muril-base-cased-finetuned-ours-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: 0.8885
• Accuracy: 0.625
• Precision: 0.5563
• Recall: 0.5570
• F1: 0.5412

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
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.0904	1.98	99	1.0762	0.245	0.0817	0.3333	0.1312
1.0535	3.96	198	1.0206	0.595	0.3940	0.5312	0.4410
0.9936	5.94	297	1.0099	0.58	0.3582	0.4833	0.4115
0.9371	7.92	396	0.9645	0.615	0.3918	0.5378	0.4507
0.8931	9.9	495	0.9333	0.615	0.3969	0.5447	0.4535
0.8237	11.88	594	0.9273	0.595	0.5451	0.5171	0.4439
0.7772	13.86	693	0.8890	0.61	0.5186	0.5512	0.4661
0.7365	15.84	792	0.8834	0.625	0.5661	0.5666	0.5112
0.6987	17.82	891	0.9026	0.615	0.5532	0.5507	0.5314
0.679	19.8	990	0.9018	0.63	0.5631	0.5596	0.5501
0.6572	21.78	1089	0.8737	0.635	0.5782	0.5806	0.5614
0.644	23.76	1188	0.8885	0.625	0.5563	0.5570	0.5412

Framework versions

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