ucyang (Unchun Yang)

upvoted a collection about 15 hours ago

NuminaMath

Datasets and models for training SOTA math LLMs. See our GitHub for training & inference code: https://github.com/project-numina/aimo-progress-prize • 6 items • Updated about 22 hours ago • 29

upvoted 2 papers 1 day ago

MiraData: A Large-Scale Video Dataset with Long Durations and Structured Captions

Paper • 2407.06358 • Published 13 days ago • 16

GRUtopia: Dream General Robots in a City at Scale

Paper • 2407.10943 • Published 7 days ago • 20

upvoted an article 2 days ago

Article

Google Cloud TPUs made available to Hugging Face users

13 days ago

• 18

upvoted a collection 3 days ago

DCLM

Collection

DCLM Models + Datasets • 6 items • Updated 4 days ago • 25

upvoted a paper 3 days ago

Shape of Motion: 4D Reconstruction from a Single Video

Paper • 2407.13764 • Published 4 days ago • 14

upvoted a paper 4 days ago

YouTube-SL-25: A Large-Scale, Open-Domain Multilingual Sign Language Parallel Corpus

Paper • 2407.11144 • Published 7 days ago • 7

upvoted an article 5 days ago

Article

How NuminaMath Won the 1st AIMO Progress Prize

11 days ago

• 67

upvoted a collection 5 days ago

🪐 SmolLM

Collection

A series of smol LLMs: 135M, 360M and 1.7B. We release base and Instruct models as well as the training corpus and some WebGPU demos • 12 items • Updated 6 days ago • 98

upvoted an article 5 days ago

Article

SmolLM - blazingly fast and remarkably powerful

6 days ago

• 151

upvoted a paper 7 days ago

SpreadsheetLLM: Encoding Spreadsheets for Large Language Models

Paper • 2407.09025 • Published 10 days ago • 105

upvoted a paper 9 days ago

Vision language models are blind

Paper • 2407.06581 • Published 13 days ago • 73

upvoted a paper 10 days ago

Q-GaLore: Quantized GaLore with INT4 Projection and Layer-Adaptive Low-Rank Gradients

Paper • 2407.08296 • Published 11 days ago • 28

upvoted 2 articles 10 days ago

Article

MMLU-Pro-NoMath

By

•

11 days ago

• 3

Article

MInference 1.0: 10x Faster Million Context Inference with a Single GPU

By

•

11 days ago

• 9

upvoted 2 papers 11 days ago

AgentInstruct: Toward Generative Teaching with Agentic Flows

Paper • 2407.03502 • Published 18 days ago • 34

PaliGemma: A versatile 3B VLM for transfer

Paper • 2407.07726 • Published 12 days ago • 58

upvoted 2 papers 12 days ago

LLaMAX: Scaling Linguistic Horizons of LLM by Enhancing Translation Capabilities Beyond 100 Languages

Paper • 2407.05975 • Published 14 days ago • 32

Unveiling Encoder-Free Vision-Language Models

Paper • 2406.11832 • Published Jun 17 • 45

upvoted 4 papers 14 days ago

Diffusion Forcing: Next-token Prediction Meets Full-Sequence Diffusion

Paper • 2407.01392 • Published 21 days ago • 39

Let the Expert Stick to His Last: Expert-Specialized Fine-Tuning for Sparse Architectural Large Language Models

Paper • 2407.01906 • Published 20 days ago • 33

InternLM-XComposer-2.5: A Versatile Large Vision Language Model Supporting Long-Contextual Input and Output

Paper • 2407.03320 • Published 19 days ago • 87

MobileLLM: Optimizing Sub-billion Parameter Language Models for On-Device Use Cases

Paper • 2402.14905 • Published Feb 22 • 103

upvoted 3 papers 15 days ago

Summary of a Haystack: A Challenge to Long-Context LLMs and RAG Systems

Paper • 2407.01370 • Published 21 days ago • 81

We-Math: Does Your Large Multimodal Model Achieve Human-like Mathematical Reasoning?

Paper • 2407.01284 • Published 21 days ago • 72

HuatuoGPT-Vision, Towards Injecting Medical Visual Knowledge into Multimodal LLMs at Scale

Paper • 2406.19280 • Published 25 days ago • 55

upvoted 13 papers 16 days ago

OMG-LLaVA: Bridging Image-level, Object-level, Pixel-level Reasoning and Understanding

Paper • 2406.19389 • Published 25 days ago • 51

Math-LLaVA: Bootstrapping Mathematical Reasoning for Multimodal Large Language Models

Paper • 2406.17294 • Published 27 days ago • 9

Towards Fast Multilingual LLM Inference: Speculative Decoding and Specialized Drafters

Paper • 2406.16758 • Published 28 days ago • 18

VideoHallucer: Evaluating Intrinsic and Extrinsic Hallucinations in Large Video-Language Models

Paper • 2406.16338 • Published 28 days ago • 23

LongRAG: Enhancing Retrieval-Augmented Generation with Long-context LLMs

Paper • 2406.15319 • Published about 1 month ago • 57

PIN: A Knowledge-Intensive Dataset for Paired and Interleaved Multimodal Documents

Paper • 2406.13923 • Published Jun 20 • 21

HARE: HumAn pRiors, a key to small language model Efficiency

Paper • 2406.11410 • Published Jun 17 • 38

nabla^2DFT: A Universal Quantum Chemistry Dataset of Drug-Like Molecules and a Benchmark for Neural Network Potentials

Paper • 2406.14347 • Published Jun 20 • 99

MMBench-Video: A Long-Form Multi-Shot Benchmark for Holistic Video Understanding

Paper • 2406.14515 • Published Jun 20 • 29

GLiNER multi-task: Generalist Lightweight Model for Various Information Extraction Tasks

Paper • 2406.12925 • Published Jun 14 • 20

upvoted a collection 21 days ago

LLM Compiler

Collection

Meta LLM Compiler is a state-of-the-art LLM that builds upon Code Llama with improved performance for code optimization and compiler reasoning. • 4 items • Updated 25 days ago • 140

upvoted a paper 21 days ago

Scaling Synthetic Data Creation with 1,000,000,000 Personas

Paper • 2406.20094 • Published 24 days ago • 85

upvoted 2 articles 22 days ago

Article

Falcon 2: An 11B parameter pretrained language model and VLM, trained on over 5000B tokens tokens and 11 languages

May 24

• 23

Article

Welcome Gemma 2 - Google's new open LLM

25 days ago

• 97

upvoted a collection 23 days ago

GLM-4

Collection

GLM-4 Open Models • 4 items • Updated Jun 5 • 88

upvoted a paper 24 days ago

A Technical Report for Polyglot-Ko: Open-Source Large-Scale Korean Language Models

Paper • 2306.02254 • Published Jun 4, 2023 • 11

upvoted a collection 24 days ago

Instruction Pre-Training

Collection

8 items • Updated about 1 month ago • 24

upvoted 2 articles 24 days ago

Article

Cosmopedia: how to create large-scale synthetic data for pre-training Large Language Models

Mar 20

• 41

Article

Ethics and Society Newsletter #6: Building Better AI: The Importance of Data Quality

28 days ago

• 28

upvoted a collection 24 days ago

Gemma 2 Release

Collection

10 items • Updated 25 days ago • 129

upvoted an article 25 days ago

Article

Decoding GPT-4'o': In-Depth Exploration of Its Mechanisms and Creating Similar AI.

By

•

May 21

• 28

upvoted a paper 25 days ago

Multimodal Needle in a Haystack: Benchmarking Long-Context Capability of Multimodal Large Language Models

Paper • 2406.11230 • Published Jun 17 • 34

upvoted 2 papers 26 days ago

LLaNA: Large Language and NeRF Assistant

Paper • 2406.11840 • Published Jun 17 • 17

VideoLLM-online: Online Video Large Language Model for Streaming Video

Paper • 2406.11816 • Published Jun 17 • 20

upvoted 2 collections 26 days ago

Cambrian Data

Collection

3 items • Updated 27 days ago • 7

Cambrian-1 Models

Collection

6 items • Updated 24 days ago • 19

upvoted 5 papers 26 days ago

Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMs

Paper • 2406.16860 • Published 28 days ago • 52

BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructions

Paper • 2406.15877 • Published 30 days ago • 43

DreamBench++: A Human-Aligned Benchmark for Personalized Image Generation

Paper • 2406.16855 • Published 28 days ago • 53

The FineWeb Datasets: Decanting the Web for the Finest Text Data at Scale

Paper • 2406.17557 • Published 27 days ago • 75

mDPO: Conditional Preference Optimization for Multimodal Large Language Models

Paper • 2406.11839 • Published Jun 17 • 36

Unchun Yang

AI & ML interests

Organizations

ucyang's activity

Google Cloud TPUs made available to Hugging Face users

How NuminaMath Won the 1st AIMO Progress Prize

SmolLM - blazingly fast and remarkably powerful

MMLU-Pro-NoMath

MInference 1.0: 10x Faster Million Context Inference with a Single GPU

Falcon 2: An 11B parameter pretrained language model and VLM, trained on over 5000B tokens tokens and 11 languages

Welcome Gemma 2 - Google's new open LLM

Cosmopedia: how to create large-scale synthetic data for pre-training Large Language Models

Ethics and Society Newsletter #6: Building Better AI: The Importance of Data Quality

Decoding GPT-4'o': In-Depth Exploration of Its Mechanisms and Creating Similar AI.