c298e056e156114f1678fe4e60096c0a

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

• Loss: 0.3011
• Data Size: 1.0
• Epoch Runtime: 10.5298
• Accuracy: 0.9375
• F1 Macro: 0.9255

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	1.9176	0	0.8533	0.2021	0.0884
No log	1	170	1.6871	0.0078	1.2174	0.325	0.1475
No log	2	340	1.6486	0.0156	1.0379	0.3771	0.1709
No log	3	510	1.4030	0.0312	1.2834	0.4583	0.2928
No log	4	680	1.0860	0.0625	1.6574	0.6021	0.4622
0.0738	5	850	0.5758	0.125	2.2227	0.8292	0.6888
0.0738	6	1020	0.6142	0.25	3.3356	0.7729	0.7175
0.3708	7	1190	0.3157	0.5	5.6539	0.9229	0.8567
0.3065	8.0	1360	0.2917	1.0	10.8257	0.9333	0.9148
0.2271	9.0	1530	0.2740	1.0	10.7049	0.9313	0.9336
0.2297	10.0	1700	0.2193	1.0	10.7394	0.9542	0.9535
0.1829	11.0	1870	0.2914	1.0	10.7309	0.9542	0.9320
0.1473	12.0	2040	0.3035	1.0	10.6042	0.9375	0.9005
0.0766	13.0	2210	0.6906	1.0	10.5800	0.8979	0.9117
0.092	14.0	2380	0.3011	1.0	10.5298	0.9375	0.9255

Framework versions

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