muril-base-cased-finetuned-combined-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: 1.5291
• Accuracy: 0.6657
• Precision: 0.6355
• Recall: 0.6275
• F1: 0.6294

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
0.9961	2.0	711	0.9148	0.5625	0.5495	0.5636	0.5265
0.8211	3.99	1422	0.8542	0.6096	0.6023	0.6071	0.5928
0.6667	5.99	2133	0.8459	0.6601	0.6366	0.6379	0.6361
0.5272	7.99	2844	0.9667	0.6517	0.6190	0.6223	0.6201
0.4327	9.99	3555	1.0185	0.6503	0.6351	0.6222	0.6229
0.3608	11.98	4266	1.1409	0.6313	0.6053	0.6100	0.6049
0.3038	13.98	4977	1.2336	0.6601	0.6287	0.6269	0.6273
0.2631	15.98	5688	1.3151	0.6503	0.6199	0.6167	0.6177
0.2368	17.97	6399	1.4230	0.6594	0.6315	0.6233	0.6251
0.2093	19.97	7110	1.4881	0.6629	0.6332	0.6220	0.6239
0.1968	21.97	7821	1.5003	0.6559	0.6279	0.6230	0.6242
0.1824	23.97	8532	1.5291	0.6657	0.6355	0.6275	0.6294

Framework versions

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