bert_base_lda_book

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

• Loss: 2.9459
• Accuracy: 0.7612

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: 96
• eval_batch_size: 96
• 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
7.6611	0.4215	10000	7.4831	0.1626
7.54	0.8431	20000	7.3893	0.1640
7.477	1.2646	30000	7.3399	0.1663
7.4478	1.6861	40000	7.3119	0.1667
5.084	2.1077	50000	4.6037	0.4982
4.1458	2.5292	60000	3.8418	0.6153
3.9219	2.9507	70000	3.6449	0.6468
3.8057	3.3723	80000	3.5337	0.6641
3.7243	3.7938	90000	3.4573	0.6761
3.6613	4.2153	100000	3.4046	0.6845
3.6198	4.6369	110000	3.3638	0.6910
3.5769	5.0584	120000	3.3235	0.6972
3.5446	5.4799	130000	3.2929	0.7021
3.5226	5.9014	140000	3.2670	0.7062
3.493	6.3230	150000	3.2465	0.7099
3.4784	6.7445	160000	3.2292	0.7130
3.4465	7.1660	170000	3.2074	0.7165
3.433	7.5876	180000	3.1960	0.7186
3.4171	8.0091	190000	3.1785	0.7212
3.4016	8.4306	200000	3.1640	0.7237
3.3853	8.8522	210000	3.1538	0.7257
3.3776	9.2737	220000	3.1404	0.7278
3.3568	9.6952	230000	3.1291	0.7295
3.3441	10.1168	240000	3.1205	0.7315
3.341	10.5383	250000	3.1095	0.7328
3.3292	10.9598	260000	3.1005	0.7345
3.3173	11.3814	270000	3.0944	0.7355
3.3162	11.8029	280000	3.0838	0.7373
3.298	12.2244	290000	3.0764	0.7386
3.293	12.6460	300000	3.0676	0.7400
3.2808	13.0675	310000	3.0613	0.7413
3.2786	13.4890	320000	3.0542	0.7423
3.2717	13.9106	330000	3.0505	0.7432
3.266	14.3321	340000	3.0427	0.7445
3.2583	14.7536	350000	3.0360	0.7454
3.2508	15.1751	360000	3.0327	0.7461
3.2454	15.5967	370000	3.0232	0.7474
3.2386	16.0182	380000	3.0186	0.7484
3.2334	16.4397	390000	3.0147	0.7492
3.2251	16.8613	400000	3.0106	0.7499
3.2228	17.2828	410000	3.0035	0.7511
3.2162	17.7043	420000	3.0003	0.7517
3.2079	18.1259	430000	2.9971	0.7524
3.2024	18.5474	440000	2.9928	0.7530
3.2014	18.9689	450000	2.9856	0.7541
3.1962	19.3905	460000	2.9826	0.7547
3.1917	19.8120	470000	2.9786	0.7555
3.1854	20.2335	480000	2.9755	0.7561
3.18	20.6551	490000	2.9711	0.7567
3.1744	21.0766	500000	2.9676	0.7573
3.1714	21.4981	510000	2.9635	0.7580
3.1664	21.9197	520000	2.9616	0.7583
3.1674	22.3412	530000	2.9592	0.7588
3.1614	22.7627	540000	2.9579	0.7591
3.1616	23.1843	550000	2.9536	0.7597
3.1579	23.6058	560000	2.9513	0.7603
3.154	24.0273	570000	2.9501	0.7605
3.1504	24.4488	580000	2.9464	0.7610
3.1482	24.8704	590000	2.9464	0.7610

Framework versions

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