5ca30c8541099a60bde5251317e53b0f

This model is a fine-tuned version of google-bert/bert-large-cased-whole-word-masking on the contemmcm/trec dataset. It achieves the following results on the evaluation set:

• Loss: 0.1939
• Data Size: 1.0
• Epoch Runtime: 18.7653
• Accuracy: 0.9729
• F1 Macro: 0.9761

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	1.9527	0	0.9522	0.1229	0.0761
No log	1	170	1.7989	0.0078	1.3222	0.1875	0.1364
No log	2	340	1.5707	0.0156	1.6749	0.3729	0.2454
No log	3	510	1.5574	0.0312	2.4780	0.3771	0.2223
No log	4	680	0.8417	0.0625	3.3871	0.8146	0.6680
0.069	5	850	0.2423	0.125	4.7374	0.9292	0.7785
0.069	6	1020	0.1945	0.25	7.1021	0.9604	0.9373
0.2407	7	1190	0.1281	0.5	11.7022	0.9646	0.9507
0.1758	8.0	1360	0.1763	1.0	19.3508	0.9667	0.9721
0.1439	9.0	1530	0.1483	1.0	18.6932	0.9708	0.9729
0.0828	10.0	1700	0.3041	1.0	19.7847	0.9354	0.8775
0.0861	11.0	1870	0.1939	1.0	18.7653	0.9729	0.9761

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1