3830bd109d48e04d03be889c0946abf7

This model is a fine-tuned version of albert/albert-xxlarge-v1 on the contemmcm/trec dataset. It achieves the following results on the evaluation set:

• Loss: 0.2092
• Data Size: 1.0
• Epoch Runtime: 10.4821
• Accuracy: 0.9688
• F1 Macro: 0.9552

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	2.2324	0	0.8350	0.1646	0.0476
No log	1	170	2.2546	0.0078	1.0334	0.2771	0.0723
No log	2	340	1.9432	0.0156	1.1140	0.2771	0.0723
No log	3	510	1.5890	0.0312	1.4418	0.2417	0.1273
No log	4	680	1.2010	0.0625	1.8216	0.5708	0.4903
0.0865	5	850	0.8811	0.125	2.3715	0.7208	0.6229
0.0865	6	1020	0.2623	0.25	3.5571	0.9354	0.7841
0.3188	7	1190	0.2526	0.5	5.9086	0.9479	0.8551
0.1945	8.0	1360	0.1477	1.0	10.7739	0.9667	0.9517
0.1706	9.0	1530	0.1666	1.0	10.5856	0.9646	0.9466
0.1284	10.0	1700	0.1398	1.0	10.5293	0.9729	0.9674
0.0773	11.0	1870	0.1698	1.0	10.6137	0.9646	0.9596
0.0375	12.0	2040	0.2128	1.0	10.5202	0.9646	0.9483
0.0357	13.0	2210	0.2166	1.0	10.5073	0.975	0.9694
0.0336	14.0	2380	0.2092	1.0	10.4821	0.9688	0.9552

Framework versions

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