bert-base-chinese-complaint-128

This model is a fine-tuned version of bert-base-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.3004

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 16

Training results

Training Loss	Epoch	Step	Validation Loss
3.3735	1.0	1250	2.4628
2.2412	2.0	2500	2.0378
1.9251	3.0	3750	1.8368
1.7407	4.0	5000	1.6972
1.6137	5.0	6250	1.5937
1.5365	6.0	7500	1.5315
1.4662	7.0	8750	1.4921
1.3985	8.0	10000	1.4517
1.3509	9.0	11250	1.4308
1.3047	10.0	12500	1.3906
1.2745	11.0	13750	1.3467
1.2377	12.0	15000	1.3306
1.2139	13.0	16250	1.3205
1.2027	14.0	17500	1.3098
1.1722	15.0	18750	1.2845
1.1697	16.0	20000	1.3004

Framework versions

• Transformers 4.8.2
• Pytorch 1.7.1
• Datasets 1.16.1
• Tokenizers 0.10.3