a484885e12db9058be88b46d3ccd2cf7

This model is a fine-tuned version of FacebookAI/roberta-large on the contemmcm/trec dataset. It achieves the following results on the evaluation set:

• Loss: 1.6885
• Data Size: 1.0
• Epoch Runtime: 27.6886
• Accuracy: 0.1792
• F1 Macro: 0.0506

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	1.8835	0	1.2843	0.0187	0.0061
No log	1	170	1.7362	0.0078	1.8577	0.2792	0.0846
No log	2	340	1.7011	0.0156	2.5661	0.2562	0.1448
No log	3	510	1.6330	0.0312	3.6341	0.2792	0.0769
No log	4	680	1.6645	0.0625	4.8753	0.1854	0.0583
0.1026	5	850	1.6887	0.125	6.7177	0.2208	0.1235
0.1026	6	1020	1.7585	0.25	9.7801	0.1792	0.0506
1.5392	7	1190	1.5047	0.5	16.3234	0.275	0.1945
1.6813	8.0	1360	1.6680	1.0	28.6093	0.2771	0.0723
1.6842	9.0	1530	1.7103	1.0	28.7736	0.1333	0.0392
1.6819	10.0	1700	1.6705	1.0	28.2001	0.2771	0.0723
1.6971	11.0	1870	1.6885	1.0	27.6886	0.1792	0.0506

Framework versions

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