geobase/geo-task-classifier-transformer

bert_tiny_trained

This model is a fine-tuned version of nreimers/BERT-Tiny_L-2_H-128_A-2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5820

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• 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
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	11	1.3285
No log	2.0	22	1.3144
No log	3.0	33	1.2977
No log	4.0	44	1.2683
No log	5.0	55	1.2476
No log	6.0	66	1.2256
No log	7.0	77	1.1961
No log	8.0	88	1.1853
No log	9.0	99	1.1625
No log	10.0	110	1.1284
No log	11.0	121	1.1036
No log	12.0	132	1.0812
No log	13.0	143	1.0573
No log	14.0	154	1.0323
No log	15.0	165	1.0092
No log	16.0	176	0.9913
No log	17.0	187	0.9725
No log	18.0	198	0.9492
No log	19.0	209	0.9269
No log	20.0	220	0.9061
No log	21.0	231	0.8869
No log	22.0	242	0.8719
No log	23.0	253	0.8521
No log	24.0	264	0.8357
No log	25.0	275	0.8169
No log	26.0	286	0.8026
No log	27.0	297	0.7936
No log	28.0	308	0.7783
No log	29.0	319	0.7677
No log	30.0	330	0.7577
No log	31.0	341	0.7516
No log	32.0	352	0.7431
No log	33.0	363	0.7355
No log	34.0	374	0.7287
No log	35.0	385	0.7220
No log	36.0	396	0.7154
No log	37.0	407	0.7119
No log	38.0	418	0.7073
No log	39.0	429	0.7025
No log	40.0	440	0.6976
No log	41.0	451	0.6931
No log	42.0	462	0.6890
No log	43.0	473	0.6859
No log	44.0	484	0.6830
No log	45.0	495	0.6807
0.7544	46.0	506	0.6785
0.7544	47.0	517	0.6774
0.7544	48.0	528	0.6769
0.7544	49.0	539	0.6768
0.7544	50.0	550	0.6767

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0