How to avoid machine learning pitfalls:
a guide for academic researchers
https://arxiv.org/pdf/2108.02497.pdf
🐦🔗: https://twitter.com/RexDouglass/status/1620123641626382338
Ayyuce Demirbas
- 13 Followers
- 17 Following
- 276 Posts
The dataset is imbalanced. Therefore, we can use the generate_class_weights function to produce class weights to apply in the fit function.
New project alert: Brain Tumor Classification and Grad-CAM visualization 🧠🔍 Get a glimpse into the results and access the code at https://github.com/ayyucedemirbas/brain_tumor_classification/blob/main/Brain_Tumor_Classification_Grad_cam.ipynb
Another thing PyTorch 2.0 helps speed up - overhead.
Overhead is everything other than the GPU doing work. It can come from Python, the ML framework, CUDA kernel launches, etc. - regardless, it's why your nvidia-smi util is so low!
So... how do we diagnose and resolve it?
(1/8)
Horace He on Twitter
“Another thing PyTorch 2.0 helps speed up - overhead. Overhead is everything other than the GPU doing work. It can come from Python, the ML framework, CUDA kernel launches, etc. - regardless, it's why your nvidia-smi util is so low! So... how do we diagnose and resolve it? (1/8)”
Amerikalı arkadaşımı kafamızda fesle doğduğumuza inandırmaya çalışıyorum. İnandı gibi sanki.
It's #PaperReadingWeekend 🙃
It's #LinuxKernelCompilingWeekend I'm compiling the Linux kernel 6.1 and reading the Ensembles of Vision Transformers paper at the same time. https://twitter.com/demirbasayyuce/status/1596545074032410624
🐦🔗: https://twitter.com/demirbasayyuce/status/1604114624114106369
Le despotisme démocratique by Alexis de Tocqueville
It's Friday! What are you reading this weekend?
🐦🔗: https://twitter.com/goodreads/status/1619032493675388928
Looking for a free MLOps course?
Check out https://madewithml.com/
🐦🔗: https://twitter.com/AssemblyAI/status/1618975867429953536
