dc6ebca90f984b4fab153b0e00d8d8d0

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

• Loss: 0.2672
• Data Size: 1.0
• Epoch Runtime: 15.2462
• Accuracy: 0.9249
• F1 Macro: 0.8771

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	1.7712	0	1.1315	0.2056	0.0948
No log	1	500	1.5903	0.0078	1.2928	0.3876	0.1583
No log	2	1000	1.5167	0.0156	1.5165	0.3533	0.0918
No log	3	1500	1.1937	0.0312	1.8105	0.5771	0.2334
No log	4	2000	0.8724	0.0625	2.2575	0.7278	0.4852
0.057	5	2500	0.4663	0.125	3.0341	0.8584	0.7708
0.3619	6	3000	0.2800	0.25	4.8857	0.9052	0.8561
0.0346	7	3500	0.2379	0.5	7.9875	0.9219	0.8781
0.1699	8.0	4000	0.1552	1.0	15.3764	0.9274	0.8864
0.1242	9.0	4500	0.1697	1.0	15.5714	0.9289	0.8863
0.1155	10.0	5000	0.2159	1.0	15.2951	0.9204	0.8809
0.0812	11.0	5500	0.2742	1.0	15.2826	0.9279	0.8864
0.0844	12.0	6000	0.2672	1.0	15.2462	0.9249	0.8771

Framework versions

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