e397a43c01ae5d125ec5dc0e3c5002b4

This model is a fine-tuned version of albert/albert-xxlarge-v2 on the ccdv/patent-classification [abstract] dataset. It achieves the following results on the evaluation set:

• Loss: 1.5873
• Data Size: 1.0
• Epoch Runtime: 191.1080
• Accuracy: 0.6735
• F1 Macro: 0.6420

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	2.5834	0	12.0978	0.1963	0.0508
No log	1	781	2.0697	0.0078	13.9520	0.2222	0.0483
No log	2	1562	2.0672	0.0156	15.2769	0.1715	0.0808
No log	3	2343	1.5716	0.0312	18.2857	0.3738	0.2166
0.046	4	3124	1.2891	0.0625	23.7702	0.5655	0.4088
1.2296	5	3905	1.1118	0.125	35.0492	0.6150	0.5408
1.138	6	4686	1.1705	0.25	57.5451	0.5974	0.5445
0.9872	7	5467	1.0099	0.5	102.1596	0.6607	0.5944
0.9517	8.0	6248	0.9450	1.0	192.3713	0.6761	0.6255
0.7609	9.0	7029	1.0006	1.0	191.8248	0.6785	0.6342
0.5183	10.0	7810	1.1087	1.0	191.5371	0.6723	0.6366
0.3436	11.0	8591	1.2662	1.0	191.7371	0.6637	0.6364
0.28	12.0	9372	1.5873	1.0	191.1080	0.6735	0.6420

Framework versions

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