bert_tiny_olda_book_5_v1

This model is a fine-tuned version of on the gokulsrinivasagan/processed_book_corpus-ld-5 dataset. It achieves the following results on the evaluation set:

• Loss: 1.4052
• Accuracy: 0.0003

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
1.4427	0.7025	10000	1.4406	0.0001
1.4355	1.4051	20000	1.4316	0.0000
1.433	2.1076	30000	1.4302	0.0000
1.4306	2.8102	40000	1.4287	0.0000
1.431	3.5127	50000	1.4299	0.0000
1.4295	4.2153	60000	1.4261	0.0000
1.4266	4.9178	70000	1.4238	0.0000
1.4205	5.6203	80000	1.4168	0.0000
1.4191	6.3229	90000	1.4146	0.0001
1.419	7.0254	100000	1.4147	0.0001
1.4181	7.7280	110000	1.4134	0.0001
1.4163	8.4305	120000	1.4129	0.0001
1.4164	9.1331	130000	1.4111	0.0002
1.4147	9.8356	140000	1.4100	0.0002
1.414	10.5381	150000	1.4106	0.0002
1.4141	11.2407	160000	1.4098	0.0002
1.413	11.9432	170000	1.4093	0.0003
1.4124	12.6458	180000	1.4087	0.0003
1.4114	13.3483	190000	1.4078	0.0002
1.4115	14.0509	200000	1.4078	0.0002
1.4104	14.7534	210000	1.4080	0.0003
1.411	15.4560	220000	1.4075	0.0003
1.411	16.1585	230000	1.4069	0.0003
1.4104	16.8610	240000	1.4066	0.0003
1.4102	17.5636	250000	1.4068	0.0003
1.4094	18.2661	260000	1.4064	0.0003
1.4096	18.9687	270000	1.4059	0.0003
1.4098	19.6712	280000	1.4059	0.0003
1.409	20.3738	290000	1.4058	0.0002
1.4081	21.0763	300000	1.4057	0.0003
1.4082	21.7788	310000	1.4052	0.0003
1.4084	22.4814	320000	1.4054	0.0003
1.4088	23.1839	330000	1.4052	0.0002
1.4088	23.8865	340000	1.4050	0.0003
1.4084	24.5890	350000	1.4052	0.0003

Framework versions

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