@Pimp_Fada@fchollet I think it's just a matter of time.. Also, remember, #tensorflow is the only framework that supports this chip.. and it's on its first version... Imagine how this evolves over time to better utilize the chip capabilities.
New code walkthrough on https://t.co/m6mT8SrKDD: data-efficient GANs with Adaptive Discriminator Augmentation. I personally really enjoyed reading through it!
https://t.co/ZGuQLO06Cw
@mazeyar Nope.. Fan kicked in later, but didn't really bother me. And can multitask on many other things without any issue. Memory pressure was around 25% maybe.. ~ 60% of CPU was idle.. GPU was pretty busy along with about 4 CPU cores out of 10.
@jessie9879 @fchollet This is a 32-core GPU Max with 64GB memory. I haven't run it on Colab myself, but this is what's reported in the article by the author.
5. @Apple#MacBookPro2021#M1Max seems versatile and can easily handle those workloads = more time saved or more work done fast!
Hope this helps someone with their purchase decision! ;)
4. Here is a link to a performance benchmark I created for ML model training, comparing #tensorflow metal vs. #pytorch CPU on the new #M1Max chip: https://t.co/TW4yAp8ovw
With the new Apple chips, I can get 64GB GPU memory and super powerful & efficient processor for DL, all in a 14" package. hmm.. I think it's time to convert back to TensorFlow.
Proud of what @WCCTrails has become.. Thanks to all the sponsors & volunteers who have helped shape it over the years. It has been my go-to for deep thinking and reflection solo sessions for years. @RegionWaterloo, please keep it going for years to come!