mirrors
/
yolov7
mirror of https://github.com/WongKinYiu/yolov7.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
167 Commits
9 Branches
1 Tag
228 MiB
Commit Graph

31 Commits (6c577e95c9897d92d8ae8d82ec4f409627b03426)

Author SHA1 Message Date
hanoch 6c577e95c9 New Train val set split Swiss-snow-2018 to training(test19g_sy_rd_ch) val:Test46c_ml_rd_il 2024-11-19 13:35:43 +02:00
hanoch 5e5dfde1f1 Fix rnd perspective fillng the padded pixels by TIR mean value to reduce edges artifacts. 
mAP per med/large exclusively
 2024-11-06 16:54:53 +02:00
hanoch 15755c31b9 dataset.py: align test() loader to train 
scaling before mosaic with predefined filling_value according to img precision chnaged by scaling before mosaic. filling uint16 with image.mean()
detect.py : align to latest changes as test.py
test.py : support in large object size metrices stats_all_large
plots.py : plot detections with th>0.1 (not 0.25)
 2024-11-04 10:08:35 +02:00
hanoch e16bd3ffad random seed, hyp for random perspective 2024-10-30 15:43:20 +02:00
hanoch e2f0c15a00 remove torch anomaly() 2024-10-22 10:32:51 +03:00
hanoch f366ed0f30 stability of denum in data norm/standard 2024-10-21 15:57:07 +03:00
hanoch 5f8046d05f claer ml debug 2024-10-20 17:08:46 +03:00
hanoch 102d731395 claer ml debug 2024-10-20 16:51:29 +03:00
hanoch 60ef66ce19 train.py save runs to incremental anywhere : /mnt/Data/<user> 2024-10-20 15:27:57 +03:00
hanoch d166717631 CLearML connect config 2024-10-20 14:40:49 +03:00
hanoch db5033a5f4 Milesone MaP_person = 82.5% 
gradient clipping with optimizer scaler
CLearML connect config
 2024-10-20 10:46:21 +03:00
hanoch 3da5c91239 CLearML connect config 2024-10-15 15:22:54 +03:00
hanoch 890f748e16 CLearML connect config 2024-10-15 14:50:31 +03:00
hanoch 65d0872a36 CLearML connect config 2024-10-15 14:40:02 +03:00
hanoch e7c36bab68 modify TIR channel expansion to be w/o augmentation 2024-10-14 14:02:24 +03:00
hanoch abdcce0e70 Test : medium size try/excepton 2024-10-14 10:32:25 +03:00
hanoch e5654f43f0 tir channel expansion, adapt mosaic to TIR or any other than UINT8/RGB , inversion aug is dtype dependant. test.py : verbose save class mAP , mAP per size. ClearML support 2024-09-23 11:23:13 +03:00
hanoch c3276b3de9 Test/ train/detect :Scaling and normalization. 2024-09-15 09:20:39 +03:00
hanoch 02f785d2dc check train dataset objects class id violation. n-ch in as param, 2024-08-27 13:58:16 +03:00
hanoch 18086f7c35 comment 2024-08-19 15:56:20 +03:00
hanoch c8c3363926 Fixed randomizer gets integers only! 2024-08-15 15:46:47 +03:00
hanoch c1a580e621 comment 2024-08-15 15:40:01 +03:00
hanoch b82fdc0574 Add support using the "parent" directory in data.yaml as parent for the images folder 2024-08-08 14:44:50 +03:00
hanoch 5bb345f112 fix annotation txtx file creation 2024-08-07 18:03:48 +03:00
Kıvanç Tezören 55b90e1119
Add option to use YOLOv5 AP metric (#775) 
* Add YOLOv5 metric option

* Inform if using v5 metric
 2022-09-16 21:14:01 +03:00
AlexeyAB84 36ce6b2087 Use compute_loss_ota() if there is not loss_ota param or loss_ota==1 2022-08-16 02:10:07 +03:00
AlexeyAB84 711a16ba57 Added param loss_ota for hyp.yaml, to disable OTA for faster training 2022-08-09 07:48:28 +03:00
Mohammad Khoshbin b8956dd5a5
fix training with frozen layers (#378) 2022-08-02 17:55:28 +03:00
Mohammad Khoshbin cc42a206ef
fix cpu-only training CUDA error (#369) 2022-07-30 18:43:18 +03:00
Dhia_Oussayed 0d882e553e
Fixed a bug for Hyperparameters Evolution. (#344) 
* Update hyp.scratch.custom.yaml

anchors parameter missing, evolution won't start without.

* Update train.py

Updated the hyperparameter evolution metadata variable to match the hyp.yaml files for the evolution to run successfully.

* Update hyp.scratch.p5.yaml

Added the anchors parameter, evolution don't start without it.

* Update hyp.scratch.p6.yaml

added the anchors parameter for the hyperparameter evolution.

* Update hyp.scratch.tiny.yaml

added the anchors parameter for the hyperparameters evolution

* Update hyp.scratch.custom.yaml

* Update hyp.scratch.tiny.yaml

* Update hyp.scratch.p5.yaml

* Update hyp.scratch.p6.yaml

* Update train.py

* Update train_aux.py
 2022-07-28 22:09:52 +03:00
Kin-Yiu, Wong a1c6e04c7c
main code 
code for inference
 2022-07-06 23:23:27 +08:00