Maizai

使用自定义数据集，loss_bbox: 0.00 loss_dfl: 0.00，网上的方法，添加类别，修改base搜没有解决。

2

yolo_world_l_dual_vlpan_2e-4_80e_8gpus_finetune_coco.py文件，我需要微调 _base_ = ( '../../third_party/mmyolo/configs/yolov8/yolov8_l_syncbn_fast_8xb16-500e_coco.py') custom_imports = dict( imports=['yolo_world'], allow_failed_imports=False) # hyper-parameters num_classes = 2 num_training_classes = 2 max_epochs = 1 # Maximum training epochs close_mosaic_epochs = 10 save_epoch_intervals =...

issa@917917:~/ax-pipeline/build$ make -j12 -- CMAKE_INSTALL_PREFIX = /home/issa/ax-pipeline/build/install -- SET AXERA_TARGET_CHIP = AX650 -- AXERA_TARGET_CHIP = AX650 -- BSP_MSP_DIR = /home/issa/ax-pipeline/build/../ax650n_bsp_sdk/msp/out -- CMD_GET_BSP_VERSION = strings /home/issa/ax-pipeline/build/../ax650n_bsp_sdk/msp/out/lib/libax_sys.so | grep 'Axera version' |...

Traceback (most recent call last): File "C:\Users\DELL\Desktop\PP-YOLOE-master\train.py", line 43, in trainer.train(num_epoch=args.num_epoch, File "C:\Users\DELL\Desktop\PP-YOLOE-master\ppyoloe\trainer.py", line 300, in train self.__train_epoch(max_epoch=num_epoch, epoch_id=epoch_id, log_interval=log_interval, local_rank=local_rank, writer=writer) File "C:\Users\DELL\Desktop\PP-YOLOE-master\ppyoloe\trainer.py", line 204, in __train_epoch output =...

按照自定义数据集进行了修改，并按照子图的切法进行了操作，为什么训练精度在85%左右，但是在板子上跑的时候就非常的低，低到10%左右，甚至根本检测不出来？有人遇到这种问题吗？是什么原因导致的呢？

pavia真值

3

您好 请问一下 我下载的pavia真值和HSI尺寸对不上 请问是怎么裁剪的呢 我看真值尺寸是320 从HSI左下角截取了320 但是两幅图像依然不匹配 请问这块是怎样处理的呢 谢谢~

This is a very impressive work. May I ask if there is any code for generating the depth map?