This model is a fine-tuned version of MCG-NJU/videomae-base on an unknown dataset. It achieves the following results on the evaluation set: The following hyperparameters were used during training: - learningrate: 5e-05 - trainbatchsize: 8 - evalbatchsize: 8 - lrschedulertype: linear - trainingsteps: 370 - Transformers 5.16.1 - Pytorch 2.14.0+cu126 - Datasets 5.0.1 - Tokenizers 0.23.2
Open-weight model · Video classification
vi-sign-language-videomae-base
by Star Duong star092304/vi-sign-language-videomae-base
vi-sign-language-videomae-base is an open-weight model for video classification from Star Duong, released under MIT License. It has 86M parameters. At 16-bit it needs about 0.2 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index. It draws 2.9k downloads a month.
This repository houses a fine-tuned VideoMAE (Base) model optimized for multi-class Vietnamese Sign Language Recognition (VSLR).
Parameters86M
Context—
Weights1.0 GB
Licensemit
AccessOpen weights
Monthly Downloads2.9k
Runs On
What it takes to serve vi-sign-language-videomae-base (86M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
|---|---|---|---|---|---|
| 16-bit | 0.2 GB | 0.2 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 0.1 GB | 0.1 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 0.0 GB | 0.1 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Oct 16, 2026.
vi-sign-language-videomae-base on every accelerator the SAVRN Index prices, at every precision
Model Card
By Star Duong, published under mit, revision 9759fc78b5d6.
Vietnamese Sign Language Recognition (VSLR) Model
This repository houses a fine-tuned VideoMAE (Base) model optimized for multi-class Vietnamese Sign Language Recognition (VSLR). The model architecture adapts self-supervised video representations to accurately classify short video clips of sign gestures into distinct Vietnamese text labels.
Model Description
- Base Architecture: MCG-NJU/videomae-base-finetuned-kinetics (Video Masked Autoencoders)
- Dataset utilized: star092304/ViSignLanguage-Video
- Task: Multi-class Video Classification (Spatiotemporal Feature Extraction)
- Target Language: Vietnamese Sign Language (VNSL)
The model processes short video sequences by partitioning them into spatiotemporal patches, mapping sequential gestures (such as "Ăn", "Bệnh viện", "Xin lỗi") to their corresponding semantic classes.
Training & Evaluation Visualizations
The training routine was monitored closely across key evaluation metrics to prevent overfitting while maximizing classification accuracy on the validation split.
1. Training and Validation Metrics
The plot below illustrates the progression of accuracy, precision, recall, and F1-score across successive training epochs.
2. Loss Curves
Configuration
- Architecture
- VideoMAEForVideoClassification
- Layers
- 12
- Hidden size
- 768
- Feed-forward size
- 3,072
- Attention heads
- 12
- Model type
- videomae
Identity and Version
- Repository
- star092304/vi-sign-language-videomae-base
- Publisher
- Star Duong
- Task
- Video classification
- Modality
- Video
- Library
- transformers
- Parameters
- 86M parameters
- Languages
- vi
- Revision
- 9759fc78b5d69625317bb658b4104040f9427fa3
- First published
- 2026-05-31
- Last updated
- 2026-05-31
Files and Weights
14 files, 1.0 GB in total. The weights are 3 files totalling 1.0 GB in pth, safetensors.
Weights3 files · 1.0 GB
Configuration2 files · 6.3 KB
Documentation1 file · 4.9 KB
Other7 files · 282.1 KB
Repository1 file · 1.5 KB
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| classifier_sequential.pth | Weights | 345.3 MB | 139dcb459071 |
| model.safetensors | Weights | 345.2 MB | c845a9922dc1 |
| videomae_best_model.pth | Weights | 345.3 MB | 050f2154d641 |
| config.json | Configuration | 5.9 KB | — |
| preprocessor_config.json | Configuration | 415 B | — |
| README.md | Documentation | 4.9 KB | — |
| inference/inference_for_colab.ipynb | Other | 30.0 KB | — |
| label_mapping.pkl | Other | 1.3 KB | d57950c25dcf |
| src/pipeline_VideoMAE.ipynb | Other | 82.0 KB | — |
| test/cv_submission.csv | Other | 60.4 KB | — |
| test/public_test.csv | Other | 34.4 KB | — |
| training_plots/loss.png | Other | 32.9 KB | — |
| training_plots/metrics.png | Other | 41.1 KB | — |
| .gitattributes | Repository | 1.5 KB | — |
License and Download
- License
- mit
- Access
- Open weights, no gate
- Download size
- 1.0 GB
Download from Star Duong
Released by Star Duong through its official repository on Hugging Face. Read the license.
Built From
- Derived from MCG-NJU/videomae-base-finetuned-kinetics
- Trained on (disclosed) star092304/ViSignLanguage-Video
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 1.0 GB |
| 16-bit | 0.2 GB |
| 8-bit | 0.1 GB |
| 4-bit | 0.0 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About vi-sign-language-videomae-base
How much GPU memory does vi-sign-language-videomae-base need?
About 0.2 GB at 16-bit and 0.1 GB at 4-bit: the weights (86M parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run vi-sign-language-videomae-base on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Can I use vi-sign-language-videomae-base commercially?
Yes. vi-sign-language-videomae-base is released under MIT License. The MIT License is a short permissive license. It permits commercial use, modification and redistribution, provided the copyright notice and permission notice are included.
Similar Models
This model is a fine-tuned version of MCG-NJU/videomae-base on an unknown dataset. It achieves the following results on the evaluation set: The following hyperparameters were used during training: - learningrate: 5e-05 - trainbatchsize: 8 - evalbatchsize: 8 - lrschedulertype: linear - trainingsteps: 370 - Transformers 5.16.1 - Pytorch 2.11.0+cu128 - Datasets 5.0.1 - Tokenizers 0.23.1
This model is a fine-tuned version of MCG-NJU/videomae-base for binary violence classification (violent / non-violent). It builds on Nikeytas/videomae-crime-detector-production-v1, which was itself fine-tuned from videomae-base on a subset of UCF Crime. Starting from that checkpoint, this model was further fine-tuned on the Bus Violence Dataset to close the domain gap to public-transport surveillance footage. - UCF Crime (jinmang2/ucfcrime) — inherited from the base checkpoint - Bus Violence Dataset (Zenodo) — real moving-bus footage, binary violent / non-violent labels, used for domain-specific fine-tuning Evaluated on a held-out Bus Violence Dataset test split (n = 280). The base…
Model · Video classification
videomae-base-finetuned-kinetics-violence-nonviolence-tuned
This model is a fine-tuned version of MCG-NJU/videomae-base-finetuned-kinetics on an unknown dataset. It achieves the following results on the evaluation set: The following hyperparameters were used during training: - learningrate: 5e-05 - trainbatchsize: 16 - evalbatchsize: 16 - lrschedulertype: linear - lrschedulerwarmupratio: 0.1 - trainingsteps: 348 - Transformers 4.49.0 - Pytorch 2.6.0+cu126 - Datasets 3.3.2 - Tokenizers 0.21.0
VideoMAEv2-Base model pre-trained for 800 epochs in a self-supervised way on UnlabeldHybrid-1M dataset. It was introduced in the paper [[CVPR23]VideoMAE V2: Scaling Video Masked Autoencoders with Dual Masking](https://arxiv.org/abs/2203.12602) by Wang et al. and first released in GitHub. You can use the raw model for video feature extraction. Here is how to use this model to extract a video feature
Model · Video classification
videomae-base-finetuned-kinetics
VideoMAE model pre-trained for 1600 epochs in a self-supervised way and fine-tuned in a supervised way on Kinetics-400. It was introduced in the paper VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training by Tong et al. and first released in this repository. Disclaimer: The team releasing VideoMAE did not write a model card for this model so this model card has been written by the Hugging Face team. VideoMAE is an extension of Masked Autoencoders (MAE) to video. The architecture of the model is very similar to that of a standard Vision Transformer (ViT), with a decoder on top for predicting pixel values for masked patches. Videos are presented to…