215 points by deeplearnerdude 6 months ago flag hide 12 comments
hackinghearts 6 months ago next
Great compendium! Looking forward to reading more about these techniques. Have you considered open sourcing the code used to generate this resource?
mindofmetis 6 months ago next
Last time I checked the plan was to release it but I haven't seen an update on that. I recommend checking their Github repo or asking them directly for more details.
fastforward 6 months ago prev next
What's the utility of knowing deep learning patterns? When should someone consider using them?
machinewhisper 6 months ago next
DL patterns are useful to improve efficiency of your models. You could consider applying them after you get the initial inference results with your baseline models.
codewave 6 months ago prev next
I'm new to deep learning. Can I still benefit from this? Or is it assuming I already have some experience?
intrinio 6 months ago next
Great question! It's well-written for both beginners and experienced users. But be ready to do independent reading and experimentation to get a solid understanding of these principles.
untame 6 months ago prev next
Thanks for sharing! I've been looking for resources exactly like this. Do you have any other reading suggestions related to deep learning techniques?
bitsource 6 months ago next
Glad it's helpful for you! I have a list of other resources I like to consult in my field, including 'Deep Learning' by Ian Goodfellow, 'The Hundred-Page Machine Learning Book' by Andriy Burkov, and 'Neural Networks and Deep Learning' by Michael Nielsen. Strongly recommend engaging with these sources!
polychromatic 6 months ago prev next
What do you think about TensorFlow vs PyTorch for applying these DL patterns?
devwiz 6 months ago next
There's no clear winner. TensorFlow has more robust features to optimize the network as a whole, and PyTorch makes it easy to build specific components of your model. Choose according to your preference and expertise.
transcendence 6 months ago prev next
I appreciate your hard work in creating this extensive resource. How do you think automating deep learning will evolve in the near future?
cpumagician 6 months ago next
Thanks! AutoML is expected to improve over time. I believe techniques in this compendium will be incorporated into these systems. But the true test will come when reliable AutoML systems are capable of learning from limited data and can be applied in real-world problems.