dao027

Caching images failed in v6.1 DDP training using --cache 'disk'

1

### Search before asking - [X] I have searched the YOLOv5 [issues](https://github.com/ultralytics/yolov5/issues) and [discussions](https://github.com/ultralytics/yolov5/discussions) and found no similar questions. ### Question it's strange my **yolov5s_v6.1** training goes wrong during Caching...

yolov5s 6.1 pytorch模型和TRT输出结果有差异 [TRT and pytorch have different inference result]

3

## Env - GPU, e.g. RTX3090 - OS, Centos 8 - Cuda version 11.2 - TensorRT version 8.4.1.5 ## About this repo - which branch/tag/commit are you using? - which...

### Before Asking - [X] I have read the [README](https://github.com/tinyvision/DAMO-YOLO/blob/master/README.md) carefully. 我已经仔细阅读了README上的操作指引。 - [X] I want to train my custom dataset, and I have read the [tutorials for finetune on...

请教mpp解码rtmp流decode_get_frame返回正常，但是frame为空的情况

4

如题，mpp解码RTSP流的时候都正常，换成一个rtmp流的时候就一直是死循环，decode_get_frame(ctx, &frame)得到的frame一直是空

请问rknn支持3d卷积吗？

2

目前模型需要conv3d的支持，环境是1808，rknn v1.7.1，请问有算子可以支持conv3d吗？

RK1808解码后的mppframe使用rga转换NV21到RGB888过程中，目标图片的fd应该怎么定义？

2

因为缺乏经验，请教一下各位大佬 目前有解码后的mppframe，根据rga文档中的效率建议，希望使用fd来进行操作，根据官方demo的程序稍作修改如下： ``` int ret = 0; int src_width, src_height, src_format; int dst_width, dst_height, dst_format; int src_dma_fd = mpp_buffer_get_fd(buffer); int dst_dma_fd; char *src_buf, *dst_buf; int src_buf_size, dst_buf_size; rga_buffer_t src_img, dst_img;...

a request in Video Object Tracking

2

is there any way to add a choice inside one video tracking? like, in one video tracking, 0-3s one object is sitting, 3-5s he is standing, with NO change to...

how to set Toggle Interpolation default NOT continuous in Video Object Tracking task?

1

i just want to add a bbox in one keyframe of the video in Video Object Tracking task.I don't want it be continuous the default setting is continuous, i need...