Modelpublic

callmesan/twitter-roberta-large-hate-latest-roman-urdu-fine-grained

sourceHugging Facemitupdated 2y agoView on Hugging Face

0likes7downloads

Model Card

twitter-roberta-large-hate-latest-roman-urdu-fine-grained

This model is a fine-tuned version of cardiffnlp/twitter-roberta-large-hate-latest on the None dataset. It achieves the following results on the evaluation set:

—Loss: 0.8408
—Accuracy: 0.8023
—Precision: 0.7245
—Recall: 0.7129
—F1: 0.7177

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
—trainbatchsize: 32
—evalbatchsize: 128
—seed: 42
—gradientaccumulationsteps: 2
—totaltrainbatch_size: 64
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.9091	1.0	113	0.8078	0.7034	0.3788	0.4083	0.3878
1.1008	2.0	226	0.9036	0.6510	0.2624	0.3397	0.2841
0.7624	3.0	339	0.6444	0.7844	0.7015	0.6274	0.6544
0.5512	4.0	452	0.5075	0.8342	0.7675	0.7453	0.7463
0.5258	5.0	565	0.3519	0.8858	0.8263	0.8081	0.8163
0.3489	6.0	678	0.3154	0.9011	0.8612	0.8260	0.8399
0.3182	7.0	791	0.2394	0.9295	0.8985	0.8864	0.8895
0.2263	8.0	904	0.1722	0.9502	0.9092	0.9223	0.9143
0.2024	9.0	1017	0.1252	0.9684	0.9474	0.9441	0.9457
0.1757	10.0	1130	0.1101	0.9736	0.9592	0.9506	0.9548

Framework versions

—Transformers 4.45.1
—Pytorch 2.4.0
—Datasets 3.0.1
—Tokenizers 0.20.0