Thank you for making these resources accessible for free and easy to understand, I am actually reading your Book (Machine Learning with PyTorch and Scikit-learn), all DL concepts are well explained with examples of codes thanks for that.
Thank you, this is a really good introduction to PyTorch which doesn't make you feel frightened of how many difficulties are inside =) If you look someday for a new topic to present, please consider making more elaborate comparison TensorFlow vs. PyTorch.
Glad it was helpful! Haha, yeah, PyTorch vs TensorFlow is a very popular question/discussion topic. A side-by-side comparison with actual code might be helpful. I am just worried that by now by Tf are way too bad :P
I expect it may turn out to be more intriguing! =) TF supports training with TPU, which several times faster than all GPUs. Also, TF seems to be faster in training and inference on CPU (while PyTorch most of the time better on GPU). Google with it's Kaggle, Colab, Cloud and HW is more powerful after all! And the most exciting combat - which one is better for production solutions, and here I lost the track, maybe you could give some insights.
@@dmitrymalishev6045 Actually, if you use PyTorch Lightning or LightningLite on top of PyTorch, you get support for all of it :). GPU, TPU, IPU, HPU. And it's super easy!
Thank you for the link! Looks like Lightning uses XLA to support TPU. Half a year ago PyTorch+XLA wasn't stable enough, some kernels just didn't work. Hope, it's better now, I'll give it a try on some upcoming Kaggle competition! =)
Thank you for making these resources accessible for free and easy to understand, I am actually reading your Book (Machine Learning with PyTorch and Scikit-learn), all DL concepts are well explained with examples of codes thanks for that.
Glad to hear that both the videos and the book are useful to you!
Such a great video! Definitely deserved more views! I like your way of explain things!
Thanks for the kind words, and I am glad to hear you liked it!
Your lectures deserves a lot more views than it has. You are a brilliant professor!
This is such a useful video with just 2k views!!
Glad you liked it!
This is super cool really !! Though I am aware about most of the concepts it was a treat watching it.
Thank you, this is a really good introduction to PyTorch which doesn't make you feel frightened of how many difficulties are inside =)
If you look someday for a new topic to present, please consider making more elaborate comparison TensorFlow vs. PyTorch.
Glad it was helpful! Haha, yeah, PyTorch vs TensorFlow is a very popular question/discussion topic. A side-by-side comparison with actual code might be helpful. I am just worried that by now by Tf are way too bad :P
I expect it may turn out to be more intriguing! =) TF supports training with TPU, which several times faster than all GPUs. Also, TF seems to be faster in training and inference on CPU (while PyTorch most of the time better on GPU). Google with it's Kaggle, Colab, Cloud and HW is more powerful after all! And the most exciting combat - which one is better for production solutions, and here I lost the track, maybe you could give some insights.
@@dmitrymalishev6045 Actually, if you use PyTorch Lightning or LightningLite on top of PyTorch, you get support for all of it :). GPU, TPU, IPU, HPU. And it's super easy!
Here's a link if you are interested: pytorch-lightning.readthedocs.io/en/stable/starter/lightning_lite.html
Thank you for the link! Looks like Lightning uses XLA to support TPU. Half a year ago PyTorch+XLA wasn't stable enough, some kernels just didn't work. Hope, it's better now, I'll give it a try on some upcoming Kaggle competition! =)
Beauty