Videomae-base-finetuned Open-source Video Understanding Model - Free Deployment, Accuracy on Evaluation Set Reaches 86.41%

Videomae Base Finetuned

Developed by LouisDT

A video understanding model fine-tuned on an unknown dataset based on the VideoMAE base model, achieving 86.41% accuracy on the evaluation set

Video Processing

Transformers

#Video Understanding #High Accuracy #Action Recognition

Downloads 15

Release Time : 2/8/2023

Model Overview

This model is a fine-tuned version of the VideoMAE base architecture, primarily used for video content understanding tasks. Specific application scenarios require further details

Model Features

Efficient Video Representation Learning

Utilizes a masked autoencoder architecture to effectively learn spatiotemporal feature representations of videos

Excellent Fine-tuning Performance

Achieves 86.41% accuracy on the evaluation set, demonstrating strong performance

Lightweight Training

Can be effectively trained with a batch size of 8

Model Capabilities

Video Feature Extraction

Video Content Classification

Spatiotemporal Pattern Recognition

Use Cases

Video Content Analysis

Action Recognition

Identify human actions or behaviors in videos

86.41% accuracy (based on evaluation set)

Scene Classification

Classify video scene content

Property	Details
Model Type	videomae - base - finetuned
Metrics	accuracy
License	cc - by - nc - 4.0

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.7163	0.21	28	0.6078	0.8098
0.7383	1.21	56	0.6975	0.4728
0.6853	2.21	84	0.6637	0.6957
0.7065	3.21	112	0.5590	0.8641
0.6673	4.17	135	0.5766	0.8587

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご

Videomae Base Finetuned

Model Overview

Model Features

Model Capabilities

Use Cases

🚀 videomae-base-finetuned

🚀 Quick Start

📚 Documentation

Model Information

Training and Evaluation

Training Hyperparameters

Training Results

Framework Versions

📄 License