0511bffb692c29030fcba400f932b363

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

• Loss: 1.3443
• Data Size: 1.0
• Epoch Runtime: 61.7330
• Accuracy: 0.6508
• F1 Macro: 0.6251

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	2.2538	0	4.3783	0.0715	0.0148
No log	1	781	1.9748	0.0078	4.9397	0.2218	0.0403
No log	2	1562	1.7445	0.0156	5.4141	0.3017	0.1939
No log	3	2343	1.4571	0.0312	6.6285	0.4459	0.2850
0.0388	4	3124	1.2784	0.0625	8.3829	0.5481	0.4046
1.2784	5	3905	1.1746	0.125	12.3872	0.5942	0.5200
1.1825	6	4686	1.0564	0.25	19.4676	0.6330	0.5631
1.054	7	5467	1.0225	0.5	34.4533	0.6450	0.5873
0.9557	8.0	6248	1.0341	1.0	65.3527	0.6516	0.6015
0.8404	9.0	7029	0.9923	1.0	62.2612	0.6627	0.5970
0.6372	10.0	7810	1.1338	1.0	63.8278	0.6462	0.5923
0.5409	11.0	8591	1.1506	1.0	62.5502	0.6619	0.6128
0.4781	12.0	9372	1.3039	1.0	64.2764	0.6466	0.6031
0.4248	13.0	10153	1.3443	1.0	61.7330	0.6508	0.6251

Framework versions

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