Partial Fine-Tuning: A Successor to Full Fine-Tuning for Vision Transformers
Fine-tuning pre-trained foundation models has gained significant popularity in various research fields. Existing methods for fine-tuning can be roughly divided into two categories, namely Parameter-Efficient Fine-Tuning and High-Performance Fine-Tuning. The former aims at improving efficiency, while...
Saved in:
Published in | arXiv.org |
---|---|
Main Authors | , , , , , , |
Format | Paper |
Language | English |
Published |
Ithaca
Cornell University Library, arXiv.org
25.12.2023
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Abstract | Fine-tuning pre-trained foundation models has gained significant popularity in various research fields. Existing methods for fine-tuning can be roughly divided into two categories, namely Parameter-Efficient Fine-Tuning and High-Performance Fine-Tuning. The former aims at improving efficiency, while the latter focuses on enhancing performance. Beyond these methods, we demonstrate that Partial Fine-Tuning can be an innovative and promising direction capable of concurrently enhancing both efficiency and accuracy. We first validate eight manually-defined partial fine-tuning strategies across kinds of datasets and vision transformer architectures, and find that some partial fine-tuning strategies (e.g., ffn only or attention only) can achieve better performance with fewer tuned parameters than full fine-tuning, and selecting appropriate layers is critical to partial fine-tuning. Thus, we propose a novel fine-tuned angle metric to guide the selection of appropriate layers for partial fine-tuning, making it flexible to be adapted to various scenarios for more practicable partial fine-tuning. Additionally, we show that partial fine-tuning can serve as a new dimension for Model Soups, improving both the model performance and generalization with fewer tuned parameters. Comprehensive experiments on a wide range of datasets and models validate the great potential of partial fine-tuning. |
---|---|
AbstractList | Fine-tuning pre-trained foundation models has gained significant popularity in various research fields. Existing methods for fine-tuning can be roughly divided into two categories, namely Parameter-Efficient Fine-Tuning and High-Performance Fine-Tuning. The former aims at improving efficiency, while the latter focuses on enhancing performance. Beyond these methods, we demonstrate that Partial Fine-Tuning can be an innovative and promising direction capable of concurrently enhancing both efficiency and accuracy. We first validate eight manually-defined partial fine-tuning strategies across kinds of datasets and vision transformer architectures, and find that some partial fine-tuning strategies (e.g., ffn only or attention only) can achieve better performance with fewer tuned parameters than full fine-tuning, and selecting appropriate layers is critical to partial fine-tuning. Thus, we propose a novel fine-tuned angle metric to guide the selection of appropriate layers for partial fine-tuning, making it flexible to be adapted to various scenarios for more practicable partial fine-tuning. Additionally, we show that partial fine-tuning can serve as a new dimension for Model Soups, improving both the model performance and generalization with fewer tuned parameters. Comprehensive experiments on a wide range of datasets and models validate the great potential of partial fine-tuning. |
Author | Ye, Peng Li, Minglei Ouyang, Wanli Tu, Chongjun Chen, Tao Huang, Yongqi He, Tong |
Author_xml | – sequence: 1 givenname: Peng surname: Ye fullname: Ye, Peng – sequence: 2 givenname: Yongqi surname: Huang fullname: Huang, Yongqi – sequence: 3 givenname: Chongjun surname: Tu fullname: Tu, Chongjun – sequence: 4 givenname: Minglei surname: Li fullname: Li, Minglei – sequence: 5 givenname: Tao surname: Chen fullname: Chen, Tao – sequence: 6 givenname: Tong surname: He fullname: He, Tong – sequence: 7 givenname: Wanli surname: Ouyang fullname: Ouyang, Wanli |
BookMark | eNqNir0KwjAURoMoWLXvcMG5cE1s_NlELA4OgsW1hJJKSr3R3Ob97eDi5nT4zndmYkye7EgkUqlVtl1LORUpc4uIUm9knqtEXK4m9M50UDiyWRnJ0WMPB7jFurbMPkDvoYjdTwDN4O-OnScogyEe9tMGXohJYzq26ZdzsSxO5fGcvYJ_R8t91foYaLgquUOtNWpE9V_1ATrvPno |
ContentType | Paper |
Copyright | 2023. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License. |
Copyright_xml | – notice: 2023. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License. |
DBID | 8FE 8FG ABJCF ABUWG AFKRA AZQEC BENPR BGLVJ CCPQU DWQXO HCIFZ L6V M7S PIMPY PQEST PQQKQ PQUKI PRINS PTHSS |
DatabaseName | ProQuest SciTech Collection ProQuest Technology Collection Materials Science & Engineering Collection ProQuest Central (Alumni) ProQuest Central ProQuest Central Essentials ProQuest Central Technology Collection ProQuest One Community College ProQuest Central Korea SciTech Premium Collection ProQuest Engineering Collection Engineering Database Publicly Available Content Database ProQuest One Academic Eastern Edition (DO NOT USE) ProQuest One Academic ProQuest One Academic UKI Edition ProQuest Central China Engineering Collection |
DatabaseTitle | Publicly Available Content Database Engineering Database Technology Collection ProQuest Central Essentials ProQuest One Academic Eastern Edition ProQuest Central (Alumni Edition) SciTech Premium Collection ProQuest One Community College ProQuest Technology Collection ProQuest SciTech Collection ProQuest Central China ProQuest Central ProQuest Engineering Collection ProQuest One Academic UKI Edition ProQuest Central Korea Materials Science & Engineering Collection ProQuest One Academic Engineering Collection |
DatabaseTitleList | Publicly Available Content Database |
Database_xml | – sequence: 1 dbid: 8FG name: ProQuest Technology Collection url: https://search.proquest.com/technologycollection1 sourceTypes: Aggregation Database |
DeliveryMethod | fulltext_linktorsrc |
Discipline | Physics |
EISSN | 2331-8422 |
Genre | Working Paper/Pre-Print |
GroupedDBID | 8FE 8FG ABJCF ABUWG AFKRA ALMA_UNASSIGNED_HOLDINGS AZQEC BENPR BGLVJ CCPQU DWQXO FRJ HCIFZ L6V M7S M~E PIMPY PQEST PQQKQ PQUKI PRINS PTHSS |
ID | FETCH-proquest_journals_29066606003 |
IEDL.DBID | 8FG |
IngestDate | Tue Sep 24 22:59:15 EDT 2024 |
IsOpenAccess | true |
IsPeerReviewed | false |
IsScholarly | false |
Language | English |
LinkModel | DirectLink |
MergedId | FETCHMERGED-proquest_journals_29066606003 |
OpenAccessLink | https://www.proquest.com/docview/2906660600/abstract/?pq-origsite=%requestingapplication% |
PQID | 2906660600 |
PQPubID | 2050157 |
ParticipantIDs | proquest_journals_2906660600 |
PublicationCentury | 2000 |
PublicationDate | 20231225 |
PublicationDateYYYYMMDD | 2023-12-25 |
PublicationDate_xml | – month: 12 year: 2023 text: 20231225 day: 25 |
PublicationDecade | 2020 |
PublicationPlace | Ithaca |
PublicationPlace_xml | – name: Ithaca |
PublicationTitle | arXiv.org |
PublicationYear | 2023 |
Publisher | Cornell University Library, arXiv.org |
Publisher_xml | – name: Cornell University Library, arXiv.org |
SSID | ssj0002672553 |
Score | 3.5195851 |
SecondaryResourceType | preprint |
Snippet | Fine-tuning pre-trained foundation models has gained significant popularity in various research fields. Existing methods for fine-tuning can be roughly divided... |
SourceID | proquest |
SourceType | Aggregation Database |
SubjectTerms | Datasets Mathematical models Parameters Performance enhancement Transformers |
Title | Partial Fine-Tuning: A Successor to Full Fine-Tuning for Vision Transformers |
URI | https://www.proquest.com/docview/2906660600/abstract/ |
hasFullText | 1 |
inHoldings | 1 |
isFullTextHit | |
isPrint | |
link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwfV3dS8MwED_miuCbn_gxR0BfQ7t-JJ0votJaxI2iVfY2kjR7knV27at_u7mwqiDsMSSEfN797nK_HMA1l1JwWYa0jEcLaqQfp-N4EVBt1AOTUrNIIVF4MmXZW_g0i2Y9yDouDIZVdjLRCuqyUugjd_FbcmbQtue5QqIXQDXu7eqTYv4ofGfdJNPYAWeEf-IhZzx9_PG2-Iwb7Bz8E7hWi6T74ORipesD6OnlIeza4Eu1PoLnHPdPfJDUQD5atOiquCF35LW12QyrmjQVQVvxbwNi0CZ5t8xwUnTo02C5Y7hKk-Iho90Q5pvjsp7_Ti44gb6x-_UpED8KFIsZ18j-9LUQIfcio5LHpRZ4Wc5gsK2n8-3VF7CHmdMxMsOPBtBv6lZfGv3ayKFduiE498k0fzGlyVfyDf4JhV8 |
link.rule.ids | 786,790,12792,21416,33408,33779,43635,43840 |
linkProvider | ProQuest |
linkToHtml | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwfV1LS8QwEB50F9GbT3ysGtBrsPaRdL2IiLVqd1mwyt5K0mZPYtc-_r8zoVVB2HNCyHO-L5OZfACXUmsldeHzIrxecLR-ko_DhccNwoPQ2oggp0ThyVTEb_7zPJh3Dre6C6vsbaI11EWZk4_8ir4lF8i2Hed2-cVJNYpeVzsJjXUY-h5CJ2WKR48_PhZXSGTM3j8za7Ej2obhTC1NtQNr5nMXNmzIZV7vQTKjVVMfLEKix9OWHBQ37I69tlbDsKxYUzK6If6twJBjsnebD87SnnMig9uHi-ghvY9534Ws2yR19jsk7wAGeNs3h8DcwMtFKKShnE_XKOVLJ0AgHhdG0RE5gtGqlo5XF5_DZpxOkix5mr6cwBZpp1NshhuMYNBUrTlFhG30mZ3Gb5wrgcI |
openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Partial+Fine-Tuning%3A+A+Successor+to+Full+Fine-Tuning+for+Vision+Transformers&rft.jtitle=arXiv.org&rft.au=Ye%2C+Peng&rft.au=Huang%2C+Yongqi&rft.au=Tu%2C+Chongjun&rft.au=Li%2C+Minglei&rft.date=2023-12-25&rft.pub=Cornell+University+Library%2C+arXiv.org&rft.eissn=2331-8422 |