From 698e2b44f59ecc2c67e9e4f95948577ef61db5d1 Mon Sep 17 00:00:00 2001 From: Xiang An Date: Sun, 2 Nov 2025 17:32:47 +0800 Subject: [PATCH] Refactor README.md for improved clarity and structure Revised the README to enhance clarity and structure, updating section headings and improving descriptions of model performance, data quality, training framework, and community resources. --- profile/README.md | 32 +++++++++++++------------------- 1 file changed, 13 insertions(+), 19 deletions(-) diff --git a/profile/README.md b/profile/README.md index da724a0..af9ce46 100644 --- a/profile/README.md +++ b/profile/README.md @@ -26,29 +26,23 @@ We are a group of researchers, with a focus on large multimodal models (LMMs). W ## Introduction -**LLaVA-OneVision1.5** introduces a novel family of **fully open-source** Large Multimodal Models (LMMs) that achieves **state-of-the-art performance** with substantially **lower cost** through training on **native resolution** images. +**LLaVA-OneVision-1.5** introduces a family of fully open-source large multimodal models (LMMs) that operate on **native-resolution images**, achieve **state-of-the-art** performance, and require comparatively **lower training costs**. -- **Superior Performance** -A family of fully open-source large multimodal models demonstrating - - Superior performance across multiple multimodal benchmarks - - outperforming **Qwen2.5-VL** in most evaluation tasks. +#### **Superior Performance** + - The model leads on multiple multimodal benchmarks and generally surpasses Qwen2.5-VL. + - Training on native-resolution images significantly improves its visual understanding. -- **High-Quality Data at Scale** -Meticulously curated **pre-training and SFT data** with rigorous filtering and quality control. - - Concept-balanced, highly diverse, high-quality caption data - - Comprehensive instruction fine-tuning data covering a wide range of tasks +#### **High-Quality Data at Scale** + - The pretraining corpus comprises large-scale, concept-balanced, diverse, and high-quality captions curated with strict filtering and quality control. + - The instruction-tuning dataset is comprehensive and covers a wide range of tasks. -- **Ultra-Efficient Training Framework** Complete end-to-end training framework designed for maximum efficiency: - - $16000 total budget for full model training on A100 GPUs ($0.6 per GPU/Hour) - - Built on **MegatronLM** with support for **MoE**, **FP8**, and **long sequence parallelization** - - Optimized codebase for cost-effective scaling +#### **Ultra-Efficient Training Framework** + - The end-to-end training cost is about $16,000 on A100 GPUs at roughly $0.60 per GPU-hour. + - The system is built on Megatron-LM with support for MoE, FP8, and long-sequence parallelism, and the codebase is optimized for cost-effective scaling. - -- **Fully Open Framework** for community access and reproducibility: - - High-quality pre-training & SFT data - - Complete training framework & code - - Training recipes & configurations - - Comprehensive training logs & metrics +#### **Fully Open Framework** + - The project releases high-quality pretraining and SFT datasets along with the complete training framework, configurations, and recipes. + - It also provides detailed training logs and metrics to enable reproducibility and community adoption.