Modelpublic

tadiecool29/MTL-FullFineTune-afri-mt5-base-joint

sourceHugging Faceafl-3.0updated 3d agoView on Hugging Face

0likes30downloads

Model Card

MTL-FullFineTune-afri-mt5-base-joint

This model is a fine-tuned version of masakhane/afri-mt5-base on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 1.8665
—Exact Match: 0.5536
—Sentiment Accuracy: 0.7082
—Sentiment Macro F1: 0.6990
—Stance Accuracy: 0.7045
—Stance Macro F1: 0.7109
—Avg Macro F1: 0.7050

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: 0.0003
—trainbatchsize: 8
—evalbatchsize: 16
—seed: 42
—gradientaccumulationsteps: 4
—totaltrainbatch_size: 32
—optimizer: Use OptimizerNames.ADAMWTORCHFUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
—lrschedulertype: cosine
—lrschedulerwarmup_steps: 300
—num_epochs: 10
—labelsmoothingfactor: 0.1

Training results

Training Loss	Epoch	Step	Validation Loss	Exact Match	Sentiment Accuracy	Sentiment Macro F1	Stance Accuracy	Stance Macro F1	Avg Macro F1
4.4331	1.0	189	2.5095	0.1110	0.3130	0.1214	0.2182	0.0750	0.0982
2.2132	2.0	378	1.9743	0.4140	0.5673	0.5116	0.5362	0.5213	0.5165
1.9958	3.0	567	1.8926	0.5374	0.6708	0.6596	0.6983	0.7026	0.6811
1.9245	4.0	756	1.8677	0.5449	0.6908	0.6855	0.6983	0.7036	0.6945
1.9002	5.0	945	1.8702	0.5499	0.6983	0.6880	0.6870	0.6869	0.6874
1.8719	6.0	1134	1.8703	0.5474	0.7057	0.6939	0.6870	0.6918	0.6928
1.8554	7.0	1323	1.8652	0.5499	0.7082	0.6998	0.7020	0.7096	0.7047
1.8604	8.0	1512	1.8656	0.5524	0.7070	0.6982	0.7045	0.7113	0.7047
1.8424	9.0	1701	1.8664	0.5561	0.7095	0.7001	0.7070	0.7131	0.7066
1.8492	10.0	1890	1.8665	0.5536	0.7082	0.6990	0.7045	0.7109	0.7050

Framework versions

—Transformers 5.16.1
—Pytorch 2.11.0+cu128
—Datasets 4.8.5
—Tokenizers 0.23.1