Alignment-enriched tuning for patch-level pre-trained document image models
Alignment between image and text has shown promising im provements on patch-level pre-trained document image mod els. However, investigating more effective or finer-grained alignment techniques during pre-training requires a large amount of computation cost and time. Thus, a question natu rally aris...
محفوظ في:
المؤلفون الرئيسيون: | WANG, Lei, HE, Jiabang, XU, Xing, LIU, Ning, LIU, Hui |
---|---|
التنسيق: | text |
اللغة: | English |
منشور في: |
Institutional Knowledge at Singapore Management University
2023
|
الموضوعات: | |
الوصول للمادة أونلاين: | https://ink.library.smu.edu.sg/sis_research/9318 https://ink.library.smu.edu.sg/context/sis_research/article/10318/viewcontent/Alignment_Enriched_pv.pdf |
الوسوم: |
إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
|
مواد مشابهة
-
Mitigating fine-grained hallucination by fine-tuning large vision-language models with caption rewrites
بواسطة: WANG, Lei, وآخرون
منشور في: (2024) -
ICL-D3IE: In-Context Learning with Diverse Demonstrations Updating for Document Information Extraction
بواسطة: HE, Jiabang, وآخرون
منشور في: (2023) -
NumGPT: Improving numeracy ability of generative pre-trained models
بواسطة: JIN, Zhihua, وآخرون
منشور في: (2023) -
ROME: Evaluating pre-trained vision-language models on reasoning beyond visual common sense
بواسطة: ZHOU, Kankan, وآخرون
منشور في: (2023) -
Cross-domain graph anomaly detection via anomaly-aware contrastive alignment
بواسطة: WANG, Qizhou, وآخرون
منشور في: (2023)