Intel/AMD CPU - Submissions for EXPORT competition #3432
Replies: 7 comments 10 replies
-
|
Are we allowed to submit an export solution using frameworks like Intel's openvino? |
Beta Was this translation helpful? Give feedback.
-
|
if the solution does not require the presence of a hardware accelerator such as an Intel stick, then it is possible |
Beta Was this translation helpful? Give feedback.
-
|
了 |
Beta Was this translation helpful? Give feedback.
-
|
https://github.com/SamSamhuns/yolov5_export_cpu Putting this repo for exporting the YOLOv5 model to OpenVINO's IR model format for Intel CPU/Edge device inference |
Beta Was this translation helpful? Give feedback.
-
|
Hi Sam, thank you for the submission. We will take a look at it |
Beta Was this translation helpful? Give feedback.
-
|
https://github.com/karanjakhar/yolov5-export-to-cpu Please have a look. |
Beta Was this translation helpful? Give feedback.
-
|
will do, thanks! |
Beta Was this translation helpful? Give feedback.
-
|
https://github.com/AnhPC03/yolov5-export-to-cpu-mnn |
Beta Was this translation helpful? Give feedback.
-
|
@glenn-jocher Thank you. I will add profiling results as soon as posible. |
Beta Was this translation helpful? Give feedback.
-
|
@glenn-jocher Sorry, could I have one question? Did you mean profiling results? What's information do you want to know? Thanks. |
Beta Was this translation helpful? Give feedback.
-
|
@AnhPC03 ah profiling just means testing the inference speed |
Beta Was this translation helpful? Give feedback.
-
|
@glenn-jocher I've compared inference speed between MNN and Pytorch in my Google Colab. I ran loop 50 times inference in one image and got following result: MNN took 52 seconds and Pytorch took 173 seconds. Please took a look the link above, and tell me if you want other information. Hope this is your need. |
Beta Was this translation helpful? Give feedback.
-
|
This looks like a quite speedup! I am looking forward for evaluation |
Beta Was this translation helpful? Give feedback.
-
|
@shivamswarnkar and I have worked on this submission for AMD/Intel CPU. https://github.com/BlueMirrors/Yolov5-ONNX We are getting identical outputs as PyTorch Inference(FP16). This uses letterboxing (so no need to forcefully resize) and you can directly test it on Colab as well. After benchmarks with 4 backends - torch, Tensorflow, tflite, and ONNX; ONNX performed the best for us. P.S. We are working on batching support. |
Beta Was this translation helpful? Give feedback.
-
|
Hi participants, Thank you all for your submissions, we have already reviewed them and are about to announce the winners! Please send me an email at stefani@ultralytics.com including your full name and GitHub username, so I can share more information with you. Stay tuned - the results are coming soon! Stefani Kovachevska |
Beta Was this translation helpful? Give feedback.
-
|
Hi everyone, We were thrilled to have received your submissions to our YOLOv5 Export Competition for the Intel/AMD CPU category. Unfortunately, the submissions in this category did not match the minimum set of requirements in each of the evaluation criteria. Therefore, no winner was selected for the category this time. However, there will be more chances for participants to compete again in the future. Thank you all for your efforts and for being a valuable member of our open-source community. Contributions from you are what makes our community great. 🙌 Be sure to check out the winning repositories in the rest of the categories in our latest blog post: https://ultralytics.com/article/export-competition-winners Stay amazing and keep creating! 🚀 |
Beta Was this translation helpful? Give feedback.
-
Use this discussion thread for YOLOv5 🚀 EXPORT Competition submissions in the Intel/AMD CPU category. Good luck!
Beta Was this translation helpful? Give feedback.
All reactions