a0f79a896e68a707c5573943a8412a7c

This model is a fine-tuned version of albert/albert-large-v1 on the dair-ai/emotion [split] dataset. It achieves the following results on the evaluation set:

• Loss: 1.5603
• Data Size: 1.0
• Epoch Runtime: 30.6626
• Accuracy: 0.3488
• F1 Macro: 0.0862

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	1.7845	0	1.7295	0.1280	0.0786
No log	1	500	1.5945	0.0078	2.1669	0.3443	0.1119
No log	2	1000	1.6931	0.0156	2.1927	0.2117	0.1254
No log	3	1500	1.5775	0.0312	2.8564	0.3483	0.0872
No log	4	2000	1.5812	0.0625	3.7361	0.3483	0.0862
0.0877	5	2500	1.6129	0.125	5.6306	0.2908	0.0751
1.6203	6	3000	1.5949	0.25	9.1866	0.3488	0.0862
0.2611	7	3500	1.5763	0.5	16.4267	0.3488	0.0862
1.5936	8.0	4000	1.5636	1.0	32.4035	0.3488	0.0862
1.599	9.0	4500	1.5623	1.0	32.0188	0.3488	0.0862
1.5993	10.0	5000	1.5638	1.0	30.5444	0.3488	0.0862
1.5884	11.0	5500	1.5584	1.0	30.3403	0.3488	0.0862
1.5615	12.0	6000	1.5633	1.0	30.4257	0.3488	0.0862
1.5618	13.0	6500	1.5572	1.0	30.2851	0.3488	0.0862
1.5887	14.0	7000	1.5608	1.0	30.6493	0.3488	0.0862
1.5723	15.0	7500	1.5674	1.0	31.3983	0.2908	0.0751
1.5818	16.0	8000	1.5597	1.0	31.8899	0.3488	0.0862
1.5675	17.0	8500	1.5603	1.0	30.6626	0.3488	0.0862

Framework versions

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