From ad4d5162857ade2b6b7a309614e5ba58a5f0e468 Mon Sep 17 00:00:00 2001 From: xiangan Date: Tue, 30 Dec 2025 19:28:04 +0800 Subject: [PATCH] docs: update README.md and add data card for OneVision Encoder training data --- README.md | 8 ++++---- docs/{datacard.md => data_card.md} | 5 ++++- 2 files changed, 8 insertions(+), 5 deletions(-) rename docs/{datacard.md => data_card.md} (94%) diff --git a/README.md b/README.md index a3fa235b..c70b0d67 100644 --- a/README.md +++ b/README.md @@ -14,9 +14,9 @@ 📝 **[Homepage](https://www.lmms-lab.com/onevision-encoder/index.html)** 🤗 **[Models](https://huggingface.co/lmms-lab-encoder/onevision-encoder-large)** | -🤗 **[Datasets](coming)** | 📄 **[Tech Report (coming)]()** | -📋 **[Model Card](docs/model_card.md)** +📋 **[Model Card](docs/model_card.md)** | +📊 **[Data Card](docs/data_card.md)** @@ -283,7 +283,7 @@ cd eval_encoder Then run the following command: ```bash -bash eval_encoder/shells_eval_ap/eval_ov_encoder_large_16frames.sh +bash shells_eval_ap/eval_ov_encoder_large_16frames.sh ``` **Sampling-Specific Parameters:** @@ -320,8 +320,8 @@ torchrun --nproc_per_node=8 --master_port=29512 attentive_probe_codec.py \ **Codec-Specific Parameters:** - `K_keep`: Number of patches to keep. -- `cache_dir`: Directory for cached codec patches. This is where the codec-selected patches will be stored/loaded. - `mv_compensate`: Motion vector compensation method (e.g., `median`). +- `cache_dir` (optional): Directory for cached codec patches. Use this to specify where codec-selected patches are stored/loaded when you want to persist or reuse them. #### Shared Parameters diff --git a/docs/datacard.md b/docs/data_card.md similarity index 94% rename from docs/datacard.md rename to docs/data_card.md index 2aae3ea7..f017a659 100644 --- a/docs/datacard.md +++ b/docs/data_card.md @@ -1,8 +1,11 @@ # Data Card: OneVision Encoder Training Data +> **📦 Data Availability Notice:** The training data requires approximately **200TB** of storage. We are currently looking for suitable storage solutions. If you need access to the data immediately, please contact [anxiangsir@outlook.com](mailto:anxiangsir@outlook.com). + + ## Overview -This document describes the datasets used for training OneVision Encoder. The training data consists of both image and video datasets, totaling approximately 754 million samples. +This document describes the datasets used for training OneVision Encoder. The training data consists of both image and video datasets. ## Dataset Summary