Colab is a cheap (free) way to start, though you won't be training very large models for very long (which you shouldn't be doing if you are a beginner). You learn a different set of skills when you put together your own rig and install/maintain the libraries, which is something I recommend everyone to try just go gain an appreciation of devops skills. But that's not a necessary diversion for a beginner (and may needlessly increase the learning curve).
Ask HN: What ML platform are you using?
41–50 of 84 posts
Re: Ask HN: What ML platform are you using?
#42I worked in Google Research for over 5 years doing Machine Learning, and recently quit to build my own ML start-up. These days, I solve a mix of NLP, computer vision, and tabular problems, all with state of the art neural network techniques. I've tried many setups.
My advice is go with Colab Pro ($50/mo) and TensorFlow/Keras. You can go with Pytorch too if you prefer.
I made the mistake of buying a 2080Ti for my desktop thinking it would be better, but no. Consumer grade hardware is nowhere near as good/fast as the server grade hardware you get in Colab. Plus you have the option to use TPUs in Colab if you want to scale up quickly.
You really don't need to get fancy with this setup. The best part of using Colab is you can work on your laptop from anywhere, and never worry about your ML model hogging all your RAM (and swap) or compute and slowing your local machine down. Trust me, this sucks when it happens, and you have to restart!
As for your data, you can host it in a GCS bucket. For small data (<1TB) even better is Google drive (I know, crazy). Colab can mount your Google drive and loads from it extremely quickly. It's like having a remote filesystem, except with a handy UI and collaboration options, and an easy way to inspect and edit your data.
Re: Ask HN: What ML platform are you using?
#43Earlier quoted context omitted.
I would say the exact same thing if the circumstances weren’t so dire, but an NVIDIA GPU is so expensive nowadays that it might be a bit better to use rented services (like the paid version of Google Colab) for now, if you have any monetary constraints. Maybe GPU prices will stabilize after Ethereum switches to POS and manufacturing pipelines get back to normal, but then I’m not that sure after seeing US trying to go…
I have a 3090 for serious (hobby) work and a 1070 from 4ish years back next to my bed. I think getting anything 1070 or better is good enough at the beginner level (training baby datasets/models from scratch such as MNIST/CIFAR, transfer learning the big models). I just don't understand the cost argument .. you can get this stuff used. Main thing is you need CUDA. The 3090 machine gets about the same use as the 1070…
The problem is that even a 1070 is ridiculously expensive these days (About $400 on Newegg, that was the cost of a RTX 2070 a few years ago!) If you can get an used GPU from a friend that would be great, but other than that you’re going to have to shell hundreds of dollars for an old GPU that you will probably have to upgrade soon.
I don’t know the OP’s financial situation, but if you’re a poor student than these things certainly matter.
Re: Ask HN: What ML platform are you using?
#44I worked in Google Research for over 5 years doing Machine Learning, and recently quit to build my own ML start-up. These days, I solve a mix of NLP, computer vision, and tabular problems, all with state of the art neural network techniques. I've tried many setups. My advice is go with Colab Pro ($50/mo) and TensorFlow/Keras. You can go with Pytorch too if you prefer. I made the mistake of buying a 2080Ti for my desk…
+1 and you can even connect your Colab to a GCP Marketplace Colab runtime that has no time limit (but will cost us) if you e.g. need to run something for a few days (although then you don't get the awesome Google drive mounting - hope they fix this eventually)
Re: Ask HN: What ML platform are you using?
#45Earlier quoted context omitted.
I have a 3090 for serious (hobby) work and a 1070 from 4ish years back next to my bed. I think getting anything 1070 or better is good enough at the beginner level (training baby datasets/models from scratch such as MNIST/CIFAR, transfer learning the big models). I just don't understand the cost argument .. you can get this stuff used. Main thing is you need CUDA. The 3090 machine gets about the same use as the 1070…
The problem is that even a 1070 is ridiculously expensive these days (About $400 on Newegg, that was the cost of a RTX 2070 a few years ago!) If you can get an used GPU from a friend that would be great, but other than that you’re going to have to shell hundreds of dollars for an old GPU that you will probably have to upgrade soon. I don’t know the OP’s financial situation, but if you’re a poor student than these thi…
Good points .. however, my advice for the cash strapped students on HN: Buying 1 good machine will last you 4+ years easily these days. Buying a system like Alienware with a 3070-3080 was under 2K Canadian over the recent Black Friday shopping event. Over 4 years, that is $500. If you are studying CS, it is absolutely worth buying a decent machine. You don't need a top of the line machine but you need to be able to study your craft on something more powerful than a rasperry pi or junky old machine.
I was quite poor growing up, and I recall buying a 3K machine when I started undergrad (that was crap hardware by today's standard). And i have no doubt that having this machine helped me get my first job, and things got better from then. If you are in CS, think of it as an investment, and make it pay off!
Btw, I am trying to be positive .. pls don't construe anything here as negative. I appreciate that money is tight for a lot of folks. Paying 10% interest to buy a 3K machine is not a good idea!! I just skimped a lot as a student, and some of it was quite pointless. I wish someone explained this to me, and hence my comment.
If you are in a situation where you'd have to take a loan to buy a system, pls don't feel like you need a GPU to do anything useful. I am certain one can make do with just colab and a web browser. Good luck to all the students out there!! Life is hard at that stage .. it gets far easier once you have a paying job in the field.
Re: Ask HN: What ML platform are you using?
#46> I am very much a beginner in the space of machine learning While the (precious and useful) advice around seem to cover mostly the bigger infrastructures, please note that you can effectively do an important slice of machine learning work (study, personal research) with just a battery-efficiency-level CPU (not GPU), in the order of minutes, on a battery. That comes before going to "Big Data". And there are lightweig…
Note that this won't work with reasonably performant CNNs. Passing an image batch through a large-ish ResNet takes half a second on our GPUs, several minutes at full load on CPU. This makes training infeasible, and most models small enough to work on CPU are so far from state-of-the-art that you can't do any worthwhile computer vision research with them.
Why start with vision? Do some language models. I used to train those all the time on my laptop.
GPT 5MB for the win. It really works.
Re: Ask HN: What ML platform are you using?
#47I worked in Google Research for over 5 years doing Machine Learning, and recently quit to build my own ML start-up. These days, I solve a mix of NLP, computer vision, and tabular problems, all with state of the art neural network techniques. I've tried many setups. My advice is go with Colab Pro ($50/mo) and TensorFlow/Keras. You can go with Pytorch too if you prefer. I made the mistake of buying a 2080Ti for my desk…
(OP, please don’t subject yourself to TensorFlow/Keras. The moment Jax became available on TPUs publicly, the moment I stopped using TF. And boy oh boy, “never looked back” is an understatement. I still cringe remembering all the time I spent trying to get tf.function to just please, please work, like a housewife alarmed that neither her partner nor herself are able to actually work.)
Re: Ask HN: What ML platform are you using?
#48Working with a large well known tech company, with surprisingly basic/non-existent ML until only very recently. Using Redshift to do a lot of the heavy lifting and initial data preparation, then SageMaker for hosting models and scoring, and Tableau for dashboards. While you can do training within SageMaker, we have a cluster of EC2 instances using H2O libraries (xgboost) to train, then wrap the resulting model as a d…
I found Redshift to be far inferior to Snowflake as a data warehouse for marshalling any tables or views you need for ML work. There's lots of statistical functions available within Snowflake that will speed things up for you if you need pre-calculations on feature sets.
Re: Ask HN: What ML platform are you using?
#49I see some great recommendations in the thread already, but I think https://cocalc.com is definitely worth checking out if you consider yourself to still be more of a learner. Their focus seems to really be on helping people who are new to the field get started. It offers familiar Jupyter Notebook-like features so you should feel right at home.
I have no affiliation with them whatsoever :) Just a fan of what they’re doing.
Re: Ask HN: What ML platform are you using?
#50I put together a linux box with a 2080ti a few years ago and have been using it consistently for personal ml research ever since. Ive found it well worth the investment and learned that the ease with which I can jump into hacking on a project is key, which is why this works so well for me. I can just ssh in at any time and start experimenting with models. Even if its not technically economical when you do the math, t…
What do you wrt CUDA and linux? I'm a linux person but every time I try and mess around with CUDA the whole thing gets super annoying. I don't want to have to reinstall everything every time there is a kernel upgrade . Maybe there is some trick with WSL2 now?