8cc0d0aed6f200dec30717cb07453458

This model is a fine-tuned version of google-bert/bert-large-cased-whole-word-masking on the dair-ai/emotion [split] dataset. It achieves the following results on the evaluation set:

• Loss: 0.2125
• Data Size: 1.0
• Epoch Runtime: 52.8025
• Accuracy: 0.9279
• F1 Macro: 0.8810

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	1.8363	0	2.6477	0.1799	0.0803
No log	1	500	1.6056	0.0078	2.9257	0.3432	0.1137
No log	2	1000	1.4672	0.0156	3.4868	0.3851	0.1253
No log	3	1500	0.9200	0.0312	4.8706	0.6653	0.3736
No log	4	2000	0.5436	0.0625	6.9309	0.8241	0.7734
0.0458	5	2500	0.4830	0.125	11.2362	0.8564	0.7954
0.2939	6	3000	0.3066	0.25	16.2892	0.8972	0.7856
0.0438	7	3500	0.2398	0.5	28.0024	0.9153	0.8655
0.249	8.0	4000	0.1330	1.0	52.7571	0.9345	0.8952
0.1161	9.0	4500	0.1790	1.0	53.8604	0.9304	0.8731
0.145	10.0	5000	0.2204	1.0	51.9929	0.9269	0.8813
0.1796	11.0	5500	0.2342	1.0	53.4157	0.9239	0.8842
0.1677	12.0	6000	0.2125	1.0	52.8025	0.9279	0.8810

Framework versions

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