e81f098cdf3e9c41569c7fc86c0a3d4d

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

• Loss: 0.4065
• Data Size: 1.0
• Epoch Runtime: 165.6836
• Accuracy: 0.8579
• F1 Macro: 0.8579
• Rouge1: 0.8577
• Rouge2: 0.0
• Rougel: 0.8579
• Rougelsum: 0.8579

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	0.6942	0	2.8078	0.5061	0.3740	0.5057	0.0	0.5061	0.5059
No log	1	3273	0.5923	0.0078	4.3789	0.6858	0.6662	0.6858	0.0	0.6860	0.6858
0.0096	2	6546	0.4991	0.0156	5.6727	0.7748	0.7700	0.7748	0.0	0.7746	0.7744
0.5	3	9819	0.4260	0.0312	8.4860	0.8158	0.8148	0.8156	0.0	0.8158	0.8156
0.4763	4	13092	0.4190	0.0625	13.3583	0.8094	0.8081	0.8094	0.0	0.8097	0.8096
0.3978	5	16365	0.3694	0.125	23.0329	0.8421	0.8420	0.8421	0.0	0.8421	0.8420
0.4019	6	19638	0.3511	0.25	43.2164	0.8588	0.8587	0.8588	0.0	0.8590	0.8585
0.3488	7	22911	0.3544	0.5	82.8278	0.8623	0.8621	0.8625	0.0	0.8625	0.8623
0.363	8.0	26184	0.3547	1.0	163.5814	0.8496	0.8491	0.8496	0.0	0.8498	0.85
0.2911	9.0	29457	0.3336	1.0	166.4596	0.8605	0.8605	0.8603	0.0	0.8607	0.8605
0.2801	10.0	32730	0.3371	1.0	165.2680	0.8640	0.8639	0.8641	0.0	0.8638	0.8642
0.2472	11.0	36003	0.4494	1.0	166.4116	0.8583	0.8582	0.8586	0.0	0.8583	0.8583
0.259	12.0	39276	0.4378	1.0	166.0218	0.8373	0.8369	0.8373	0.0	0.8373	0.8373
0.1931	13.0	42549	0.4065	1.0	165.6836	0.8579	0.8579	0.8577	0.0	0.8579	0.8579

Framework versions

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