Lysandre's picture

Lysandre

lysandre

·

http://lysand.re

AI & ML interests

chief open-source officer @ hf

Recent Activity

posted an update about 3 hours ago

SmolVLM-2 and SigLIP-2 are now part of `transformers` in dedicated releases! They're added on top of the v4.49.0 release, and can be installed from the following tags: `v4.49.0-SmolVLM-2` and `v4.49.0-SigLIP-2`. This marks a new beginning for the release process of transformers. For the past five years, we've been doing monthly releases featuring many models (v4.49.0, the latest release, features 9 new architectures). Starting with SmolVLM-2 & SigLIP2, we'll now additionally release tags supporting new models on a stable branch. These models are therefore directly available for use by installing from the tag itself. These tags will continue to be updated with fixes applied to these models. Going forward, continue expecting software releases following semantic versioning: v4.50.0 will have ~10 new architectures compared to v4.49.0, as well as a myriad of new features, improvements and bug fixes. Accompanying these software releases, we'll release tags offering brand new models as fast as possible, to make them accessible to all immediately.

updated a dataset about 7 hours ago

huggingface/transformers-metadata

updated a dataset 1 day ago

huggingface/transformers-metadata

View all activity

Organizations

lysandre's activity

upvoted an article 10 days ago

Article

From Llasa to Llasagna 🍕: Finetuning LLaSA to generates Italian speech and other languages

By

and 1 other •

10 days ago

• 22

upvoted an article 13 days ago

Article

Welcome to Inference Providers on the Hub 🔥

25 days ago

• 380

upvoted a paper 4 months ago

Lina-Speech: Gated Linear Attention is a Fast and Parameter-Efficient Learner for text-to-speech synthesis

Paper • 2410.23320 • Published Oct 30, 2024 • 8

upvoted 2 articles 4 months ago

Article

Transformers.js v3: WebGPU support, new models & tasks, and more…

Oct 22, 2024

• 67

Article

Tool Use, Unified

Aug 12, 2024

• 83

upvoted a collection 5 months ago

Llama3-8B-1.58

A trio of powerful models: fine-tuned from Llama3-8b-Instruct, with BitNet architecture! • 3 items • Updated Sep 14, 2024 • 11

upvoted an article 5 months ago

Article

Fine-tuning LLMs to 1.58bit: extreme quantization made easy

Sep 18, 2024

• 223

upvoted 3 articles 6 months ago

Article

Don't repeat yourself - 🤗 Transformers Design Philosophy

Apr 5, 2022

• 18

Article

MobileNet Baselines

By

•

Jul 26, 2024

• 23

Article

MobileNet-V4 (now in timm)

By

•

Jun 17, 2024

• 43

upvoted an article 7 months ago

Article

WWDC 24: Running Mistral 7B with Core ML

Jul 22, 2024

• 58

upvoted a collection 8 months ago

Nemotron 4 340B

Nemotron-4: open models for Synthetic Data Generation (SDG). Includes Base, Instruct, and Reward models. • 4 items • Updated Jan 17 • 162

upvoted a collection 9 months ago

Embedding Model Datasets

A curated subset of the datasets that work out of the box with Sentence Transformers: https://huggingface.co/datasets?other=sentence-transformers • 68 items • Updated 7 days ago • 108

upvoted an article 9 months ago

Article

License to Call: Introducing Transformers Agents 2.0

May 13, 2024

• 129

upvoted an article 10 months ago

Article

LLM Comparison/Test: Llama 3 Instruct 70B + 8B HF/GGUF/EXL2 (20 versions tested and compared!)

By

•

Apr 24, 2024

• 61

upvoted a collection 11 months ago

Gemma release

Groups the Gemma models released by the Google team. • 40 items • Updated Dec 13, 2024 • 330

upvoted 2 collections about 1 year ago

Canonical models

This collection lists all the historical (pre-"Hub") canonical model checkpoints, i.e. repos that were not under an org or user namespace • 68 items • Updated Feb 13, 2024 • 14

SigLIP

Contrastive (sigmoid) image-text models from https://arxiv.org/abs/2303.15343 • 10 items • Updated 2 days ago • 53

upvoted a paper about 1 year ago

Exponentially Faster Language Modelling

Paper • 2311.10770 • Published Nov 15, 2023 • 118

upvoted a collection about 1 year ago

Switch-Transformers release

This release included various MoE (Mixture of expert) models, based on the T5 architecture . The base models use from 8 to 256 experts. • 9 items • Updated Dec 13, 2024 • 17