d370c079423cc0ca0fb2f8ea0a17a30f

This model is a fine-tuned version of openai-community/gpt2-large on the dair-ai/emotion [split] dataset. It achieves the following results on the evaluation set:

• Loss: 0.2662
• Data Size: 1.0
• Epoch Runtime: 108.8033
• Accuracy: 0.9264
• F1 Macro: 0.8804

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	1.7306	0	5.2325	0.2581	0.1535
No log	1	500	1.6054	0.0078	6.1709	0.3679	0.1563
No log	2	1000	1.4602	0.0156	8.0060	0.4698	0.2092
No log	3	1500	0.8167	0.0312	11.2500	0.7198	0.5275
No log	4	2000	0.5123	0.0625	15.4503	0.8377	0.7730
0.0369	5	2500	0.3432	0.125	23.6348	0.8916	0.8542
0.2337	6	3000	0.2065	0.25	37.4914	0.9239	0.8543
0.0266	7	3500	0.1835	0.5	61.4971	0.9259	0.8851
0.144	8.0	4000	0.1397	1.0	111.6013	0.9289	0.8751
0.1002	9.0	4500	0.1341	1.0	107.6271	0.9330	0.8996
0.0917	10.0	5000	0.1911	1.0	108.0822	0.9330	0.8968
0.0326	11.0	5500	0.2329	1.0	107.5770	0.9320	0.8923
0.0382	12.0	6000	0.2385	1.0	108.0987	0.9315	0.8935
0.0476	13.0	6500	0.2662	1.0	108.8033	0.9264	0.8804

Framework versions

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