Stories by tosh

Mononoke

Sapling: A Scalable, User-Friendly Source Control System

Intel Lunar Lake-MX SoC with On-Package LPDDR5X Memory Detailed

Hound: Fast code searching made easy

Chord

Deploy and serve any LLM on HF with a single command

MisterHouse

Tulu-2-Dpo-70B

Video-LLaVA

UIx, a story behind yet another React wrapper [video]

Yarn-Llama-2-70B-32k

Hydra – Model of Experts

Assisted Generation: a new direction toward low-latency text generation

The language strangeness budget (2015)

A Note to OpenAI Developers

Switch Transformers C – 2048 experts (1.6T params for 3.1 TB) (2022)

TinyLlama at 2T of 3T

310 311 312 313 314 315 316 317 318 319