deoxykev

7 ms·

1.
Solving Automata Cam Profiling with Grasshopper(youtube.com)
2 points by deoxykev 7mo ago|0 comments
2.
by deoxykev 10mo ago
I live there in that city. There are hardly any homeless at all here. Not like other cities at least. I could see it being a major problem in other places.
3.
by deoxykev 1y ago
How about LLM chat over DNS? https://github.com/accupham/llm-dns-proxy
4.
by deoxykev 1y ago
And it typically works on captive portals too before payment.
5.
by deoxykev 1y ago
Meta-commentary always leans nerdier.
6.
by deoxykev 2y ago
Curious to hear what kind of work you do. Because there are definitely fields where productivity as 10x'd because of AI tools.
7.
by deoxykev 2y ago
HTMX and shoelace is an awesome combo. Super fast to prototype things and tweak as needed. Being able to copy paste snippets and directly inject data in a straightforward way is a nice way of working. It limits cognitive overhead so you can
8.
by deoxykev 2y ago
Don't forget to finetune the reranker too if you end up doing the embedding model. That tends to have outsized effects on performance for out of distribution content.
9.
by deoxykev 2y ago
Interesting, I had never heard about min-p until now. From what I understand, it's like a low-pass filter for the token sampling pool which boosts semantic coherence. Like removing static from the radio. Do you have any benchmarks of m
10.
by deoxykev 2y ago
Yeah, there is a clear bottleneck somewhere in llama.cpp. Even high end hardware is struggling to get good numbers. The theoretical limit should be higher, but it's not yet. Benchmarks: https://github.com/ggerganov&#x2F
11.
by deoxykev 2y ago
I don't think autoregressive models have a fundemental difference in terms of reasoning capability in latent space vs token space. Latent space enables abstract reasoning and pattern recognition, while token space acts as both the disc
12.
by deoxykev 2y ago
The fundemental challenge of using log probabilities to measure LLM certainty is the mismatch between how language models process information and how semantic meaning actually works. The current models analyze text token by token-- fragment
13.
by deoxykev 2y ago
My take: the distills under 32B aren’t worth running. Quants seem to impact quality much more than other models. 32B and 70B unquantized are very good. 671B is SOTA.
14.
by deoxykev 2y ago
8x 3090 will net you around 10-12tok/s
15.
by deoxykev 2y ago
Have you hit any non-determinism errors keeping workflow state outside temporal?
16.
by deoxykev 2y ago
Hey, I’m building agents on top of temporal as well. One of the main limitations is child workflows can not spawn other child workflows. Are you doing an activity for every prompt execution and passing those through other activities? Or som
17.
by deoxykev 2y ago
Imhex is a really great frontend for Capstone. https://github.com/WerWolv/ImHex
18.
by deoxykev 2y ago
Are you able to run 405B? 4Bit quant vram requirements are just shy of 192GB.
19.
by deoxykev 2y ago
4 bit quants should require 85GB VRAM, so this will fit nicely on 4x 24G consumer GPUs, plus some leftover for KV cache optimization.
20.
by deoxykev 2y ago
How does this compare to LayoutLMv3? Was it trained on forms at all?
21.
by deoxykev 3y ago
Hi there, I would be interested in a chat about those back-office patterns and use cases. Could you send an email to a2V2aW4gQCBkZW94eSAuIG5ldA==
22.
by deoxykev 3y ago
I use ansible to deploy and sync scripts, services, etc. I think it would work well for you use case as well.
23.
Subnet 6 Leaderboard by NousResearch(huggingface.co)
1 points by deoxykev 3y ago|0 comments
24.
by deoxykev 3y ago
Here is the list from the thread: --- 1. A Man Named Pearl 2. Once Upon a Time in Northern Ireland 3. Microcosmos 4. Crip Camp! 5. Keep The River On Your Right 6. All The Beauty And The Bloodshed 7. Harlan County USA 8. Stay on Board: The L
25.
by deoxykev 3y ago
A default key expiry of 3 days would be helpful here too, as it would mitigate the threat of someone compromising the endpoint and extracting secrets from browser history.
26.
by deoxykev 3y ago
That’s a great idea, I think something like this would work very well as a installable PWA even.
27.
by deoxykev 3y ago
This is cool. Do you mind me asking what videos you use it for?
28.
by deoxykev 3y ago
Here’s a similar project, but for windows AD networks https://github.com/byt3bl33d3r/DeathStar
29.
by deoxykev 3y ago
I would like to know this as well.
30.
by deoxykev 3y ago
I would love an API, and more flexible pricing options. A subscription just won’t work for my use case, which would be just once in a while for a hobby chatbot.

More ›