mirrors
/
yolov7
mirror of https://github.com/WongKinYiu/yolov7.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Commit Graph

  • 9203361e11 Deleting Projects files Gorkem KOCATURK 2025-04-16 11:13:25 +0300
  • 5f19b9e76b Readme file changing, test bash script changing, Folder order changing (PROJECTS/DATA and PROJECTS/RESULTS) Gorkem KOCATURK 2025-04-15 15:08:03 +0300
  • d0dbbb5ef5 doc(Dockerfile): add comment for accure training barisx 2025-04-14 19:02:17 +0300
  • 491711c114 feat(Dockerfile): add preweights for training barisx 2025-04-14 18:41:57 +0300
  • f6ddbc6ee2 doc(Readme), feat(run-docker-container): update readme cuda error, add docker script barisx 2025-04-14 18:38:29 +0300
  • 3d20bb578c minor fix barisx 2025-04-14 10:44:27 +0300
  • 940c31f5a6 minor fix barisx 2025-04-14 10:08:41 +0300
  • 93139c0425 add some comments and update readme barisx 2025-04-12 22:37:55 +0300
  • 27f2d413a1 init barisx 2025-04-12 19:51:58 +0300
  • af63103b71
         Update loss.py training error fix Barış Şenyerli 2025-04-12 19:47:39 +0300
  • 25b959e9b4
         Merge b00bc47dac into a207844b1c hanochk 2025-04-08 16:03:11 +0800
  • b00bc47dac OHEM : periodic mechanism best_model weights most the mAP@50 summary reposrter: merge_results.py of best model remove aggresive truncated objects center crop from training list hanoch3 2025-04-08 10:32:58 +0300
  • 27e5ea3f8e
         Update train.py SouhailaK 2025-03-31 18:27:14 +0200
  • 4576d5166a Se crea la ruta dataset SouhailaK 2025-03-31 15:44:27 +0200
  • 1669991d6a Cambio en la ruta de data.yaml SouhailaK 2025-03-31 14:40:41 +0200
  • 6a58e7f4da Se cambian rutas en data.yaml SouhailaK 2025-03-31 13:28:03 +0200
  • 71a6c68d24 Se modifican los archivos de configuración .yaml SouhailaK 2025-03-31 12:45:40 +0200
  • 83d9d02998 OHEM : datasets.py-resample indices according to the top-k Fix : random seed test.py : accumulate loss per image hard to run natchwise hanoch3 2025-03-31 10:50:00 +0300
  • 565c17e6a4 incoroperating CrossEntropy() loss with designated FocalLoss. w/ weighted loss Adding Asymentrical loss for Multi-label hasn;t tested yet! hanoch3 2025-03-30 11:14:54 +0300
  • 0a12c34633 Remove center_crop /s that had samll truncated objects +missing png/s and other tiff from the final annotation list : tir_tiff_seq_png_3_class_fixed_whether_copied_dataset_label.xlsx hanoch3 2025-03-24 16:24:29 +0200
  • 2cc7e05cc8 Update validation set : Test36B 13/11 training-set Test36B 12/11 +missing png/s and other tiff from the final annotation list : tir_tiff_seq_png_3_class_fixed_whether_copied_dataset_label.xlsx hanoch3 2025-03-24 14:57:46 +0200
  • c90fa54f1e Update validation set : Test36B 13/11 training-set Test36B 12/11 +missing png/s and other tiff from the final annotation list : tir_tiff_seq_png_3_class_fixed_whether_copied_dataset_label.xlsx hanoch3 2025-03-24 14:57:24 +0200
  • 4fdfef3ffe add missing png files to the training set training_set_yolov7941_w_center_roi_no_swiss_no_38B_20b_22c_n_png_train_3_cls_jan_25.txt hanoch3 2025-03-23 17:45:04 +0200
  • ab6b4bd341 BCE loss with weight vector - fix Trigger when AP fall hanoch3 2025-03-23 16:05:37 +0200
  • 1fcfbb6f5b
         Merge 8a5d9730ab into a207844b1c Chakra Timilsina 2025-03-18 14:21:54 +0545
  • 8a5d9730ab fix:touch load required weight only false Chakra Timilsina 2025-03-18 14:14:39 +0545
  • 7fbdec325e
         Merge dc195aa6f9 into a207844b1c Roman KC 2025-03-16 12:22:29 +0545
  • dc195aa6f9
         fix: pythorch 2.6 updates for all Roman KC 2025-03-16 12:14:34 +0545
  • 41da49e028
         fix: torch.load requires weights_only Roman KC 2025-03-16 12:03:25 +0545
  • b4f20a583b
         Merge e89974e4c9 into a207844b1c Shuqi Xiao 2025-03-05 17:26:19 +0800
  • e89974e4c9 Fix for torch 2.6 Shuqi XIAO 2025-03-05 17:25:51 +0800
  • 8766685f94 Fix image_data_creator.py : truncated objects in lower part of image: fix effective height computation datasets.py : Remove empty annotated image entries weights file moved to HDD hanoch 2025-02-27 12:45:47 +0200
  • 52320239d4 ONNX reorder output labels No. by pre-NMs processing, model name over PR hanoch 2025-02-20 10:38:08 +0200
  • 78cd0cb8f6 detect.py: support of txt file list FL with configured weighted loss : alpha test.py: tSNE over the cropped bbox features hanoch 2025-02-16 10:05:06 +0200
  • 6b8d33fa31 detect.py: support of txt file list FL with configured weighted loss : alpha test.py: tSNE over the cropped bbox features hanoch 2025-02-16 10:03:45 +0200
  • 30c9907884 gignore add xml hanoch 2025-02-03 15:20:20 +0200
  • d00653795f True prediction equal the same range cell as GT. Fix bug where no preds but GT exist. whether statistics, ONNX save pre-processed file, hanoch 2025-02-03 15:14:16 +0200
  • 4c0190e554 Test51a_Test40A_test_set.txt : remove truncated/small objects on the right crop. Whether condition based breakdown hanoch 2025-01-30 13:34:19 +0200
  • b8e71e931a adding debug filename of each prediction/GT. Legend to P/R per bin @ threshold hanoch 2025-01-29 15:46:14 +0200
  • f4266d2161 adding debug filename of each prediction/GT. Legend to P/R per bin @ threshold hanoch 2025-01-29 15:35:29 +0200
  • e88bd2fd10 Fix bug : np.array.size() condition, pre range iteration, clear all range p/r stats, height based range, p/r per range per class hanoch 2025-01-28 15:00:05 +0200
  • 59b7be6b91 Fix bug hanoch 2025-01-27 14:03:51 +0200
  • 085ee89bfc mAP breakdown sensor/time in day/ mAP per range : Dataset : add csv with metadata of each image, caching labels omit label w/ class no not in the list. range_bar plot, add plot class support. Onnx RT: adding noise to dataset, dataset with ext. noise hanoch 2025-01-27 10:10:38 +0200
  • 0f8f0772fe Onnx for Prediction w/o GT hanoch 2025-01-13 15:44:41 +0200
  • 973e43446d Onnx for Prediction w/o GT hanoch 2025-01-13 15:40:19 +0200
  • ab1f39e24a Overfit 640 settings commit before checkout old hanoch 2025-01-08 11:27:05 +0200
  • b8a2a10ffe Union train list mp4/tiff/seq hanoch 2025-01-07 17:05:07 +0200
  • 772cee9084 datasets.py: 768x1024, random padding, roi based scaling . still open issue the prediction over the random padding area should be omitted from calculation hanoch 2025-01-06 17:28:32 +0200
  • 221429a81b datasets.py: 768x1024, random padding, roi based scaling . still open issue the prediction over the random padding area should be omitted from calculation hanoch 2025-01-06 17:26:49 +0200
  • f1ff652dcf
         Merge f263929660 into a207844b1c tejareddy 2025-01-05 21:27:29 +0530
  • f263929660 Fixed import path issue in experimental.py TejaReddy 2025-01-05 20:50:16 +0530
  • 76e6315e99 Onnx model P/R curve, fixed, 1st IOU criterion per image than mAP overall. hanoch 2024-12-24 12:32:20 +0200
  • 54a0cc226b NMS-ORT built in Onnx model P/R curve fits theory. hanoch 2024-12-23 17:03:39 +0200
  • ee65c332a3 Fix mAP over Onnx wrapper for old TIR model hanoch 2024-12-16 18:04:06 +0200
  • 9aec3afb56 adapt to TIR env export.py. Fix threshold computation for Precision recall. test.py : adding preds/gts per image to csv for error analysis hanoch 2024-12-16 17:02:01 +0200
  • 5d99ec8b2c Onnx model OnnxRT for eval old TIR model, validate annotations over PNG hanoch 2024-12-11 17:28:56 +0200
  • abe52501fe comments hanoch 2024-12-05 17:17:47 +0200
  • 6421910e14 center_roi hanoch 2024-12-03 17:02:27 +0200
  • 542cc99ada PNG/MP$ dataset __getitem__() adaptation. Fix paste_in loading many patches by avoiding normalizing the PNG hanoch 2024-12-01 17:37:59 +0200
  • 705aa5ea87 Fix random padding disabled on mosaic that ovverides ROI hanoch 2024-11-26 14:53:41 +0200
  • 6ab8732fbe hyp effective in ClearML, center_roi train/val/test list hanoch 2024-11-26 12:05:31 +0200
  • 13a313c649 Gamma augmentation robustness to Random padding, clip 0<image<1, new training set excluding Test19G swiss with bad NUC compensation hanoch 2024-11-26 09:51:51 +0200
  • c6e82c42f6 Random padding- post processing w/ all augmentations permutations.standartization Stability increasing epsilon to 1e-5. Random padding. Replace Albumentation Gamma aug by skimage one more stable hanoch 2024-11-25 15:51:22 +0200
  • 6c577e95c9 New Train val set split Swiss-snow-2018 to training(test19g_sy_rd_ch) val:Test46c_ml_rd_il hanoch 2024-11-19 13:35:43 +0200
  • 2de1a2593c Random padding, Val/train new split append to previous list hanoch 2024-11-17 15:10:26 +0200
  • 4afc52b1cd Filter validation set from RF leakage USA. Add tag per PR curve hanoch 2024-11-13 16:49:31 +0200
  • 7391d5fc94 Clean bad recordings from France hanoch 2024-11-11 11:22:01 +0200
  • 0dbd6c532a
         Merge dcb84232e9 into a207844b1c Mazen-Embaby 2024-11-11 16:03:42 +0800
  • dcb84232e9
         Merge branch 'WongKinYiu:main' into main Mazen-Embaby 2024-11-11 12:01:59 +0400
  • ddeff5a8fc increased training/val set 24500/5346 hanoch 2024-11-10 17:36:16 +0200
  • 5e5dfde1f1 Fix rnd perspective fillng the padded pixels by TIR mean value to reduce edges artifacts. mAP per med/large exclusively hanoch 2024-11-06 16:54:53 +0200
  • b2bc3773a1 hyp : scaling and mosaic w/ scliang before mosaic hanoch 2024-11-04 10:13:35 +0200
  • 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) hanoch 2024-11-04 10:08:35 +0200
  • c040cf8265 enable random perspective hanoch 2024-10-31 11:13:08 +0200
  • e16bd3ffad random seed, hyp for random perspective hanoch 2024-10-30 15:43:20 +0200
  • 5639572615 plot PR curve thresholds and R@P remove no_aug hanoch 2024-10-27 14:49:26 +0200
  • e2f0c15a00 remove torch anomaly() hanoch 2024-10-22 10:32:51 +0300
  • f366ed0f30 stability of denum in data norm/standard hanoch 2024-10-21 15:57:07 +0300
  • 5f8046d05f claer ml debug hanoch 2024-10-20 17:08:46 +0300
  • 102d731395 claer ml debug hanoch 2024-10-20 16:51:29 +0300
  • 60ef66ce19 train.py save runs to incremental anywhere : /mnt/Data/<user> hanoch 2024-10-20 15:27:57 +0300
  • d166717631 CLearML connect config hanoch 2024-10-20 14:40:49 +0300
  • db5033a5f4 Milesone MaP_person = 82.5% gradient clipping with optimizer scaler CLearML connect config hanoch 2024-10-20 10:46:21 +0300
  • 3da5c91239 CLearML connect config hanoch 2024-10-15 15:22:54 +0300
  • 890f748e16 CLearML connect config hanoch 2024-10-15 14:50:31 +0300
  • 65d0872a36 CLearML connect config hanoch 2024-10-15 14:40:02 +0300
  • 14fd22ef35 train/val/test data split hanoch 2024-10-14 14:18:40 +0300
  • 292ca3b308 train/val/test data split hanoch 2024-10-14 14:17:35 +0300
  • e7c36bab68 modify TIR channel expansion to be w/o augmentation hanoch 2024-10-14 14:02:24 +0300
  • abdcce0e70 Test : medium size try/excepton hanoch 2024-10-14 10:32:25 +0300
  • 9b7cc0c659
         Update requirements.txt Abhijeet Ghadiyaram 2024-09-23 14:26:20 -0500
  • 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 hanoch 2024-09-23 11:23:13 +0300
  • c3276b3de9 Test/ train/detect :Scaling and normalization. hanoch 2024-09-15 09:20:39 +0300
  • d558bf24de
         Merge 40c8731b50 into 44f30af0da Nomaanrizvi 2024-09-08 11:44:16 +0530
  • 40c8731b50 fixed the plot.py and dataloader.py files Nomaanrizvi 2024-09-08 11:41:09 +0530
  • 22a3f8b979
         Merge b389738a8a into 44f30af0da Nomaanrizvi 2024-09-05 17:45:17 +0530
  • b389738a8a Fixed np.int issue in u7: corrected model training bug by changing the data-type of the batch index(line 488) Nomaanrizvi 2024-09-05 17:34:24 +0530
  • 02f785d2dc check train dataset objects class id violation. n-ch in as param, hanoch 2024-08-27 13:58:16 +0300
  • 18086f7c35 comment hanoch 2024-08-19 15:56:20 +0300
  • c8c3363926 Fixed randomizer gets integers only! hanoch 2024-08-15 15:46:47 +0300