vishaal27

tiktoken

tiktoken is a fast BPE tokeniser for use with OpenAI's models.

Inflection-Benchmarks

Public Inflection Benchmarks

dbrx

Code examples and resources for DBRX, a large language model developed by Databricks

LAMM

[NeurIPS 2023 Datasets and Benchmarks Track] LAMM: Multi-Modal Large Language Models and Applications as AI Agents

attention-interpolation-diffusion

Interpolation Between Text-to-Image Generation!

modelgauge

Make it easy to automatically and uniformly measure the behavior of many AI Systems.

LLM-SLERP-Merge

Spherical Merge Pytorch/HF format Language Models with minimal feature loss.

Visual-CoT

Visual CoT: Advancing Multi-Modal Language Models with a Comprehensive Dataset and Benchmark for Chain-of-Thought Reasoning

DreamLIP

[ECCV 2024] Official PyTorch implementation of DreamLIP: Language-Image Pre-training with Long Captions

ALLaVA

Harnessing 1.4M GPT4V-synthesized Data for A Lite Vision-Language Model

mixinglaws

visual_diversity_budget

Annotations on a Budget: Leveraging Geo-Data Similarity to Balance Model Performance and Annotation Cost

emergent_in_context_learning

skerch

Sketched matrix decompositions for PyTorch

mup

maximal update parametrization (µP)

Long-CLIP

[ECCV 2024] official code for "Long-CLIP: Unlocking the Long-Text Capability of CLIP"

evolutionary-model-merge

Official repository of Evolutionary Optimization of Model Merging Recipes

VL-ICL

Code for paper: VL-ICL Bench: The Devil in the Details of Benchmarking Multimodal In-Context Learning

scaling_on_scales

When do we not need larger vision models?

DiversitySSL

ml-tic-clip

Repository for the paper: "TiC-CLIP: Continual Training of CLIP Models".

clipcov-data-efficient-clip

Code for AISTATS Eff MML

Chain-of-Spot

Chain-of-Spot: Interactive Reasoning Improves Large Vision-language Models

img2img-turbo

One-step image-to-image with Stable Diffusion turbo: sketch2image, day2night, and more

Mind2Web

[NeurIPS'23 Spotlight] "Mind2Web: Towards a Generalist Agent for the Web"

lilac

Curate better data for LLMs

that_is_good_data

HPT

HPT - Open Multimodal LLMs from HyperGAI

mnms

m&ms: A Benchmark to Evaluate Tool-Use for multi-step multi-modal tasks

Meta-Prompting

Meta-Prompting for Automating Zero-shot Visual Recognition with LLMs (ECCV 2024)