8c7369ec0ea2ac7e6ebb149b4a095a7f

This model is a fine-tuned version of google-bert/bert-base-cased on the ccdv/patent-classification [abstract] dataset. It achieves the following results on the evaluation set:

• Loss: 1.4563
• Data Size: 1.0
• Epoch Runtime: 40.2647
• Accuracy: 0.6286
• F1 Macro: 0.5885

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
No log	0	0	2.2724	0	2.8149	0.1056	0.0334
No log	1	781	1.9525	0.0078	3.4525	0.3017	0.1438
No log	2	1562	1.6928	0.0156	3.4530	0.3696	0.1960
No log	3	2343	1.5029	0.0312	4.2744	0.4357	0.2884
0.0386	4	3124	1.3079	0.0625	5.4587	0.5477	0.4123
1.2767	5	3905	1.2149	0.125	7.9862	0.5661	0.4584
1.1469	6	4686	1.0732	0.25	12.2712	0.6288	0.5500
0.9983	7	5467	1.0156	0.5	21.6274	0.6508	0.5944
0.8947	8.0	6248	1.0242	1.0	40.4642	0.6589	0.6005
0.6881	9.0	7029	1.0714	1.0	40.4926	0.6538	0.5959
0.4692	10.0	7810	1.2413	1.0	39.7291	0.6506	0.6127
0.346	11.0	8591	1.4563	1.0	40.2647	0.6286	0.5885

Framework versions

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