How to Weight Multitask Finetuning? Fast Previews via Bayesian Model-Merging
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910739656081408 |
|---|---|
| author | Maldonado, Hugo Monzón Möllenhoff, Thomas Daheim, Nico Gurevych, Iryna Khan, Mohammad Emtiyaz |
| author_facet | Maldonado, Hugo Monzón Möllenhoff, Thomas Daheim, Nico Gurevych, Iryna Khan, Mohammad Emtiyaz |
| contents | When finetuning multiple tasks altogether, it is important to carefully weigh them to get a good performance, but searching for good weights can be difficult and costly. Here, we propose to aid the search with fast previews to quickly get a rough idea of different reweighting options. We use model merging to create previews by simply reusing and averaging parameters of models trained on each task separately (no retraining required). To improve the quality of previews, we propose a Bayesian approach to design new merging strategies by using more flexible posteriors. We validate our findings on vision and natural-language transformers. Our work shows the benefits of model merging via Bayes to improve multitask finetuning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_08147 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | How to Weight Multitask Finetuning? Fast Previews via Bayesian Model-Merging Maldonado, Hugo Monzón Möllenhoff, Thomas Daheim, Nico Gurevych, Iryna Khan, Mohammad Emtiyaz Machine Learning Artificial Intelligence When finetuning multiple tasks altogether, it is important to carefully weigh them to get a good performance, but searching for good weights can be difficult and costly. Here, we propose to aid the search with fast previews to quickly get a rough idea of different reweighting options. We use model merging to create previews by simply reusing and averaging parameters of models trained on each task separately (no retraining required). To improve the quality of previews, we propose a Bayesian approach to design new merging strategies by using more flexible posteriors. We validate our findings on vision and natural-language transformers. Our work shows the benefits of model merging via Bayes to improve multitask finetuning. |
| title | How to Weight Multitask Finetuning? Fast Previews via Bayesian Model-Merging |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2412.08147 |