moralstories-bart-moralaction.situation-gen

This model is a fine-tuned version of facebook/bart-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0824

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-06
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_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: linear
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.9984	312	1.7346
No log	2.0	625	0.1568
No log	2.9984	937	0.0850
No log	4.0	1250	0.0824
No log	4.9984	1562	0.0816
No log	6.0	1875	0.0811
1.1476	6.9984	2187	0.0807
1.1476	8.0	2500	0.0805
1.1476	8.9984	2812	0.0804
1.1476	10.0	3125	0.0803
1.1476	10.9984	3437	0.0803
1.1476	12.0	3750	0.0803
0.0822	12.9984	4062	0.0804
0.0822	14.0	4375	0.0804
0.0822	14.9984	4687	0.0803
0.0822	16.0	5000	0.0805
0.0822	16.9984	5312	0.0804
0.0822	18.0	5625	0.0806
0.0822	18.9984	5937	0.0807
0.0769	20.0	6250	0.0806
0.0769	20.9984	6562	0.0807
0.0769	22.0	6875	0.0808
0.0769	22.9984	7187	0.0809
0.0769	24.0	7500	0.0810
0.0769	24.9984	7812	0.0810
0.0733	26.0	8125	0.0811
0.0733	26.9984	8424	0.0813
0.0733	28.0	8737	0.0813
0.0733	28.9984	9049	0.0814
0.0733	30.0	9362	0.0814
0.0733	30.9984	9674	0.0815
0.0733	32.0	9987	0.0816
0.0704	32.9984	10299	0.0817
0.0704	34.0	10612	0.0817
0.0704	34.9984	10924	0.0819
0.0704	36.0	11237	0.0819
0.0704	36.9984	11549	0.0820
0.0704	38.0	11862	0.0821
0.0686	38.9984	12174	0.0821
0.0686	40.0	12487	0.0821
0.0686	40.9984	12792	0.0822
0.0686	42.0	13105	0.0823
0.0686	42.9984	13417	0.0823
0.0686	44.0	13730	0.0823
0.0673	44.9984	14042	0.0823
0.0673	46.0	14355	0.0824
0.0673	46.9984	14667	0.0824
0.0673	48.0	14980	0.0824
0.0673	48.9984	15292	0.0824
0.0673	49.984	15600	0.0824

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1
• Datasets 3.2.0
• Tokenizers 0.20.1