0fc23bb14ab994dacee4a1a4c0cd8b43

This model is a fine-tuned version of distilbert/distilbert-base-cased-distilled-squad on the dair-ai/emotion [split] dataset. It achieves the following results on the evaluation set:

• Loss: 0.2628
• Data Size: 1.0
• Epoch Runtime: 14.3628
• Accuracy: 0.9229
• F1 Macro: 0.8745

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	1.7813	0	1.1389	0.2445	0.0961
No log	1	500	1.5856	0.0078	1.3428	0.3478	0.1360
No log	2	1000	1.5292	0.0156	1.3492	0.3488	0.0862
No log	3	1500	1.1536	0.0312	1.6680	0.5907	0.2690
No log	4	2000	0.7951	0.0625	2.1807	0.7485	0.5522
0.0538	5	2500	0.4097	0.125	3.0200	0.8755	0.8209
0.3276	6	3000	0.2715	0.25	4.6168	0.9108	0.8571
0.0335	7	3500	0.2275	0.5	7.8049	0.9173	0.8775
0.1705	8.0	4000	0.1569	1.0	14.9012	0.9289	0.8867
0.1238	9.0	4500	0.1763	1.0	14.5510	0.9299	0.8850
0.1172	10.0	5000	0.1938	1.0	14.1771	0.9325	0.8939
0.0754	11.0	5500	0.1972	1.0	14.5075	0.9279	0.8901
0.0754	12.0	6000	0.2628	1.0	14.3628	0.9229	0.8745

Framework versions

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