Robbert-v2-dutch-base-finetuned-emotion-valence Open Source Model - Accurately Predict Emotional Valence Scores of Dutch Texts

Robbert V2 Dutch Base Finetuned Emotion Valence

Developed by antalvdb

This model is a fine-tuned version of the RobBERT-v2 Dutch base model for emotional valence prediction tasks, designed to predict the valence score (emotional intensity from negative to positive) of Dutch text.

Text Classification

Transformers

Open Source License:MIT #Dutch Sentiment Analysis #Emotional Valence Prediction #High-Precision Regression

Downloads 61

Release Time : 7/26/2024

Model Overview

This is a fine-tuned model for Dutch sentiment analysis tasks, capable of predicting the emotional valence (intensity from negative to positive) of text. Based on the RobBERT-v2 Dutch base model, it excels in emotional valence prediction.

Model Features

Dutch Sentiment Analysis

Specialized for predicting emotional valence in Dutch text

Efficient Fine-Tuning

Fine-tuned for 50 epochs, reducing validation loss to 0.0317

Stable Performance

Demonstrated stable convergence and strong generalization during training

Model Capabilities

Dutch Text Sentiment Analysis

Emotional Valence Score Prediction

Text Emotion Recognition

Use Cases

Sentiment Analysis

Social Media Sentiment Monitoring

Analyze the emotional tone of Dutch social media posts

Provides emotional valence scores on a 0-1 scale

Customer Feedback Analysis

Evaluate the emotional valence of Dutch customer reviews

Helps quantify customer satisfaction

🚀 robbert-v2-dutch-base-finetuned-emotion-valence

This model is a fine - tuned version of pdelobelle/robbert-v2-dutch-base. It aims to achieve better performance on emotion valence tasks, and has achieved good results on the evaluation set.

📚 Documentation

Model Information

Property	Details
Model Type	Fine - tuned version of pdelobelle/robbert-v2-dutch-base
License	MIT

Model Performance

This model achieves the following results on the evaluation set:

Loss: 0.0317
Rmse: 0.1781

Training Procedure

Training Hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e-05
train_batch_size: 32
eval_batch_size: 32
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 50

Training Results

Training Loss	Epoch	Step	Validation Loss	Rmse
0.0813	1.0	25	0.0510	0.2258
0.0445	2.0	50	0.0381	0.1952
0.0409	3.0	75	0.0466	0.2158
0.0308	4.0	100	0.0351	0.1874
0.0257	5.0	125	0.0393	0.1983
0.0231	6.0	150	0.0442	0.2103
0.0203	7.0	175	0.0447	0.2115
0.0191	8.0	200	0.0372	0.1929
0.0156	9.0	225	0.0425	0.2061
0.0154	10.0	250	0.0367	0.1917
0.0138	11.0	275	0.0365	0.1910
0.0128	12.0	300	0.0432	0.2078
0.0137	13.0	325	0.0329	0.1814
0.0118	14.0	350	0.0327	0.1809
0.0118	15.0	375	0.0378	0.1945
0.0109	16.0	400	0.0360	0.1897
0.0103	17.0	425	0.0325	0.1803
0.0096	18.0	450	0.0327	0.1809
0.0091	19.0	475	0.0430	0.2072
0.0081	20.0	500	0.0345	0.1856
0.0094	21.0	525	0.0365	0.1912
0.0084	22.0	550	0.0350	0.1870
0.0075	23.0	575	0.0324	0.1800
0.0069	24.0	600	0.0330	0.1816
0.0087	25.0	625	0.0347	0.1863
0.0079	26.0	650	0.0297	0.1722
0.0071	27.0	675	0.0311	0.1763
0.0076	28.0	700	0.0322	0.1795
0.0064	29.0	725	0.0338	0.1839
0.0067	30.0	750	0.0326	0.1806
0.0061	31.0	775	0.0327	0.1808
0.0064	32.0	800	0.0339	0.1842
0.0062	33.0	825	0.0300	0.1732
0.0062	34.0	850	0.0331	0.1819
0.0055	35.0	875	0.0318	0.1782
0.0059	36.0	900	0.0323	0.1797
0.0056	37.0	925	0.0311	0.1765
0.0055	38.0	950	0.0310	0.1762
0.0053	39.0	975	0.0325	0.1802
0.0056	40.0	1000	0.0310	0.1761
0.0054	41.0	1025	0.0323	0.1799
0.0057	42.0	1050	0.0351	0.1873
0.0053	43.0	1075	0.0347	0.1861
0.0054	44.0	1100	0.0330	0.1816
0.0059	45.0	1125	0.0313	0.1769
0.0053	46.0	1150	0.0312	0.1766
0.0051	47.0	1175	0.0325	0.1804
0.0057	48.0	1200	0.0304	0.1745
0.0048	49.0	1225	0.0317	0.1782
0.005	50.0	1250	0.0317	0.1781

Framework Versions

Transformers 4.42.4
Pytorch 2.3.1+cu121
Datasets 2.20.0
Tokenizers 0.19.1

📄 License

This model is released under the MIT 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