11950907cd8ead80f741495e5107b632

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

• Loss: 1.4652
• Data Size: 1.0
• Epoch Runtime: 22.7181
• Accuracy: 0.6438
• F1 Macro: 0.6014

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	2.2284	0	1.9242	0.0675	0.0193
No log	1	781	1.9595	0.0078	2.4514	0.2272	0.0538
No log	2	1562	1.6313	0.0156	2.3766	0.4455	0.2698
No log	3	2343	1.4143	0.0312	2.8401	0.4645	0.2870
0.0373	4	3124	1.2584	0.0625	3.4011	0.5547	0.4002
1.2599	5	3905	1.1884	0.125	4.8737	0.5771	0.4750
1.1543	6	4686	1.0718	0.25	7.5413	0.6280	0.5237
0.9985	7	5467	1.0131	0.5	12.6351	0.6490	0.5752
0.8882	8.0	6248	1.0183	1.0	23.3045	0.6520	0.5988
0.6949	9.0	7029	1.0816	1.0	22.3227	0.6548	0.6035
0.5016	10.0	7810	1.2134	1.0	22.7539	0.6448	0.5891
0.3539	11.0	8591	1.4652	1.0	22.7181	0.6438	0.6014

Framework versions

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