Synth$^2$: Boosting Visual-Language Models with Synthetic Captions and Image Embeddings

• URL: http://arxiv.org/abs/2403.07750v2
• Date: Fri, 7 Jun 2024 12:10:47 GMT
• Title: Synth$^2$: Boosting Visual-Language Models with Synthetic Captions and Image Embeddings
• Authors: Sahand Sharifzadeh, Christos Kaplanis, Shreya Pathak, Dharshan Kumaran, Anastasija Ilic, Jovana Mitrovic, Charles Blundell, Andrea Banino,
• Abstract summary: We create synthetic image-text pairs for efficient and effective Visual-Language Models (VLMs) training. Our method employs a pretrained text-to-image model to synthesize image embeddings from captions generated by an LLM. Our VLM, finetuned on synthetic data achieves comparable performance to models trained solely on human-annotated data.
• Score: 16.28853186016663
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The creation of high-quality human-labeled image-caption datasets presents a significant bottleneck in the development of Visual-Language Models (VLMs). In this work, we investigate an approach that leverages the strengths of Large Language Models (LLMs) and image generation models to create synthetic image-text pairs for efficient and effective VLM training. Our method employs a pretrained text-to-image model to synthesize image embeddings from captions generated by an LLM. Despite the text-to-image model and VLM initially being trained on the same data, our approach leverages the image generator's ability to create novel compositions, resulting in synthetic image embeddings that expand beyond the limitations of the original dataset. Extensive experiments demonstrate that our VLM, finetuned on synthetic data achieves comparable performance to models trained solely on human-annotated data, while requiring significantly less data. Furthermore, we perform a set of analyses on captions which reveals that semantic diversity and balance are key aspects for better downstream performance. Finally, we show that synthesizing images in the image embedding space is 25\% faster than in the pixel space. We believe our work not only addresses a significant challenge in VLM training but also opens up promising avenues for the development of self-improving multi-modal models.

Related papers

• Meissonic: Revitalizing Masked Generative Transformers for Efficient High-Resolution Text-to-Image Synthesis [62.06970466554273]
  We present Meissonic, which non-autoregressive masked image modeling (MIM) text-to-image elevates to a level comparable with state-of-the-art diffusion models like SDXL. We leverage high-quality training data, integrate micro-conditions informed by human preference scores, and employ feature compression layers to further enhance image fidelity and resolution. Our model not only matches but often exceeds the performance of existing models like SDXL in generating high-quality, high-resolution images.
  arXiv  Detail & Related papers  (2024-10-10T17:59:17Z)
• SynthVLM: High-Efficiency and High-Quality Synthetic Data for Vision Language Models [39.55942000935765]
  We introduce SynthVLM, a novel data synthesis pipeline for Vision Large Language Models (VLLMs) Unlike existing methods that generate captions from images, SynthVLM employs advanced diffusion models and high-quality captions to automatically generate and select high-resolution images from captions. We achieve state-of-the-art (SoTA) performance on various vision question answering tasks, maintaining high alignment quality and preserving advanced language abilities.
  arXiv  Detail & Related papers  (2024-07-30T11:57:40Z)
• Is Synthetic Image Useful for Transfer Learning? An Investigation into Data Generation, Volume, and Utilization [62.157627519792946]
  We introduce a novel framework called bridged transfer, which initially employs synthetic images for fine-tuning a pre-trained model to improve its transferability. We propose dataset style inversion strategy to improve the stylistic alignment between synthetic and real images. Our proposed methods are evaluated across 10 different datasets and 5 distinct models, demonstrating consistent improvements.
  arXiv  Detail & Related papers  (2024-03-28T22:25:05Z)
• Diversify, Don't Fine-Tune: Scaling Up Visual Recognition Training with Synthetic Images [37.29348016920314]
  We present a new framework leveraging off-the-shelf generative models to generate synthetic training images. We address class name ambiguity, lack of diversity in naive prompts, and domain shifts. Our framework consistently enhances recognition model performance with more synthetic data.
  arXiv  Detail & Related papers  (2023-12-04T18:35:27Z)
• Image Captions are Natural Prompts for Text-to-Image Models [70.30915140413383]
  We analyze the relationship between the training effect of synthetic data and the synthetic data distribution induced by prompts. We propose a simple yet effective method that prompts text-to-image generative models to synthesize more informative and diverse training data. Our method significantly improves the performance of models trained on synthetic training data.
  arXiv  Detail & Related papers  (2023-07-17T14:38:11Z)
• UniDiff: Advancing Vision-Language Models with Generative and Discriminative Learning [86.91893533388628]
  This paper presents UniDiff, a unified multi-modal model that integrates image-text contrastive learning (ITC), text-conditioned image synthesis learning (IS), and reciprocal semantic consistency modeling (RSC) UniDiff demonstrates versatility in both multi-modal understanding and generative tasks.
  arXiv  Detail & Related papers  (2023-06-01T15:39:38Z)
• Generating Images with Multimodal Language Models [78.6660334861137]
  We propose a method to fuse frozen text-only large language models with pre-trained image encoder and decoder models. Our model demonstrates a wide suite of multimodal capabilities: image retrieval, novel image generation, and multimodal dialogue.
  arXiv  Detail & Related papers  (2023-05-26T19:22:03Z)
• DiVAE: Photorealistic Images Synthesis with Denoising Diffusion Decoder [73.1010640692609]
  We propose a VQ-VAE architecture model with a diffusion decoder (DiVAE) to work as the reconstructing component in image synthesis. Our model achieves state-of-the-art results and generates more photorealistic images specifically.
  arXiv  Detail & Related papers  (2022-06-01T10:39:12Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.