PhoBert_content_256 Open-Source Vietnamese Text Classification Model

Phobert Content 256

Developed by RonTon05

A Vietnamese text classification model fine-tuned from vinai/phobert-base-v2, achieving 89.62% accuracy on the validation set

Large Language Model

Transformers

#Vietnamese text classification #PhoBERT fine-tuning #High accuracy

Downloads 64

Release Time : 4/9/2025

Model Overview

This model is an optimized BERT variant for Vietnamese text content classification tasks, suitable for short text classification scenarios

Model Features

Vietnamese optimization

Based on the PhoBERT architecture specifically designed for Vietnamese, it performs excellently in Vietnamese text processing

Efficient fine-tuning

Fine-tuned for 10 epochs with a learning rate of 2e-5, achieving nearly 90% accuracy on the validation set

Lightweight deployment

Moderate parameter size of the base model makes it suitable for real-world production deployment

Model Capabilities

Vietnamese text classification

Short text content analysis

Multi-category prediction

Use Cases

Content moderation

User-generated content classification

Classifying Vietnamese user comments on social media

89.62% accuracy, 88.69% F1 score

Customer service

Ticket auto-classification

Automatically classifying Vietnamese customer service requests

🚀 PhoBert_content_256

This model, PhoBert_content_256, is a fine - tuned version of vinai/phobert-base-v2. It offers enhanced performance on specific tasks, leveraging the pre - trained knowledge of the base model.

🚀 Quick Start

This model is a fine - tuned version of vinai/phobert-base-v2 on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.5561
Accuracy: 0.8962
F1: 0.8869

📚 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: 2e - 05
train_batch_size: 64
eval_batch_size: 64
seed: 42
optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon = 1e - 08 and optimizer_args = No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	0.4673	200	0.3165	0.8774	0.8648
No log	0.9346	400	0.2647	0.8947	0.8875
0.3125	1.4019	600	0.2729	0.8947	0.8861
0.3125	1.8692	800	0.2917	0.8981	0.8885
0.2303	2.3364	1000	0.2730	0.8976	0.8893
0.2303	2.8037	1200	0.2565	0.9026	0.8958
0.1856	3.2710	1400	0.3230	0.8929	0.8831
0.1856	3.7383	1600	0.3062	0.8988	0.8901
0.143	4.2056	1800	0.3220	0.8997	0.8926
0.143	4.6729	2000	0.3811	0.8926	0.8819
0.1124	5.1402	2200	0.4578	0.8900	0.8785
0.1124	5.6075	2400	0.4192	0.8935	0.8835
0.0889	6.0748	2600	0.4115	0.8986	0.8910
0.0889	6.5421	2800	0.4469	0.8959	0.8883
0.072	7.0093	3000	0.4771	0.8912	0.8804
0.072	7.4766	3200	0.4829	0.8947	0.8855
0.072	7.9439	3400	0.4922	0.8943	0.8850
0.0559	8.4112	3600	0.5391	0.8941	0.8840
0.0559	8.8785	3800	0.5356	0.8941	0.8848
0.0451	9.3458	4000	0.5583	0.8943	0.8848
0.0451	9.8131	4200	0.5561	0.8962	0.8869

Framework versions

Transformers 4.51.1
Pytorch 2.5.1+cu121
Datasets 3.3.1
Tokenizers 0.21.0

📄 License

The model is released under the AGPL - 3.0 license.

📦 Metadata

Property	Details
Library Name	transformers
Model Type	PhoBert_content_256
Base Model	vinai/phobert-base-v2
Tags	generated_from_trainer
Metrics	accuracy, f1

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