bert-base-japanese-ghost_rate-weighted-0605

This model is a fine-tuned version of cl-tohoku/bert-base-japanese on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.4588
• Accuracy: 0.4300
• F1: 0.3863

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	0.9980	253	1.5043	0.2988	0.2334
1.5403	2.0	507	1.4588	0.4300	0.3863
1.5403	2.9980	760	1.4944	0.3984	0.3676
1.1252	4.0	1014	1.5833	0.4103	0.3885
1.1252	4.9980	1267	1.6723	0.4024	0.3889
0.7911	6.0	1521	1.8026	0.4083	0.3978
0.7911	6.9980	1774	1.9235	0.3984	0.3910
0.5514	8.0	2028	2.0213	0.3955	0.4014
0.5514	8.9980	2281	2.0726	0.4034	0.4096
0.4123	9.9803	2530	2.1018	0.4034	0.4074

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1