Jungmin_ML_DL_CV_LLM

Results 19 comments of Jungmin_ML_DL_CV_LLM

Can StableDiffusionBrushNetPipeline be connected to the openpose ControlNet like the StableDiffusionControlNetPipeline?

can you share the code?

Integration of dinov3, including optional train and inference scripts.

changing backbone with DinoV3 is super better for object detection, might be latency issue...

Using track() to handle multi-rtsp stream input with TensorRT + Triton Inference Server

Thanks for lots of intersted shares. I would like to use detection, not tracking model. Just change model.track to model.detect? Also I want to use several models concurrently. utilizing multiple-RSTP...

Using track() to handle multi-rtsp stream input with TensorRT + Triton Inference Server

> Yes, you can change `model.track` to `model.detect` for detection tasks. To use several models concurrently with multiple RTSP streams, consider using threading to handle each stream with its own...

Using track() to handle multi-rtsp stream input with TensorRT + Triton Inference Server

Can you give me a help code to use multiple RTSP stream and Multi-models? it will be better to understand what you mean. Thanks, Jungmin

Using track() to handle multi-rtsp stream input with TensorRT + Triton Inference Server

![스크린샷 2024-09-07 231629](https://github.com/user-attachments/assets/f72876a1-93af-4552-9a2b-f954018e9782) ![스크린샷 2024-09-07 231605](https://github.com/user-attachments/assets/ee96484f-af02-49cd-8d08-50e2d8a8b174) It is only working on one model and one stream. Can you tell me how to process multi-stream and multi-model?

Using track() to handle multi-rtsp stream input with TensorRT + Triton Inference Server

> To process multiple streams with multiple models, use threading to create separate threads for each stream and model combination. This allows concurrent processing of each stream with its respective...

Using track() to handle multi-rtsp stream input with TensorRT + Triton Inference Server

Above my code, I already followed your simple code, but results only show one stream and one model. Can you fix my code to use multiple models and multiple stream?

About stylemix

stylemix is within generate_video.py. you can just follow step-by-step. But some of mix is not working well.