foxhop

7 ms·

31.
Comparing Node.js and Python Performance with the Official OpenAI Client(russell.ballestrini.net)
1 points by foxhop 2y ago|4 comments
32.
by foxhop 2y ago
I conducted a performance comparison between Node.js and Python using the official OpenAI client, and the results were quite revealing. Node.js outperformes Python in this specific use case.
33.
by foxhop 3y ago
that's where I arrive too
34.
by foxhop 3y ago
yeah I just tested this, autoplay only works on the video tag if muted which seems like it defeats the point.
35.
by foxhop 3y ago
did you know that browsers will not autoplay video tags unless the video is also muted?
36.
by foxhop 3y ago
these ships have to be boarded by special port captians to make this maneuver.
37.
by foxhop 3y ago
no but current LLM can already almost give 1:1 support and differentiated assistance based on how the student comes. I'm in edtech building this. it's like having 1 teacher with 25 assistants.
38.
by foxhop 3y ago
a set of patches to run on the Repo works but it breaks when a big refactor happens upstream.
39.
by foxhop 3y ago
If you want to see a cool beta game in a related genre check out https://luckeyproductions.itch.io/park I've been messing with the game engine, Dry the successor to Urho3D on my blog lately
40.
by foxhop 3y ago
I think kubernetes is a mistake and should have went with AWS ECS (using fargate or backed by autoscaling ec2), if single change he wouldn't need to even thing about a bunch of other topics on his list. Something to think about, AWS La
41.
by foxhop 3y ago
https://github.com/russellballestrini/flask-socketio-llm-com... This project is a chatroom application that allows users to join different chat rooms, send messages, and interact with multiple language models in real-t
42.
by foxhop 3y ago
let's start by turning the water and electricity back on
43.
by foxhop 3y ago
vLLM doesn't support quantized models at this time so you need 2x 4090 to run Mixtral. llama.cpp supports quantized models so that makes sense, ollama must have picked a quantized model to make it fit?
44.
by foxhop 3y ago
openchat crashes when you ask it to import certain javascript libraries because it will hallucinate a never ending integrity string, you can give it examples and ask it not to but yeah it locks up the inference server.
45.
by foxhop 3y ago
I'm working on this problem in my spare time, it requires a GPU (or two) RTX 3090 or 4090 because of the need for vRAM memory but I could also forsee APIs in front that allow worker nodes to run the quantized models for cheaper. the pr
46.
by foxhop 3y ago
are you using a quantized version of the model and if so which one?
47.
by foxhop 3y ago
great article I read a few posts but you say you coined the term "prompt injection" but I'm fairly certain that was a thing discovered by anyone who thinks about using LLMs in production. sql injection/prompt injection
48.
by foxhop 3y ago
performant != high performance he used it twice so I felt compelled to state this
49.
by foxhop 3y ago
I use automatic1111 hires fix with SDXL-turbo on a RTX 4090 and it looks the best by far for high quality images once you find a good prompt/seed (but heres the thing turning it on makes all the prompts/seeds that much better...)
50.
by foxhop 3y ago
Thanks for the summary.
51.
by foxhop 3y ago
I don't understand why there are 31 layers in this tool. I don't understand why we only see 4 charts (I would expect to see 8, one for each 7B mistral tiny model)
52.
by foxhop 3y ago
next year I'm focusing on learning more about AI, I bought a $3.5k computer with an nvidia RTX 4090 24G of memory, it arrived today, I plan to extend my playground chat app to use local models for text generation and image generation i
53.
by foxhop 3y ago
https://www.youtube.com/watch?v=pd3shNtSojY short video showing how it works
54.
Show HN: Alternative Front End to GPT, Claude, or Mistral Flask/Socketio(github.com)
4 points by foxhop 3y ago|2 comments
55.
by foxhop 3y ago
Hey all, I've been lucky enough to sprint on this open source project during the last two weeks during work, a bunch of new features, you can now edit any message and also Mistral inference platform was added.
56.
by foxhop 3y ago
This feels like something an LLM would be good at producing and also catching.
57.
by foxhop 3y ago
https://github.com/russellballestrini/flask-socketio-llm-com... I'm building a similar app but uses python/socket.io
58.
by foxhop 3y ago
let me know if you try it out, and if you are looking for a dangerous mode checkout the python-execution branch, it adds a `/python` command to run the most recent code block in a python exec
59.
Show HN: Chatroom app to interact with multiple LLMs in real-time(github.com)
3 points by foxhop 3y ago|1 comments
60.
by foxhop 3y ago
Human: I have a message and I need to categorize it into the right bucket. Can you help? General Inquiry Feedback & Suggestions Technical Issues Billing & Payments Appointments & Scheduling Others re

More ›