deberta-v2-base-japanese-finetuned-kyoto

This model is a fine-tuned version of ku-nlp/deberta-v2-base-japanese on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6094
• Precision: 0.7567
• Recall: 0.7654
• F1: 0.7574

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: 120
• eval_batch_size: 120
• seed: 42
• 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	Precision	Recall	F1
No log	1.0	21	0.7848	0.7113	0.5890	0.4384
No log	2.0	42	0.6595	0.6998	0.7330	0.7159
No log	3.0	63	0.6561	0.7189	0.7136	0.7007
No log	4.0	84	0.6023	0.7166	0.7508	0.7333
No log	5.0	105	0.6213	0.7325	0.7427	0.7283
No log	6.0	126	0.6261	0.7652	0.7460	0.7349
No log	7.0	147	0.5984	0.7602	0.7540	0.7481
No log	8.0	168	0.6204	0.7642	0.7573	0.7507
No log	9.0	189	0.6094	0.7567	0.7654	0.7574
No log	10.0	210	0.6167	0.7609	0.7605	0.7554

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu117
• Datasets 2.10.1
• Tokenizers 0.13.2