g413n

5 ms·

1.
by g413n 7mo ago
it's a pretty general policy but this is all super early, it's great at exploring websites so fuzzing was easy, for CAD it has good enough base rates with the few-shot prompt when we do the repetitive stuff, and we gave it checkpo
2.
by g413n 7mo ago
relevant note is that we finetuned by having the human also use arrow keys which keeps it in-distribution but also slower to collect
3.
by g413n 7mo ago
yeah we actually had some wacky ideas with ctc + a reverse-causal mask but diffusion does just make it all a bit more simple
4.
by g413n 7mo ago
we do exponential binning but fwiw I think we can do way better just hasn't been the main research area initially
5.
by g413n 7mo ago
yeah we've done audio work in the past so we'll def merge the recipes at some point, long term should have full io that a human has (except maybe not generating video for video calls that seems a bit much)
6.
by g413n 9mo ago
so cool :)
7.
by g413n 9mo ago
what's the basis for conversion between hours of neural data to number of tokens? is that counting the paired text tokens?
8.
by g413n 1y ago
yep this. we just turned off management
9.
by g413n 1y ago
we would be so down to buy s3 junkyard tbh we were going around begging various storage clouds to offer us this before giving up and building it ourselves
10.
by g413n 1y ago
yep this was the case for us.
11.
by g413n 1y ago
I think <2x more drives than needed, not 20x (24 vs 14TB), but the racks holding the drives could've been denser. Around the same cost in any case and our colo doesn't charge for space, so it's not a big deal and we were j
12.
by g413n 1y ago
yeah we weren't sure about putting that number esp whether it includes all the image attachments, but in any case it's at least around the right reference class for the largest text data operations.
13.
by g413n 1y ago
def agree on the setup fees, that was just a price crunch to get it done within the weekend. (too short-notice for professional services, too sensitive for craigslist, so basically just paying a bunch of folks we already knew and trusted) f
14.
by g413n 1y ago
yeah that's why we started paying people near the second half- not super clearly stated in the blogpost, but the novelty definitely wore off with plenty of drives left to stack, so we switched strategies to get it done in time. I think
15.
by g413n 1y ago
yeah the cost calculus is very different for gpus, it absolutely makes sense for us to be using cloud there. also hardly any datacenters can support the power density, esp in downtown sf
16.
by g413n 1y ago
it's not just in sf it's across the street from our office this has been incredibly nice for our first hardware project, if we ever expand substantially then we'd def care more about the colo costs.
17.
by g413n 1y ago
yeah it's on the wishlist to try
18.
by g413n 1y ago
our training stack doesn't make strong assumptions about data integrity, it's chill
19.
by g413n 1y ago
When I was working at sfcompute prior to this we saw multiple datacenters literally catch on fire bc the industry was not experienced with the power density of h100s. Our training chips just aren't a standard package in the way JBODs a
20.
by g413n 1y ago
just general research work. Once the recipes are efficient enough the modality is a smaller detail. On the product side we're trying to orient more towards 'productive work assistant' rather than the default pull of audio mod
21.
by g413n 1y ago
yeah this it means that even after negotiating much better terms than baseline we run into the fact that cloud providers just have a higher cost basis for the more premium/general product.
22.
by g413n 1y ago
yeah misunderstanding we'll update the post-- separately it's true that we aren't network specialists and the network wrangling was prob disproportionately hard for us/ shouldn't have taken so long.
23.
by g413n 1y ago
around 250
24.
by g413n 1y ago
egress costs are the crux for AWS and they didn't budge when we tried to negotiate that we them, it's just entirely unusable for AI training otherwise. I think the cloudflare private quote is pretty representative of the cheaper e
25.
by g413n 1y ago
yeah it's totally plausible that we go with something like this in the future. We have similar offers where we could separate out either the financing, the build-out, or both and just do the software. (for Hetzner in particular it was
26.
by g413n 1y ago
worth stressing that we literally could not afford pretraining without this, approx our entire seed round would go into cloud storage costs
27.
by g413n 1y ago
atm we don't and we're a bit unsure whether it's a free lunch wrt adding complexity. there's a really nice property of having isolated hard drives where you can take any individual one and `sudo mount` it and you have a
28.
by g413n 1y ago
oh we use cloud gpus, infiniband h100s absolutely aren't something we want to self-host. not aws tho, they're crazy overpriced; mithril and sfcompute! we also use cloudflare extensively for everything that isn't the core heap
29.
by g413n 1y ago
we don't have perfect metrics here but this seems to match our experience; a lot of failures happened shortly after install before the bulk of the data download onto the heap, so actual data loss is lower than hardware failure rates
30.
by g413n 1y ago
we're in a pretty unique situation in that very early on we fundamentally can't afford the hyperscaler clouds to cover operations, so we're forced to develop some expertise. turned out to be reasonably chill and we'll

More ›