Faisal Waris
@fwaris7
Followers
192
Following
1K
Media
57
Statuses
1K
FsOperator updated to have to both voice and text chat modes, where voice uses the OpenAI real-time API.
github.com
A CUA (computer-use-agent) sample in F#. Contribute to fwaris/FsOperator development by creating an account on GitHub.
0
0
1
A CUA (computer-use-agent) sample in F#. It utilizes the new 'responses' API from OpenAI along with the new computer_use_preview model.
github.com
A CUA (computer-use-agent) sample in F#. Contribute to fwaris/FsOperator development by creating an account on GitHub.
1
0
7
F# (windows) service and GUI client (Avalonia, FuncUI) for transcription and diarization.
linkedin.com
I wrote a small utility service to transcribe and diarize video (.mp4) files (e.g. WebEx recordings) https://lnkd.in/gqweR5RE The service is meant to run on a GPU-enabled box. A small GPU box is...
0
0
3
Pretty impressive performance by o3 #OpenAI12Days. Still, o3 can consume $1K+ worth of electricity on complex tasks (@ megawatts). Human brains are in @ 25 watts range, so machines have some way to go, yet. However, with Googles quantum leap, the gap may be closing fast.
0
0
0
Experimental 'chat with code evaluator' feature added to FsOpenAI to generate and evaluate F# code interactively. (need own OpenAI key). Evaluation sandbox utilizes typed AST to only allow types from set namespaces.
0
1
11
OpenAI's o1 seems like a pretty capable F# code generator. See this post for more.
linkedin.com
OpenAI 'o1' model is a very capable code generator. I used o1 for a moderately complex code gen task and it accomplished it in the first try, given a relatively straightforward prompt! The task was...
0
2
8
New look for F# FsOpenAI rag/chat app. Test link below (need to use own OpenAI key) .Now with Radzen Blazor components.Source: (
2
6
13
Check out my latest article: Phi-3 Vision is a Surprisingly Useful Gem via @LinkedIn.
linkedin.com
My work involves building RAG applications for question-answering over highly technical internal company documents. The documents contain a wide variety of images, e.
0
0
0
F# example to serve ONNX vision models (e.g. phi-3 vision) for online and batch modes. Supports multiple parallel request if GPU hardware allows.
github.com
Vision processing with ONNX based GenAI vision models - fwaris/OnnxVision
0
11
23
RT @jeremyphoward: There's a new bill, SB-1047 "Safe and Secure Innovation for Frontier Artificial Intelligence Models Act". I think it co….
answer.ai
By imposing the restrictions on open-source AI, SB-1047 could reduce AI safety, through reducing transparency, collaboration, diversity, and resilience.
0
282
0
Check out my latest article: Constrained and Provable LLM Code Generation via @LinkedIn.
linkedin.com
LLMs are now good at generating code but human intervention is still required. Can't accept the generated code blindly - except perhaps for well-studied, academic tasks or trivial tasks.
0
1
1
Have been using F# interactive in VS Code to work on a large Linux VM from my local desktop using "Remote SSH". Works great! One issue: A long line of code may not compile (not sure why exactly). Works if you split code over multiple lines. Beats using Terminal for sure.
0
0
1
tussling with internal teams to acquire more compute capacity for my LLM apps. LLM compute capacity is the new scarce resource.
0
0
0
GitHub Co-Pilot is *very* responsive. Not sure exactly how it works - but knowing something about LLMs - I can surmise they (MS) are running it on pretty beefy infrastructure.
0
0
2