b82bf19d1990c8487a3af5d296e44474

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

• Loss: 0.2969
• Data Size: 1.0
• Epoch Runtime: 7.0773
• Accuracy: 0.9396
• F1 Macro: 0.9156

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
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Accuracy	F1 Macro
No log	0	0	1.8181	0	0.7914	0.2917	0.1088
No log	1	170	1.6968	0.0078	1.0215	0.2771	0.0723
No log	2	340	1.6595	0.0156	0.9251	0.3229	0.2235
No log	3	510	1.2829	0.0312	1.0391	0.5083	0.3556
No log	4	680	0.9331	0.0625	1.2147	0.7146	0.6218
0.0652	5	850	0.4361	0.125	1.6323	0.9021	0.7538
0.0652	6	1020	0.5383	0.25	2.4263	0.8313	0.7076
0.3391	7	1190	0.3894	0.5	3.8828	0.9375	0.9159
0.2746	8.0	1360	0.2471	1.0	7.0792	0.9437	0.9326
0.2831	9.0	1530	0.2849	1.0	6.9488	0.9271	0.9171
0.1722	10.0	1700	0.4727	1.0	7.0109	0.8896	0.8683
0.1523	11.0	1870	0.2840	1.0	6.9373	0.9375	0.9222
0.0999	12.0	2040	0.2969	1.0	7.0773	0.9396	0.9156

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.1