. Features required for Hyper-V will not be displayed.

Attempt 3 ↓

Command: python detect.py --weights yolov7.pt --conf 0.25 --img-size 640 --source inference/images/horses.jpg

Instead of running detect.py using my own image, I used the default image that came with the repo and still reproduced the same issue.

Output image ↓

vishnuvskvkl commented 1 year ago

I'm facing similar issue even if I downgrade the cuda version to 11.3. Even though it works fine when using the CPU version

evan-kolberg commented 1 year ago

Command: python detect.py --source city-people.jpg --device cpu Same with me. I am able to get a visual by using the cpu, but it isn't exactly ideal since I have a perfectly fine CUDA gpu that can be used.

WongKinYiu commented 1 year ago

Please force set half to False if your GPU does not have tensor core.

evan-kolberg commented 1 year ago

Setting the var half to False on line 31 worked! Thank you!

WongKinYiu / yolov7

Program "detect.py" won't produce visual output using a GPU without tensor cores [solved] #646

Main issue ↓

Attempt 1 ↓

As pictured above, no changes were saved onto the output image

Attempt 2 ↓

Nothing was saved in labels and no changes were made to the city-people.jpg

Situation conditions ↓

Attempt 3 ↓

Instead of running detect.py using my own image, I used the default image that came with the repo and still reproduced the same issue.