মিডিয়া
- টুইট
- টুইট এবং উত্তর
- মিডিয়া, বর্তমান পৃষ্ঠা।
-
Excellent talk indeed! After I watched it,
@seanmylaw pointed out there's also a longer version (/director's cut?!) from PyData Ann Arbor on YouTube:https://www.youtube.com/watch?v=YPJQydzTLwQ&t=2636s … -
Comparing some more biologically inspired learning algos to backprop by Bartunov et al: https://arxiv.org/abs/1807.04587 .TP and FA look ok-ish on MNIST and CIFAR for a first try but we probably also need different architectures to go with these (or any non-BP algo) to beat BP performancepic.twitter.com/CfW7e9BDzb
এই থ্রেডটি দেখান -
Heard great things about UMAP (Uniform Manifold Approx. & Projection for Dim. Reduction) vs e.g., T-SNE, but haven't had a chance to read up on it, yet. Conveniently, I just see that there's a recording from a UMAP talk at SciPy 2018 on YouTube: https://www.youtube.com/watch?v=nq6iPZVUxZU …. Perfect.
-
Also has a good set of references on when/how the term was coined:pic.twitter.com/fqssPaOb28
এই থ্রেডটি দেখান -
Yay, finally LaTeX equation support for Apple Keynote
pic.twitter.com/Oz4gP7IlaC
-
Apologies if your tool does not do what Safari shows me what it could do. The naming was basically part of to identify which extension I was talking about as it's displayed in the title.pic.twitter.com/khLIwU5PPl
-
If you have nbextensions installed, there's an extension for that included:pic.twitter.com/8cZs1nAU64
-
yeah good point. I use an open port most of the time anyway (I add a password lock to my notebooks though) since I have a dynamic IP when I use my laptop. But I actually do like the comment feature of the collab-style notebookspic.twitter.com/i0OvHPJb1X
-
"Comparing Google’s TPUv2 against Nvidia’s V100 on ResNet-50" Interesting benchmark. Not much of a difference actually for the larger batch sizes (the missing y-axis label is the number of images) https://blog.riseml.com/comparing-google-tpuv2-against-nvidia-v100-on-resnet-50-c2bbb6a51e5e …pic.twitter.com/2ylA3LlRHt
এই থ্রেডটি দেখান -
A handy model summary function for PyTorch: https://github.com/sksq96/pytorch-summary …pic.twitter.com/lcsramkoTB
-
Just put out a new version of mlxtend today for the Journal of Open Source Software (was accepted today, yay). Not too many changes this time, but the colorblind-friendly decision region plots are now finally in the release version :) http://rasbt.github.io/mlxtend/changelog/ …pic.twitter.com/VPx2PFmaTk
-
besides the email, no idea. I actually didn't even bother reading that and thought that deleting it is way easier (and "better safe then sorry") option -- never logged in into google analytics anyway. Below is a screenshot of the contents if you haven't got that mailpic.twitter.com/w2Rdxj8OUh
-
Just got home and found an unexpected package from my colleague
@karlrohe my doorstep ... Thanks a lot!!! Having never seen one in real life, I am amazed how small these things are. I guess it's playtime now ;)pic.twitter.com/RXDIHgbvf1
-
Caffe64 -- a no-dependency replacement of Caffe deep learning framework written in assembly. Looks like a cool project :) https://github.com/dfouhey/caffe64 pic.twitter.com/RIBU1WyDaz
-
"Fundamentals of Data Visualization" a great free online-ebook by
@ClausWilke that's worth checking out! http://serialmentor.com/dataviz/ pic.twitter.com/X9O4QM1rny
-
Thanks
@rheartpython I thought you may also be interested in this little write-up on entropy as a criterion (analogous to gini): https://sebastianraschka.com/faq/docs/decisiontree-error-vs-entropy.html …pic.twitter.com/qlFCNkRWNr
-
skope-rules -- a neat scikit-learn contrib project that lets you extract logical, interpretable rules for "scoping" a target class https://github.com/scikit-learn-contrib/skope-rules … (reminds me a bit of Learning Classifier Systems
@DocUrbs ) https://github.com/scikit-learn-contrib/skope-rules …pic.twitter.com/gl3qaAPYyn
-
Thx! yeah, I was thinking about "sequential" color maps, but the problem really is to go >3 classes. Actually, turned out the matplotlib v2.0 defaults were quite good regarding the different color blindness tests via Coblis. What I ended up with is the following:pic.twitter.com/qkyWcA7mZ9
-
"Deep Learning Framework Examples" -- A nice collection of Jupyter Notebooks implementing benchmarks for different deep learning libraries https://github.com/ilkarman/DeepLearningFrameworks …pic.twitter.com/NRQLXAHnCc
এই থ্রেডটি দেখান -
লোড হতে বেশ কিছুক্ষণ সময় নিচ্ছে।
টুইটার তার ক্ষমতার বাইরে চলে গেছে বা কোনো সাময়িক সমস্যার সম্মুখীন হয়েছে আবার চেষ্টা করুন বা আরও তথ্যের জন্য টুইটারের স্থিতি দেখুন।