patentClassificationLongFormerTextrank

This model is a fine-tuned version of eclec/patentClassificationLongFormer2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4294
• Accuracy: 0.7959
• F1: 0.6187

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: 1.330504416591152e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 3
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.24934655263987432
• lr_scheduler_warmup_steps: 90
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.4444	1.0	2059	0.4397	0.7947	0.6100
0.3942	2.0	4119	0.4294	0.7959	0.6187
0.3331	3.0	6177	0.4607	0.7999	0.6078

Framework versions

• Transformers 4.32.1
• Pytorch 2.0.1+cu117
• Datasets 2.12.0
• Tokenizers 0.13.2