bert_tiny_lda_20_v1_book

This model is a fine-tuned version of on the gokulsrinivasagan/processed_book_corpus-ld-20 dataset. It achieves the following results on the evaluation set:

• Loss: 3.8712
• Accuracy: 0.6793

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: 0.0001
• train_batch_size: 160
• eval_batch_size: 160
• seed: 10
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 10000
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
8.6049	0.7025	10000	8.4377	0.1653
5.6593	1.4051	20000	5.2243	0.5031
5.1606	2.1076	30000	4.7778	0.5589
4.9169	2.8102	40000	4.5539	0.5885
4.7607	3.5127	50000	4.4085	0.6088
4.6405	4.2153	60000	4.3027	0.6216
4.5578	4.9178	70000	4.2156	0.6307
4.496	5.6203	80000	4.1619	0.6375
4.457	6.3229	90000	4.1256	0.6425
4.4199	7.0254	100000	4.0918	0.6468
4.3953	7.7280	110000	4.0677	0.6504
4.3703	8.4305	120000	4.0441	0.6538
4.3437	9.1331	130000	4.0295	0.6560
4.3295	9.8356	140000	4.0084	0.6594
4.3125	10.5381	150000	3.9955	0.6612
4.3048	11.2407	160000	3.9842	0.6627
4.2863	11.9432	170000	3.9727	0.6645
4.276	12.6458	180000	3.9592	0.6663
4.2651	13.3483	190000	3.9543	0.6669
4.2573	14.0509	200000	3.9438	0.6683
4.247	14.7534	210000	3.9343	0.6699
4.2387	15.4560	220000	3.9274	0.6712
4.2331	16.1585	230000	3.9226	0.6718
4.2238	16.8610	240000	3.9161	0.6727
4.2171	17.5636	250000	3.9106	0.6735
4.2098	18.2661	260000	3.9046	0.6740
4.2083	18.9687	270000	3.9001	0.6749
4.1991	19.6712	280000	3.8949	0.6759
4.1961	20.3738	290000	3.8903	0.6766
4.1893	21.0763	300000	3.8864	0.6772
4.1866	21.7788	310000	3.8808	0.6779
4.181	22.4814	320000	3.8782	0.6784
4.18	23.1839	330000	3.8763	0.6785
4.1771	23.8865	340000	3.8731	0.6790
4.1765	24.5890	350000	3.8704	0.6795

Framework versions

• Transformers 4.46.3
• Pytorch 2.2.1+cu118
• Datasets 2.17.0
• Tokenizers 0.20.3