Co Fine Tuning With Vision Language Preservation

via “vision-language model evaluation with unified vlm interface”

Microsoft's unified LLM evaluation and prompt robustness benchmark.

Unique: Implements VLMModel as a parallel factory to LLMModel, maintaining architectural consistency while handling image preprocessing, encoding, and provider-specific vision APIs. Automatically normalizes image inputs across providers with different resolution and format requirements.

vs others: More specialized than LangChain's vision support because it's optimized for systematic evaluation of vision robustness rather than general-purpose multimodal chaining, enabling fine-grained control over image perturbations and evaluation metrics.

via “instruction-tuned multimodal generation with alignment”

Meta's largest open multimodal model at 90B parameters.

Unique: Provides both base and instruction-tuned variants, allowing users to choose between raw model capability and aligned behavior, with torchtune framework enabling custom fine-tuning on proprietary instruction datasets

vs others: Open-weight instruction-tuned variants enable custom alignment without relying on proprietary API providers, though fine-tuning infrastructure requirements are higher than using managed APIs

via “fine-tuning with torchtune framework”

Meta's multimodal 11B model with text and vision.

Unique: Integrated torchtune support enables local fine-tuning without proprietary cloud training APIs. Framework abstracts distributed training complexity, allowing single-GPU fine-tuning with gradient checkpointing and memory optimization. Instruction-tuned base variants available as starting points for task-specific alignment.

vs others: Local fine-tuning with torchtune avoids vendor lock-in and cloud training costs of alternatives like OpenAI fine-tuning API or Anthropic Claude fine-tuning, while maintaining full control over training data and process.

via “fine-tuning on custom vision tasks”

Microsoft's unified model for diverse vision tasks.

Unique: Supports fine-tuning on custom vision tasks while preserving multi-task capabilities through task-specific prompt tokens, enabling domain adaptation without losing general-purpose vision abilities

vs others: More flexible than task-specific fine-tuning (e.g., YOLO fine-tuning) because it preserves multi-task functionality; LoRA fine-tuning is more efficient than full fine-tuning but with slight accuracy trade-offs

via “fine-tuning and model adaptation for custom tasks”

Tiny vision-language model for edge devices.

Unique: Modular fine-tuning system that freezes vision encoder and adapts text encoder/decoder and region encoder independently, reducing training data and compute requirements; includes reference dataset loaders for document VQA and chart QA, enabling task-specific adaptation without custom data pipeline engineering.

vs others: Faster fine-tuning than full model retraining due to frozen vision encoder; more flexible than fixed pre-trained models, though requires more engineering than simple prompt engineering.

via “visual-question-answering-with-instruction-tuning”

Open multimodal model for visual reasoning.

Unique: Uses GPT-4-generated synthetic instruction-tuning data (158K samples) rather than human-annotated datasets, enabling rapid training in ~1 day on 8 A100 GPUs while maintaining strong performance; frozen CLIP encoder + learned projection matrix is simpler than full vision encoder fine-tuning but trades adaptability for training efficiency

vs others: Faster to train and deploy than full vision-language models like BLIP-2 or Flamingo because it freezes the vision encoder and uses synthetic training data, while achieving competitive VQA performance at lower computational cost

via “vision encoder + language model alignment via instruction tuning”

150K visual instruction examples for multimodal model training.

Unique: Demonstrates that instruction tuning with GPT-4V-generated examples can effectively align independent vision and language components without end-to-end pre-training. The dataset is specifically structured to bridge the modality gap through instruction-following rather than contrastive or generative pre-training objectives.

vs others: More efficient than end-to-end vision-language pre-training (BLIP, ALBEF) because it reuses frozen encoders; more practical than datasets requiring human annotation at scale; stronger alignment signal than generic image-text pairs because examples are instruction-grounded.

via “co-fine-tuning-with-vision-language-preservation”

Google's vision-language-action model for robotics.

Unique: Implements co-fine-tuning by representing actions as text tokens within the language modeling framework, allowing the same transformer architecture to simultaneously optimize for vision-language understanding and robotic action prediction without separate policy heads

vs others: Preserves semantic understanding from web-scale vision-language pretraining better than standard fine-tuning by maintaining both vision and text encoder knowledge, while avoiding the computational overhead of separate policy networks or adapter modules

via “transfer learning and domain-specific fine-tuning with frozen vision encoder”

image-to-text model by undefined. 5,97,442 downloads.

Unique: Enables parameter-efficient fine-tuning by freezing the ViT encoder (which contains ~86M parameters) and only updating Q-Former (~190M) and OPT decoder (~2.7B), reducing memory footprint and training time by ~40% compared to full model fine-tuning while maintaining strong performance on downstream tasks.

vs others: More efficient than fine-tuning full vision-language models like BLIP-2-OPT-6.7B; more flexible than fixed-feature extraction because the Q-Former and decoder can adapt to domain-specific patterns.

via “vision-language image-to-image editing instruction refinement”

[CVPR 2026] PromptEnhancer is a prompt-rewriting tool, refining prompts into clearer, structured versions for better image generation.

Unique: Implements multi-modal chain-of-thought reasoning that jointly analyzes image content and editing instructions, grounding the instruction refinement in actual visual elements rather than processing text in isolation. This enables spatial awareness and visual context integration that text-only prompt enhancement cannot achieve.

vs others: Produces more spatially-aware and visually-grounded editing instructions than text-only prompt enhancement because it analyzes the actual image content, reducing ambiguity and improving downstream image-to-image model performance on complex edits.

via “vision-language task adaptation with minimal fine-tuning”

* ⭐ 09/2022: [PaLI: A Jointly-Scaled Multilingual Language-Image Model (PaLI)](https://arxiv.org/abs/2209.06794)

Unique: Leverages the unified representation space created during joint vision-language pretraining, where images and text are encoded in the same semantic space. This enables task adaptation without separate vision and language encoders, reducing model complexity and improving cross-modal reasoning.

vs others: Requires less task-specific fine-tuning than dual-encoder approaches (CLIP-based systems) because the shared transformer has already learned to align visual and linguistic patterns, making it easier to adapt to new vision-language tasks.

via “vision-language model instruction tuning via image-text pair alignment”

* ⭐ 04/2023: [Align your Latents: High-Resolution Video Synthesis with Latent Diffusion Models (VideoLDM)](https://arxiv.org/abs/2304.08818)

Unique: Introduces a systematic two-stage alignment approach that decouples vision encoding from language understanding, using adapter modules and LoRA-style parameter-efficient fine-tuning to maintain frozen pre-trained weights while achieving strong instruction-following performance. This contrasts with end-to-end training approaches by reducing memory overhead and enabling faster iteration on instruction datasets.

vs others: More parameter-efficient and faster to train than full model fine-tuning (e.g., BLIP-2, LLaVA v1.0 early approaches) while achieving comparable or superior instruction-following accuracy through explicit alignment objectives rather than implicit joint training.

via “vision-language-model-architecture-patterns”


Unique: Systematically covers architectural trade-offs (frozen vs. trainable, early vs. late fusion, adapter design) specific to vision-language systems, rather than treating them as straightforward combinations of existing models

vs others: More practical than individual model papers because it abstracts patterns across CLIP, BLIP, LLaVA, and other systems, enabling builders to make informed architectural choices

via “visio-linguistic alignment probing and diagnostic evaluation”

* ⭐ 04/2022: [Winoground: Probing Vision and Language Models for Visio-Linguistic... (Winoground)](https://arxiv.org/abs/2204.03162)

Unique: Introduces Winoground benchmark specifically designed to test visio-linguistic alignment through minimal-difference contrastive pairs, moving beyond standard image-text retrieval metrics to probe fine-grained semantic understanding — distinct from generic vision-language benchmarks that measure retrieval or generation quality

vs others: More sensitive to semantic alignment failures than Flickr30K or COCO retrieval benchmarks because it uses adversarial minimal-difference pairs that expose brittleness in learned representations