b00f4a20ff414d1eac400fa711d62e96

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

• Loss: 1.5091
• Data Size: 1.0
• Epoch Runtime: 33.6551
• Accuracy: 0.6336
• F1 Macro: 0.5939

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	2.2730	0	2.9031	0.0715	0.0148
No log	1	781	1.9864	0.0078	3.1189	0.2218	0.0403
No log	2	1562	1.6899	0.0156	3.3978	0.3672	0.2065
No log	3	2343	1.4297	0.0312	3.9684	0.4249	0.2596
0.0386	4	3124	1.3480	0.0625	4.9948	0.5278	0.3852
1.308	5	3905	1.1956	0.125	7.0079	0.5753	0.4683
1.1918	6	4686	1.0904	0.25	11.1012	0.6196	0.5430
1.0552	7	5467	1.0691	0.5	18.7857	0.6308	0.5699
0.9404	8.0	6248	1.0493	1.0	35.0722	0.6410	0.5868
0.8081	9.0	7029	1.0377	1.0	35.0726	0.6552	0.5956
0.6622	10.0	7810	1.1103	1.0	33.9116	0.6532	0.5987
0.5362	11.0	8591	1.2328	1.0	34.0819	0.6488	0.5972
0.4803	12.0	9372	1.3523	1.0	33.7882	0.6346	0.5955
0.3479	13.0	10153	1.5091	1.0	33.6551	0.6336	0.5939

Framework versions

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