d4c77fdcca51d32bc3ed461f53d08c9d

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

• Loss: 0.6577
• Data Size: 1.0
• Epoch Runtime: 546.3732
• Accuracy: 0.6320
• F1 Macro: 0.3872
• Rouge1: 0.6318
• Rouge2: 0.0
• Rougel: 0.6319
• Rougelsum: 0.6317

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	0.6879	0	17.4783	0.5375	0.4847	0.5375	0.0	0.5375	0.5374
0.6438	1	11370	0.5936	0.0078	23.1011	0.6320	0.3872	0.6318	0.0	0.6319	0.6317
0.5588	2	22740	0.5311	0.0156	25.6909	0.7140	0.6788	0.7140	0.0	0.7141	0.7139
0.5315	3	34110	0.5885	0.0312	33.8598	0.7105	0.6300	0.7104	0.0	0.7105	0.7102
0.4884	4	45480	0.4737	0.0625	50.0975	0.7602	0.7423	0.7602	0.0	0.7601	0.7601
0.6657	5	56850	0.6618	0.125	82.4035	0.6320	0.3872	0.6318	0.0	0.6319	0.6317
0.6617	6	68220	0.6578	0.25	148.3036	0.6320	0.3872	0.6318	0.0	0.6319	0.6317
0.6635	7	79590	0.6587	0.5	277.7724	0.6320	0.3872	0.6318	0.0	0.6319	0.6317
0.6563	8.0	90960	0.6577	1.0	546.3732	0.6320	0.3872	0.6318	0.0	0.6319	0.6317

Framework versions

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