Romance-cleaned-1

This model is a fine-tuned version of gpt2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 4.7175

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 1000
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.97	29	9.9497
No log	1.97	58	9.1816
No log	2.97	87	8.5947
No log	3.97	116	8.2217
No log	4.97	145	7.8354
No log	5.97	174	7.5075
No log	6.97	203	7.2112
No log	7.97	232	6.9077
No log	8.97	261	6.5994
No log	9.97	290	6.3077
No log	10.97	319	6.0416
No log	11.97	348	5.8126
No log	12.97	377	5.6197
No log	13.97	406	5.4789
No log	14.97	435	5.3665
No log	15.97	464	5.2738
No log	16.97	493	5.1942
No log	17.97	522	5.1382
No log	18.97	551	5.0784
No log	19.97	580	5.0347
No log	20.97	609	4.9873
No log	21.97	638	4.9514
No log	22.97	667	4.9112
No log	23.97	696	4.8838
No log	24.97	725	4.8468
No log	25.97	754	4.8221
No log	26.97	783	4.7996
No log	27.97	812	4.7815
No log	28.97	841	4.7606
No log	29.97	870	4.7394
No log	30.97	899	4.7167
No log	31.97	928	4.7140
No log	32.97	957	4.6910
No log	33.97	986	4.6844
No log	34.97	1015	4.6765
No log	35.97	1044	4.6687
No log	36.97	1073	4.6721
No log	37.97	1102	4.6724
No log	38.97	1131	4.6629
No log	39.97	1160	4.6772
No log	40.97	1189	4.6795
No log	41.97	1218	4.6788
No log	42.97	1247	4.6832
No log	43.97	1276	4.6954
No log	44.97	1305	4.7009
No log	45.97	1334	4.7082
No log	46.97	1363	4.7140
No log	47.97	1392	4.7158
No log	48.97	1421	4.7181
No log	49.97	1450	4.7175

Framework versions

• Transformers 4.24.0
• Pytorch 1.12.1+cu113
• Datasets 2.6.1
• Tokenizers 0.13.1