Hope all is well!
Nice Work!
Cheers mrfabulous1 
1 Like
really, cldnt find it
Built a web application “insightsR” based on old machine learning course from fast.ai. Dealt with providing automated insights for tabular data, taking just 2 inputs from the user (dataset and the target). Completely based on Jeremy’s lessons - added automation, putting things together and using streamlit for a better appealing web application…
Was first hesitant to display as my addition wasn’t much…
Details below:-
Github: GitHub - Vinothsuku/insightsR: automated insights for tabular data
Blog: insightsR — automated insights for tabular data | by Vinoth Sukumaran | Analytics Vidhya | Apr, 2021 | Medium
Online: https://insightsr.herokuapp.com/
Please let me know your comments.
2 Likes
Hi everyone,
I released a new, fastai2 compatible, version of my library for neural network compression, shamelessly called Fasterai.
Focus has been done on sparse neural network training, i.e. replacing most of the weights in the network by zeroes, but other techniques such as Knowledge Distillation are also available.
It has all been made possible thanks to the magic of the fastai callbacks system.
Details:
Give it a try, you would be surprised by how many weights you can remove from your networks while keeping performance intact !
11 Likes
Congrats! Also I love the theme colors! 
1 Like
For the first project, I took the fastai tutorial and created an image classifier that attempts to classify native trees in Kentucky. USA. It works to a point depending on how you take the picture. The inspiration came from my recent move out of the suburbs and into a more wooded are with lots of wildlife. I really didn’t know how to determine what tree was which type, so I developed this app.
Hey! I started implementing what I have learned, here is my first medium article explaining how I tackle the problem of classifying Malware.
Have a nice day!
I like that, I like how you added interaction to this site. How did you do that? did you upload a video?
I’ve documented my work with medium for those who may be interested.
1 Like
Hi all, sharing a mini-project on utilizing an external dataset in performing a multi-label classification. Enjoy!
Here’s a simple step-by-step process in determining what the metric accuracy_multi measures.
1 Like
FastServe - Generate API endpoints from fast.ai models
Hi all,
We’ve just launched the private beta for FastServe: a service to turn pre-trained fast.ai models into APIs, that you can then plug into your applications. You upload a model file (e.g., export.pkl), and get back an API endpoint that can serve inference. We’re hoping this will help data scientists deploy models more quickly and easily.
If you are interested in trying it out, please check out the introduction video on YouTube (link below), and sign up for the beta at https://launchable.ai/fastserve-beta. We’d love to hear what folks think!
Thanks!
Is it designed around fastai v1? Or fastai v2?
fastai v2
2 Likes
Using Images and an Algorithm to Triage Ill Babies
1 Like
Maybe you have seen thispersondoesnotexist.
Me and my partner are trying to do this for interiors/housing images.
So, we are using this repository: https://github.com/lucidrains/stylegan2-pytorch
We are running on amazon P2 instances recommended for this application.
As a rehearsal, we used the following:
- 50 images
- Instance: p2.xlarge =** 1 GPUs, 4vCPUs, 61 GiB RAM
Now we want to use our actual dataset and a faster machine:
- 15K images
- Instance: p2.16xlarge =** 16 GPUs, 64vCPUs, 732 GiB RAM
- 0.5% after 1 hours
We are concerned about the progress so far. The progress speed is similar to our rehearsal run even with this monster instance. We increased GPUs - but also the training data set. How can we estimate the time this will take?
What metrics on the AWS instance can we look at to make sure we are using its full capacity? Like GPU utilization?
1 Like
This is probably my first post here on this forum. I am learning through the 2019 course and created a simple donut vs bagel vs vada classifier. While the model itself is simple, rather silly, I used this opportunity to experiment with deploying this as a serverless deep learning inferencing function using AWS Lambda.
The web application can be accessed here: https://bit.ly/donutornot.
You can read about how I did it on my blog: Donut or Not! - Deploying Deep Learning Inference as Serverless Functions | Atma's blog or checkout the GitHub repo at: https://github.com/AtmaMani/donut_or_not
7 Likes
I was doing something pretty similar to this before. however, I switched from lucidrains (although he is awesome) to nvidias official repo.
As far as estimating how long something will take, you might be able to roughly calculate based on this GitHub - NVlabs/stylegan2-ada-pytorch: StyleGAN2-ADA - Official PyTorch implementation
but you never really know with gans, so I can’t comment to your resources/timing. i did find I had to checkpointed saved outputs for manual checking, since the % isn’t the same as , say a loss.
to check GPU usage, you can I used to use nvidia-smi (don’t remember the flags) to make sure.
