ae957c2c7de5b941ec4d602bd76d787e

This model is a fine-tuned version of google-bert/bert-base-cased on the nyu-mll/glue [qqp] dataset. It achieves the following results on the evaluation set:

• Loss: 0.3412
• Data Size: 1.0
• Epoch Runtime: 547.1133
• Accuracy: 0.8551
• F1 Macro: 0.8489
• Rouge1: 0.8552
• Rouge2: 0.0
• Rougel: 0.8552
• Rougelsum: 0.8551

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	Rouge1	Rouge2	Rougel	Rougelsum
No log	0	0	0.7233	0	17.7426	0.3680	0.2690	0.3682	0.0	0.3681	0.3683
0.5923	1	11370	0.4791	0.0078	21.9407	0.7624	0.7489	0.7624	0.0	0.7624	0.7624
0.4721	2	22740	0.4433	0.0156	25.8439	0.7907	0.7776	0.7907	0.0	0.7908	0.7906
0.4223	3	34110	0.4047	0.0312	34.1094	0.8058	0.7850	0.8059	0.0	0.8059	0.8057
0.3741	4	45480	0.4180	0.0625	50.8946	0.8201	0.7972	0.8200	0.0	0.8201	0.8201
0.3549	5	56850	0.3498	0.125	84.2947	0.8432	0.8362	0.8432	0.0	0.8432	0.8431
0.3469	6	68220	0.3156	0.25	150.4334	0.8618	0.8526	0.8619	0.0	0.8618	0.8619
0.3227	7	79590	0.3255	0.5	289.4957	0.8643	0.8574	0.8643	0.0	0.8644	0.8643
0.3462	8.0	90960	0.3285	1.0	558.3491	0.8610	0.8542	0.8610	0.0	0.8610	0.8609
0.3021	9.0	102330	0.3144	1.0	540.2528	0.8543	0.8495	0.8543	0.0	0.8544	0.8543
0.2517	10.0	113700	0.3215	1.0	546.0557	0.8715	0.8649	0.8715	0.0	0.8715	0.8714
0.2616	11.0	125070	0.3174	1.0	546.6105	0.8731	0.8654	0.8730	0.0	0.8731	0.8731
0.2343	12.0	136440	0.3443	1.0	546.2910	0.8657	0.8599	0.8658	0.0	0.8657	0.8657
0.2567	13.0	147810	0.3412	1.0	547.1133	0.8551	0.8489	0.8552	0.0	0.8552	0.8551

Framework versions

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