# Action Recognition

A video understanding model fine-tuned based on MCG-NJU/videomae-base, with average performance on the evaluation set (accuracy 50%)

A video understanding model fine-tuned from MCG-NJU/videomae-base, trained on an unknown dataset with an accuracy of 11.1%

A video understanding model fine-tuned based on MCG-NJU/videomae-base, trained on an unknown dataset

A video understanding model based on the TimeSformer architecture, fine-tuned on an unknown dataset with an accuracy of 88.52%

A sign language classification model fine-tuned based on Google Vision Transformer (ViT), achieving an accuracy of 80.56%

A vision model fine-tuned from the VideoMAE base model on an unknown video dataset, primarily used for video understanding tasks, achieving 85.59% accuracy on the evaluation set.

VideoMAE is a video pretraining model based on Masked Autoencoder (MAE), fine-tuned on the Kinetics-400 dataset through self-supervised learning, suitable for video classification tasks.

A video understanding model fine-tuned on an unknown dataset based on MCG-NJU/videomae-base, achieving an F1 score of 0.7147

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

TimeSformer is a video understanding model based on spatio-temporal attention mechanisms, pre-trained and fine-tuned on the Kinetics-400 dataset.

X-CLIP is an extended version of CLIP for general video-language understanding, supporting video classification and video-text retrieval tasks.

X-CLIP is an extended version of CLIP for general video-language understanding, trained through contrastive learning on video-text pairs, suitable for video classification and video-text retrieval tasks.

X-CLIP is an extended version of CLIP for general video-language understanding, trained via contrastive learning on video-text pairs, supporting zero-shot, few-shot, and fully supervised video classification tasks.

VideoMAE is a video self-supervised pretraining model based on Masked Autoencoder (MAE), fine-tuned on the Something-Something-v2 dataset for video classification tasks.

VideoMAE is a video self-supervised pre-training model based on Masked Autoencoder (MAE), fine-tuned on the Kinetics-400 dataset for video classification tasks.

A hybrid CNN-RNN architecture-based video classification model for action recognition tasks