Videomae-base-finetuned-subset Open-source Video Understanding Model - Free Deployment with High-precision Video Understanding

Videomae Base Finetuned Subset

Developed by Joy28

A video understanding model fine-tuned on an unknown dataset based on the MCG-NJU/videomae-base model, with an accuracy of 67.13%

Video Processing

Transformers

#Video understanding #Behavior recognition #Temporal modeling

Downloads 2

Release Time : 12/30/2023

Model Overview

This model is a fine-tuned version of the VideoMAE base model, focusing on video content understanding tasks. Specific application scenarios need to be supplemented.

Model Features

Video feature extraction

Pretrained on video masked autoencoding based on the VideoMAE architecture, good at capturing video temporal features

Efficient fine-tuning

Perform transfer learning on the base model to adapt to specific video understanding tasks

Model Capabilities

Video content classification

Temporal feature extraction

Video semantic understanding

Use Cases

Video analysis

Action recognition

Recognize human actions or behaviors in videos

Accuracy of 67.13% (based on the current evaluation set)

Scene classification

Classify the content of video scenes

🚀 videomae-base-finetuned-subset

This model is a fine - tuned version of [MCG - NJU/videomae - base](https://huggingface.co/MCG - NJU/videomae - base) on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.7700
Accuracy: 0.6713

🚀 Quick Start

No specific quick - start content is provided in the original document.

✨ Features

No feature descriptions are provided in the original document.

📦 Installation

No installation steps are provided in the original document.

💻 Usage Examples

No code examples are provided in the original document.

📚 Documentation

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e - 05
train_batch_size: 4
eval_batch_size: 4
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
training_steps: 11100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.638	0.01	112	1.5736	0.1567
1.5845	1.01	224	1.5841	0.2719
1.4522	2.01	336	1.6293	0.2350
1.3111	3.01	448	1.0450	0.6037
1.2849	4.01	560	1.3186	0.4608
1.3246	5.01	672	1.1759	0.5161
1.3801	6.01	784	1.2188	0.4608
1.3228	7.01	896	0.9895	0.6406
0.9706	8.01	1008	1.1265	0.6129
1.2483	9.01	1120	1.2352	0.5484
0.9394	10.01	1232	1.2345	0.4977
0.8285	11.01	1344	0.8702	0.6682
1.1175	12.01	1456	0.9073	0.6406
1.093	13.01	1568	0.9210	0.5576
0.8364	14.01	1680	0.9316	0.6590
0.766	15.01	1792	0.7628	0.7742
0.7702	16.01	1904	0.8982	0.6682
0.9184	17.01	2016	1.1010	0.6221
0.7309	18.01	2128	0.8245	0.6866
0.9575	19.01	2240	0.9029	0.7097
0.8233	20.01	2352	1.2445	0.5161
0.7643	21.01	2464	0.9558	0.6498
0.6722	22.01	2576	1.1864	0.5714
0.8441	23.01	2688	0.9690	0.7235
0.7971	24.01	2800	0.9349	0.6774
0.8296	25.01	2912	1.4574	0.4516
0.8613	26.01	3024	0.8688	0.7189
0.5614	27.01	3136	1.2101	0.6083
0.6971	28.01	3248	1.3006	0.4654
0.9642	29.01	3360	0.9573	0.6313
0.836	30.01	3472	1.1268	0.6221
0.7166	31.01	3584	1.2384	0.5622
0.9302	32.01	3696	1.0620	0.5991
0.7729	33.01	3808	1.3253	0.5622
0.8005	34.01	3920	1.4979	0.4931
0.8025	35.01	4032	0.9786	0.5668
0.881	36.01	4144	0.8477	0.6544
0.5343	37.01	4256	1.3107	0.6544
0.5611	38.01	4368	0.9520	0.6866
0.6824	39.01	4480	0.7909	0.7281
0.6146	40.01	4592	1.0886	0.6175
1.0098	41.01	4704	1.0434	0.6313
0.5555	42.01	4816	0.9603	0.6912
0.4578	43.01	4928	1.2341	0.5945
0.5883	44.01	5040	1.2559	0.6359
0.3579	45.01	5152	1.2459	0.5622
0.7936	46.01	5264	1.2685	0.6083
0.4331	47.01	5376	0.9118	0.7097
0.8989	48.01	5488	1.3406	0.5806
0.7674	49.01	5600	1.5231	0.5484
0.8136	50.01	5712	1.2210	0.6221
0.6583	51.01	5824	0.9262	0.7051
0.4305	52.01	5936	1.0339	0.6959
0.7197	53.01	6048	1.1948	0.6682
0.7143	54.01	6160	1.1851	0.6774
0.5441	55.01	6272	1.0351	0.6636
0.6443	56.01	6384	1.0297	0.6866
0.7747	57.01	6496	1.5174	0.5991
0.5943	58.01	6608	1.1961	0.6452
0.5781	59.01	6720	1.2187	0.7143
0.6913	60.01	6832	1.1590	0.6728
0.6186	61.01	6944	1.0495	0.7235
0.5185	62.01	7056	0.9844	0.7051
0.4077	63.01	7168	1.3194	0.6313
0.8217	64.01	7280	1.2620	0.6636
0.5273	65.01	7392	1.0395	0.7373
0.9002	66.01	7504	1.5225	0.5806
0.5763	67.01	7616	1.2559	0.6406
1.0535	68.01	7728	1.2646	0.6498
1.0064	69.01	7840	1.1533	0.6866
0.332	70.01	7952	1.0438	0.7005
0.3978	71.01	8064	1.0248	0.7051
0.4459	72.01	8176	1.0926	0.7465
0.511	73.01	8288	1.1233	0.7143
0.7933	74.01	8400	1.1535	0.7189
0.3739	75.01	8512	1.3056	0.6912
0.6976	76.01	8624	1.3159	0.6682
0.5453	77.01	8736	1.4541	0.6359
0.2915	78.01	8848	1.2601	0.7051
0.6552	79.01	8960	1.5338	0.6544
0.5067	80.01	9072	1.6630	0.6037
0.5134	81.01	9184	1.4740	0.6406
0.7271	82.01	9296	1.2171	0.7097
0.719	83.01	9408	1.3653	0.6406
0.1955	84.01	9520	1.4696	0.6544
0.5761	85.01	9632	1.3334	0.6636
0.7094	86.01	9744	1.2673	0.6912
0.5186	87.01	9856	1.3147	0.6866
0.6876	88.01	9968	1.2622	0.7051
0.4912	89.01	10080	1.3054	0.7189
0.194	90.01	10192	1.3244	0.6959
0.6916	91.01	10304	1.1800	0.7327
0.5735	92.01	10416	1.1056	0.7419
0.2122	93.01	10528	1.1070	0.7281
0.1434	94.01	10640	1.1776	0.7097
0.4681	95.01	10752	1.1505	0.7327
0.2856	96.01	10864	1.1203	0.7235
0.6509	97.01	10976	1.1502	0.7189
0.1881	98.01	11088	1.1474	0.7189
0.5577	99.0	11100	1.1473	0.7189

Framework versions

Transformers 4.36.2
Pytorch 1.13.1
Datasets 2.16.1
Tokenizers 0.15.0

🔧 Technical Details

No technical details are provided in the original document.

📄 License

This model is licensed under the CC - BY - NC - 4.0 license.

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