kevinlu1248

5 ms·

1.
by kevinlu1248 8mo ago
Ya definitely, that makes total sense. It feels to me that currently the labs have great researchers, who only care about making models perform better across raw intel and then they have incompetent applied AI engineers / FDE's wh
2.
by kevinlu1248 8mo ago
We have an explanation here: https://blog.sweep.dev/posts/next-edit-jetbrains#next-edit-a... But basically suggesting changes away from your cursor position
3.
by kevinlu1248 8mo ago
It's a bit undertrained on C#, we'll continue improving on this!
4.
by kevinlu1248 8mo ago
Also wish Fleet took off, not a fan of installing a new IDE for every separate repo that's in a different language
5.
by kevinlu1248 8mo ago
I've done some testing before and many of the new Jetbrains internal plugins cause memory leaks which really lags down my IDE...
6.
by kevinlu1248 8mo ago
Definitely agree here, have had so many cases where I would like ask Claude for XYZ, then ask for XYZ again but with a small change. Instead of abstracting out the common code it would just duplicate the code with the small change.
7.
by kevinlu1248 8mo ago
We currently integrate with Jetbrains' PSI
8.
by kevinlu1248 8mo ago
Honestly I think we can improve our training throughput drastically via a few more optimizations but we've been spending most of our time on model quality improvements instead.
9.
by kevinlu1248 8mo ago
You can see that Qwen3 does worse than Qwen2.5 on our benchmark. Reason is it's never been pretrained for FIM / autocomplete.
10.
by kevinlu1248 8mo ago
Not at the moment but we do host it for our Jetbrains plugin
11.
by kevinlu1248 8mo ago
Yes, I've used it to write blog posts / large user-facing copy.
12.
by kevinlu1248 8mo ago
Are you using the right format? https://huggingface.co/sweepai/sweep-next-edit-1.5B/blob/mai...
13.
by kevinlu1248 8mo ago
^ these were pretty much the main reasons. The other one is that constrained decoding only works on CFGs (simpler grammars like JSON schemas) since only these ones can produce automatas which can be used for constrained decoding. Programmin
14.
by kevinlu1248 8mo ago
Unfortunately, the main optimization (3x speedup) is using n-gram spec dec which doesn't run on CPUs. But I believe it works on Metal at least.
15.
by kevinlu1248 8mo ago
Haha, we never trained it for chat but I would bet it works regardless. Also that's crazy, M4 Mac?
16.
by kevinlu1248 8mo ago
Our hosted autocomplete is coming to Zed in a few weeks.
17.
by kevinlu1248 8mo ago
We'll push to Ollama
18.
by kevinlu1248 8mo ago
Let me know if you have any questions! What hardware are you on?
19.
by kevinlu1248 8mo ago
We're using an internal fork of trl for some of the steps.
20.
by kevinlu1248 8mo ago
Personally, I think usable AI is more valuable than simply more intelligence. Many of the labs are pushing towards models that are 1% better on CodeForces and AIME if you just let it think and use tools for hours, instead of more user-frien
21.
by kevinlu1248 8mo ago
Double-check if you're using the right format. Example here: https://huggingface.co/sweepai/sweep-next-edit-1.5B/blob/mai...
22.
by kevinlu1248 8mo ago
Woops meant to say open-weight. We put open-weight in the title and but accidentally wrote open-source in the description.
23.
by kevinlu1248 8mo ago
Someone in this thread already built a Neovim plugin connecting to this model I believe.
24.
by kevinlu1248 8mo ago
Also more technical details on SFT data here: https://blog.sweep.dev/posts/next-edit-jetbrains#building-au...
25.
by kevinlu1248 8mo ago
Thanks! Let us know if you have any questions / feedback.
26.
by kevinlu1248 8mo ago
Not at the moment, if you install the hosted Sweep AI Jetbrains plugin it uses our hosted (larger) model.
27.
by kevinlu1248 8mo ago
Let me know if you have any questions. We have a lot of harness code that cleans up many bad behaviours that makes it a lot more usable (like token healing: https://blog.sweep.dev/posts/token-healing-autocomplete ).
28.
by kevinlu1248 8mo ago
this is awesome, i'm going to try this out
29.
by kevinlu1248 10mo ago
I'm building a supermaven competitor for jetbrains. We use the Jetbrains PSI (basically Jetbrain's version of the LSP) to pull definitions into context to make the autocomplete smarter. My colleague wrote a blog on this here: htt
30.
Building sub-100ms autocompletion for JetBrains IDEs(blog.sweep.dev)
8 points by kevinlu1248 1y ago|1 comments

More ›