8ce5fea051c60e45b5bfb442818b11cc

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

• Loss: 1.5598
• Data Size: 1.0
• Epoch Runtime: 34.7128
• 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	2.2256	0	1.5822	0.1114	0.0592
No log	1	500	1.7762	0.0078	1.9378	0.3352	0.1293
No log	2	1000	1.6559	0.0156	2.1808	0.3311	0.1161
No log	3	1500	1.5656	0.0312	2.8947	0.3488	0.0862
No log	4	2000	1.5816	0.0625	4.0184	0.3483	0.0961
0.0875	5	2500	1.5799	0.125	5.9544	0.2908	0.0751
1.5937	6	3000	1.5651	0.25	10.0182	0.2908	0.0751
0.259	7	3500	1.5625	0.5	18.5547	0.3488	0.0862
1.5848	8.0	4000	1.5587	1.0	35.0657	0.3488	0.0862
1.5873	9.0	4500	1.5591	1.0	35.2483	0.3488	0.0862
1.5936	10.0	5000	1.5660	1.0	34.8636	0.2908	0.0751
1.58	11.0	5500	1.5570	1.0	34.8525	0.3488	0.0862
1.5586	12.0	6000	1.5558	1.0	34.8399	0.3488	0.0862
1.562	13.0	6500	1.5581	1.0	34.7223	0.3488	0.0862
1.5844	14.0	7000	1.5560	1.0	34.8527	0.3488	0.0862
1.5679	15.0	7500	1.5622	1.0	34.7357	0.3488	0.0862
1.5766	16.0	8000	1.5598	1.0	34.7128	0.3488	0.0862

Framework versions

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