bert_tiny_lda_100_v1_book

This model is a fine-tuned version of on the gokulsrinivasagan/processed_book_corpus-ld-100 dataset. It achieves the following results on the evaluation set:

• Loss: 4.9524
• Accuracy: 0.6766

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 OptimizerNames.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
10.2267	0.7025	10000	10.0665	0.1636
7.2173	1.4051	20000	6.7451	0.4823
6.4475	2.1076	30000	6.0308	0.5509
6.1492	2.8102	40000	5.7522	0.5804
5.9566	3.5127	50000	5.5702	0.6009
5.8259	4.2153	60000	5.4546	0.6137
5.7396	4.9178	70000	5.3672	0.6239
5.6664	5.6203	80000	5.3074	0.6312
5.6155	6.3229	90000	5.2622	0.6366
5.5704	7.0254	100000	5.2177	0.6416
5.5381	7.7280	110000	5.1869	0.6460
5.5072	8.4305	120000	5.1572	0.6495
5.476	9.1331	130000	5.1399	0.6520
5.4586	9.8356	140000	5.1144	0.6554
5.4395	10.5381	150000	5.0980	0.6573
5.4279	11.2407	160000	5.0854	0.6588
5.4084	11.9432	170000	5.0694	0.6610
5.3943	12.6458	180000	5.0544	0.6627
5.3829	13.3483	190000	5.0477	0.6636
5.374	14.0509	200000	5.0361	0.6652
5.3602	14.7534	210000	5.0257	0.6666
5.3506	15.4560	220000	5.0155	0.6681
5.3443	16.1585	230000	5.0103	0.6687
5.3334	16.8610	240000	5.0030	0.6697
5.3252	17.5636	250000	4.9964	0.6705
5.3187	18.2661	260000	4.9904	0.6711
5.3167	18.9687	270000	4.9849	0.6723
5.3068	19.6712	280000	4.9791	0.6731
5.3031	20.3738	290000	4.9740	0.6736
5.2947	21.0763	300000	4.9701	0.6742
5.2931	21.7788	310000	4.9633	0.6752
5.2875	22.4814	320000	4.9602	0.6756
5.2841	23.1839	330000	4.9582	0.6758
5.2815	23.8865	340000	4.9541	0.6762
5.2811	24.5890	350000	4.9512	0.6766

Framework versions

• Transformers 4.46.1
• Pytorch 2.2.0+cu121
• Datasets 3.1.0
• Tokenizers 0.20.1