ilovefood

9 ms·

31.
by ilovefood 2y ago
Looks pretty cool, congrats so far! Do you allow downloading the fine tuned model for local inference?
32.
by ilovefood 2y ago
It isn't, the blog was just a high level recap of what I built, as I unfortunately don't have enough time to turn it into a product that can be used. I'm very happy to see that someone is building it though, I really think th
33.
by ilovefood 2y ago
Looks great :) it's exactly what I've built a year ago: LLM in the middle / client side content filtering using LLms. - https://karimjedda.com/llms-in-the-middle-content-aware-clie...
34.
by ilovefood 2y ago
For solo projects, my code is the project planning. I use a bunch of #TODO or //TODO in different files throughout each project's codebase, that a small Python script picks up and compiles into a single HTML file. As I'm
35.
Data Engineering with Rust(preview.datawithrust.com)
2 points by ilovefood 2y ago|0 comments
36.
Data Engineering with Rust – An online book(preview.datawithrust.com)
1 points by ilovefood 2y ago|1 comments
37.
by ilovefood 2y ago
Hi all, after a lot of feedback I've pushed the new version of my Data Engineering with Rust online book. It's free for now and I've ported the first chapters from the previous version, would love to hear what you think ! Kar
38.
Mix and match local and online LLMs with Open WebUI(twitter.com)
2 points by ilovefood 2y ago|0 comments
39.
by ilovefood 3y ago
Beautiful and well written article. Unfortunately I made a small mistake in the second quizz :P
40.
by ilovefood 3y ago
I love Pandoc! I recently learned you can use LUA to write custom plugins and change some of the converting behavior. I'm using it for example to create slides similar to the "sent" program. It helps me bootstrap new presenta
41.
by ilovefood 3y ago
The article is good, and I get the point. However, from experience, this: > POST on /queries/enlisted-students-on-joining-date/version/1 { "date": "2023-09-22" } to retrieve all students that joine
42.
by ilovefood 3y ago
> There is promise in constraining output to be valid JSON. One new trick that the open-source llama.cpp project has popularized is generative grammars This has been working for months now and is the best method for this type of stuff, a
43.
by ilovefood 3y ago
Thank you very much for the supporting words, I've been getting lots of positive feedback from this. The end of year workload has made it such though that I need to be mindful of time. I think one of the two first options will be the w
44.
by ilovefood 3y ago
I would say, bring it on! Nothing will make it past my phi-2 or mistral-7B-v0.1 ^^, at least for now. I think what this could lead to is homogenization of the content serving layer, since all you'd really need is to get content to the
45.
by ilovefood 3y ago
I'm using it to filter out the content that's displayed in my browser screen as I browse: https://karimjedda.com/llms-in-the-middle-content-aware-clie... Essentially, I wrote a small browser extension, that takes
46.
by ilovefood 3y ago
Amazing, I was eagerly waiting for this one. Loading extensions in previous DuckDB-WASM releases didn't work seamlessly. Looks like now it's the case :D ref: https://github.com/duckdb/duckdb-wasm/issues&#
47.
LLMs in the middle: Filtering social-media content in the browser using LLMs(karimjedda.com)
2 points by ilovefood 3y ago|0 comments
48.
by ilovefood 3y ago
Full agreement, but since these sort of approaches are new, I'm treading very carefully here. It was very important to me to have this working locally first and prove the concept is sound. One thing is sure: my browser activity changed
49.
LLMs in the middle: Content aware client-side filtering(karimjedda.com)
2 points by ilovefood 3y ago|2 comments
50.
by ilovefood 3y ago
I have been using this with a lot of success for a while now: https://github.com/KoljaB/RealtimeSTT/tree/master , it works in real time, without any delays on an old Nvidia card. I tried it with German &
51.
by ilovefood 3y ago
It's okay I have to say. I just ran out of memory on my 4090, so I had to retry on an A100. Here's an extract: https://pastebin.com/pzLfCFWt I think something might be off with the example. Can't wait for thi
52.
by ilovefood 3y ago
This is working relatively well, the code is really worth a read. If you run it locally, consider the open PR and install sentencepiece as well. It's been generating text for the past 10 minutes now :D Some of the instructions are igno
53.
by ilovefood 3y ago
I've been using this library a lot, it's amazing. However, I noticed a very considerable degradation (time taken + generation quality) with versions > 0.0.58 when used with local LLMs. I haven't taken time to compare betwe
54.
Why many smaller bets might beat a single big one(karimjedda.com)
1 points by ilovefood 3y ago|0 comments
55.
by ilovefood 3y ago
I've been tinkering quite a bit with WasmCloud and WasmEdge lately, for some data workloads (scraping, ingesting, apis etc) mainly as an excuse to learn WebAssembly. What I want to see for myself is how Unit compares to those for this
56.
by ilovefood 3y ago
This is really great! What I'm really look forward to trying is the server side WebAssembly: https://unit.nginx.org/configuration/#configuration-wasm . Huge potential!
57.
by ilovefood 3y ago
Thanks for the positivity, I'll give it a shot.
58.
by ilovefood 3y ago
Thank you for the kind words. I wouldn't even know where to start. There's a lot I can code and do, but making something into a product is something I have no experience with. What's your recommended approach? Even if open-so
59.
by ilovefood 3y ago
I took this a step further, turning charts into infographics using stable diffusion (SDXL) https://karimjedda.com/beautiful-data-visualizations-powered...
60.
Transform boring charts into beautiful information with Stable Diffusion(karimjedda.com)
2 points by ilovefood 3y ago|0 comments

More ›