DReAMy-lib/t5-base-DreamBank-Generation-Emot-Char

t5-base-DreamBank-Generation-Char

This model is a fine-tuned version of t5-base on the DB emotion classification. It achieves the following results on the evaluation set (please note they refer to best uploaded model):

• Loss: 0.3047
• Rouge1: 0.8609
• Rouge2: 0.7956
• Rougel: 0.8476
• Rougelsum: 0.8578

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: 0.0002
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
No log	1.0	24	0.4863	0.7670	0.6655	0.7575	0.7634
No log	2.0	48	0.4284	0.6870	0.5207	0.6846	0.6875
No log	3.0	72	0.3541	0.7659	0.6742	0.7600	0.7625
No log	4.0	96	0.3211	0.8147	0.7251	0.7965	0.8078
No log	5.0	120	0.3103	0.8400	0.7747	0.8313	0.8371
No log	6.0	144	0.3220	0.8538	0.7867	0.8285	0.8515
No log	7.0	168	0.3047	0.8609	0.7956	0.8476	0.8578
No log	8.0	192	0.3106	0.8574	0.7836	0.8401	0.8509
No log	9.0	216	0.3054	0.8532	0.7857	0.8378	0.8481
No log	10.0	240	0.3136	0.8455	0.7789	0.8282	0.8432

Framework versions

• Transformers 4.25.1
• Pytorch 1.12.1
• Datasets 2.5.1
• Tokenizers 0.12.1

Cite

Should you use our models in your work, please consider citing us as:

@article{BERTOLINI2024406,
title = {DReAMy: a library for the automatic analysis and annotation of dream reports with multilingual large language models},
journal = {Sleep Medicine},
volume = {115},
pages = {406-407},
year = {2024},
note = {Abstracts from the 17th World Sleep Congress},
issn = {1389-9457},
doi = {https://doi.org/10.1016/j.sleep.2023.11.1092},
url = {https://www.sciencedirect.com/science/article/pii/S1389945723015186},
author = {L. Bertolini and A. Michalak and J. Weeds}
}