OPEN SOURCE

–

08 November 2024 7h34

I went for Triton because it was already closely integrated into PyTorch. Taichi looks really polished in comparison, but not tried it yet!

→ View original post on X — @_akhaliq

8 November 2024

Ollama Model Interaction Gradio App for Text and Image Analysis

By

@_akhaliq

–

08 November 2024 6h13

Ollama-Model-Interaction A simple Gradio-based app for interacting with @ollama models, supporting image analysis, text completion, and model pulling

8 November 2024

Etched Oasis 500M Model Released on Hugging Face

By

@_akhaliq

–

07 November 2024 16h34

model: https://
huggingface.co/Etched/oasis-5
00m
…

→ View original post on X — @_akhaliq

→ View original post on X — @_akhaliq

Oasis 500M: Open Model for Video Game Generation Trending

By

@_akhaliq

–

07 November 2024 16h34

Oasis 500M, the open model for video game generation from @DecartAI and @Etched is #1 trending on @huggingface

→ View original post on X — @mer__edith

Signal’s Secure Model Outpaces Stagnating Big Tech Industry

By

@mer__edith

–

07 November 2024 13h24

….which shows how broken the surveillance biz model powering the current centralized tech industry is. Signal is the nervous system of secure comms globally. We are innovating while big tech is stagnating. Yet we’re fundraising donations while boring apps white labeling big

Triton Library Lacks 3.0-Level Maturity: API Limitations

By

–

07 November 2024 12h19

I don't just mean the limitations of the GPU, that's part of it but done that before. Triton itself doesn't feel like a 3.0-level library. Stuff like http://
tl.dot() doesn't work when you specify the output array as a parameter (per docs), you have to use the retval.

Triton MLP Implementation Achieves 70x Performance Improvement

By

–

07 November 2024 12h10

Triton is exciting to work with. I built a type of MLP with it and it's 70x faster than the Torch version… Conversely, the set of features that actually works is so razor thin you have to code-golf every single decision for hours! https://
github.com/triton-lang/tr
iton
…

ReMoE Implementation Optimization Challenges Wall-Clock Performance

By

–

07 November 2024 12h00

Thought the same thing and was trying to reimplement ReMoE over the past days… Achieving the same level of utilization as the default codebase, and thus beating the wall-clock record, will not be trivial.