Albert-Base

This model is a fine-tuned version of albert/albert-base-v2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2950
• Accuracy: 0.9234

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: 128
• eval_batch_size: 512
• seed: 42
• 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
• num_epochs: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.4359	1.0	158	0.3263	0.9016
0.2711	2.0	316	0.3180	0.9063
0.1583	3.0	474	0.3126	0.9117
0.1463	4.0	632	0.3113	0.9132
0.0722	5.0	790	0.2950	0.9234
0.0627	6.0	948	0.3461	0.9178
0.0528	7.0	1106	0.3541	0.9177
0.0336	8.0	1264	0.3469	0.9251
0.0416	9.0	1422	0.4367	0.9130
0.0356	10.0	1580	0.3814	0.9226
0.0092	11.0	1738	0.4591	0.9206
0.0127	12.0	1896	0.4355	0.9277
0.0055	13.0	2054	0.4419	0.9264
0.0019	14.0	2212	0.4578	0.9297
0.0009	15.0	2370	0.4947	0.9284
0.0017	16.0	2528	0.4893	0.9318
0.0003	17.0	2686	0.5132	0.9278
0.0013	18.0	2844	0.4901	0.9307
0.0029	19.0	3002	0.5176	0.9301
0.0064	20.0	3160	0.5325	0.9321
0.0001	21.0	3318	0.5367	0.9294
0.0001	22.0	3476	0.5389	0.9309
0.008	23.0	3634	0.5888	0.9274
0.0001	24.0	3792	0.5542	0.9315
0.0003	25.0	3950	0.5350	0.9323
0.0	26.0	4108	0.5480	0.9312
0.0	27.0	4266	0.5337	0.9343
0.0	28.0	4424	0.5453	0.9331
0.0	29.0	4582	0.5434	0.9333
0.0	30.0	4740	0.5443	0.9335
0.0	31.0	4898	0.5457	0.9338
0.0	32.0	5056	0.5468	0.9339
0.0	33.0	5214	0.5481	0.9339
0.0	34.0	5372	0.5492	0.9339
0.0	35.0	5530	0.5503	0.9343
0.0	36.0	5688	0.5511	0.9345
0.0	37.0	5846	0.5519	0.9343
0.0	38.0	6004	0.5525	0.9343
0.0	39.0	6162	0.5529	0.9343
0.0	40.0	6320	0.5530	0.9343

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Tokenizers 0.21.0