huihui6666

Hi,I found it detection only achieve about 1.5FPS on RTX4090,it's much lower than the official.do you find the reason for the low speed?

> The results of tensorRT10.3 are different from those of TensorRT 10.8. The results of 10.8 are as expected, so the results of 10.3 are wrong. I want to use...

I use the config file `yolo_world_v2_l_efficient_neck_2e-4_80e_8gpus_mask-refine_finetune_coco.py` `_base_ = ( '../../third_party/mmyolo/configs/yolov8/' 'yolov8_l_mask-refine_syncbn_fast_8xb16-500e_coco.py') custom_imports = dict( imports=['yolo_world'], allow_failed_imports=False) # hyper-parameters num_classes = 57 num_training_classes = 57 max_epochs = 80 # Maximum training...

some of my train log as follow 2024/11/26 11:33:52 - mmengine - INFO - Exp name: yolo_world_v2_m_vlpan_bn_2e-4_80e_8gpus_mask-refine_finetune_coco_my_20241126_091045 2024/11/26 11:34:01 - mmengine - INFO - Epoch(train) [35][7050/7676] base_lr: 2.0000e-04 lr: 1.1833e-04...

I encounter the similar problem, do you solve it?

> > 我遇到了类似的问题，你解决了吗？ > > After I changed the system to ubuntu20.04 and JetPack5 version, everything was normal. I don’t know why it can’t run normally on ubuntu22.04 and JetPack6....

> > 您好，对于需要金字塔特征图的头部，我们通常遵循 MTP 的实现（[https://arxiv.org/abs/2403.13430）。MTP](https://arxiv.org/abs/2403.13430)在骨干网络中添加了一个小型 ViTdet 风格的 FPN。https [://github.com/ViTAE-Transformer/MTP/blob/962f7fd8781c095eb26db65ead3016e666b6d417/RS_Tasks_Finetune/Semantic_Segmentation/mmseg/models/backbones/vit_rvsa_mtp.py#L631](https://github.com/ViTAE-Transformer/MTP/blob/962f7fd8781c095eb26db65ead3016e666b6d417/RS_Tasks_Finetune/Semantic_Segmentation/mmseg/models/backbones/vit_rvsa_mtp.py#L631) > > 这是在本期 Feature2Pyramid 类中实现的：[#54](https://github.com/facebookresearch/dinov3/issues/54) > > 我已经使用上述 MTP github 中的配置文件测试了带有 mmseg 和 mmdet 的 dinov3，并在 DINOv3 主干中接线，使用 get_intermediate_layers 获取 7B...