4c4acb58743284ea8a317c7555518635

This model is a fine-tuned version of albert/albert-base-v2 on the nyu-mll/glue dataset. It achieves the following results on the evaluation set:

• Loss: 0.4300
• Data Size: 1.0
• Epoch Runtime: 421.3780
• Accuracy: 0.8788
• F1 Macro: 0.8688
• Rouge1: 0.8787
• Rouge2: 0.0
• Rougel: 0.8788
• Rougelsum: 0.8787

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	Rouge1	Rouge2	Rougel	Rougelsum
No log	0	0	0.7313	0	15.6065	0.3804	0.3430	0.3804	0.0	0.3804	0.3806
0.6539	1	11370	0.6248	0.0078	19.1777	0.6751	0.5924	0.6751	0.0	0.6751	0.6750
0.5853	2	22740	0.5155	0.0156	22.0095	0.7505	0.7337	0.7506	0.0	0.7505	0.7504
0.468	3	34110	0.4447	0.0312	28.5059	0.7901	0.7791	0.7900	0.0	0.7900	0.7901
0.4386	4	45480	0.4299	0.0625	41.3563	0.7999	0.7753	0.7998	0.0	0.8000	0.7999
0.3925	5	56850	0.3993	0.125	65.6416	0.8175	0.8112	0.8175	0.0	0.8176	0.8176
0.3677	6	68220	0.3622	0.25	120.6779	0.8338	0.8261	0.8339	0.0	0.8339	0.8338
0.3276	7	79590	0.3478	0.5	221.8931	0.8390	0.8345	0.8391	0.0	0.8390	0.8391
0.3221	8.0	90960	0.3095	1.0	426.9949	0.8638	0.8538	0.8639	0.0	0.8638	0.8639
0.2799	9.0	102330	0.3082	1.0	426.1677	0.8663	0.8544	0.8663	0.0	0.8663	0.8663
0.242	10.0	113700	0.3059	1.0	427.5854	0.8722	0.8631	0.8722	0.0	0.8723	0.8722
0.2155	11.0	125070	0.3139	1.0	426.4683	0.8714	0.8614	0.8713	0.0	0.8715	0.8714
0.185	12.0	136440	0.3005	1.0	429.3517	0.8770	0.8675	0.8769	0.0	0.8770	0.8770
0.1774	13.0	147810	0.3372	1.0	427.6157	0.8796	0.8708	0.8797	0.0	0.8797	0.8795
0.1644	14.0	159180	0.3492	1.0	425.8594	0.8738	0.8671	0.8739	0.0	0.8739	0.8738
0.1591	15.0	170550	0.3838	1.0	427.9141	0.8742	0.8657	0.8742	0.0	0.8742	0.8742
0.1474	16.0	181920	0.4300	1.0	421.3780	0.8788	0.8688	0.8787	0.0	0.8788	0.8787

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.1