muril-base-cased-finetuned-non-code-mixed-DS

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

• Loss: 1.2867
• Accuracy: 0.6214
• Precision: 0.6081
• Recall: 0.6009
• F1: 0.6034

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: 1e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 43
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.0861	2.0	463	1.0531	0.3506	0.1169	0.3333	0.1731
0.99	3.99	926	0.9271	0.5836	0.4310	0.5200	0.4502
0.8759	5.99	1389	0.9142	0.5965	0.5788	0.5907	0.5802
0.7726	7.98	1852	0.8726	0.6095	0.6079	0.6078	0.6027
0.6659	9.98	2315	0.9145	0.6246	0.6139	0.6174	0.6140
0.5727	11.97	2778	0.9606	0.6311	0.6180	0.6109	0.6133
0.4889	13.97	3241	1.0342	0.6170	0.6059	0.6054	0.6045
0.4267	15.97	3704	1.0539	0.6170	0.6089	0.6081	0.6066
0.3751	17.96	4167	1.1740	0.6343	0.6255	0.6074	0.6112
0.3402	19.96	4630	1.2021	0.6192	0.6078	0.6013	0.6031
0.318	21.95	5093	1.2875	0.6181	0.6007	0.5946	0.5965
0.2977	23.95	5556	1.2867	0.6214	0.6081	0.6009	0.6034

Framework versions

• Transformers 4.20.1
• Pytorch 1.10.1+cu111
• Datasets 2.3.2
• Tokenizers 0.12.1