3e178a39e286163561f9a670e11465a1

This model is a fine-tuned version of albert/albert-large-v1 on the ccdv/patent-classification [abstract] dataset. It achieves the following results on the evaluation set:

• Loss: 1.2288
• Data Size: 1.0
• Epoch Runtime: 48.9799
• Accuracy: 0.6400
• F1 Macro: 0.5982

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	2.2511	0	3.4305	0.1635	0.0531
No log	1	781	2.0006	0.0078	4.0544	0.2392	0.0804
No log	2	1562	1.8515	0.0156	4.3115	0.3255	0.1657
No log	3	2343	1.5473	0.0312	4.8879	0.3846	0.2308
0.04	4	3124	1.3605	0.0625	6.2615	0.5218	0.3745
1.3709	5	3905	1.3125	0.125	9.3563	0.5415	0.3865
1.2754	6	4686	1.1844	0.25	14.9678	0.5807	0.4390
1.1531	7	5467	1.1276	0.5	26.2589	0.6096	0.5406
1.0392	8.0	6248	1.0762	1.0	51.5742	0.6274	0.5668
0.9615	9.0	7029	1.0177	1.0	49.7864	0.6458	0.5840
0.8491	10.0	7810	1.0766	1.0	51.6486	0.6400	0.5918
0.745	11.0	8591	1.0970	1.0	51.0736	0.6396	0.5917
0.6966	12.0	9372	1.1589	1.0	49.6696	0.6464	0.5859
0.5373	13.0	10153	1.2288	1.0	48.9799	0.6400	0.5982

Framework versions

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