# Image-to-Text Generation

Dimple is the first discrete diffusion multimodal large language model (DMLLM) that combines autoregressive and diffusion training paradigms. After training on the same dataset as LLaVA-NEXT, it outperforms LLaVA-NEXT-7B by 3.9%.

Magma-8B is an image-text-to-text conversion model based on the GGUF format, suitable for multimodal task processing.

This model is a GGUF format conversion of llava-hf/llava-1.5-7b-hf, supporting image-to-text generation tasks.

This model is a 3-bit quantized version converted from google/gemma-3-27b-it-qat-q4_0-unquantized to the MLX format, suitable for image-to-text tasks.

Gemma 3 27B IT QAT 4bit is an MLX-format model converted from Google's original model, supporting image-to-text tasks.

Gemma 3 is a lightweight open-source multimodal model series by Google, supporting text and image inputs with text generation capabilities. This version is a 27B parameter instruction-tuned model using quantization-aware training, offering lower memory requirements while maintaining near-original quality.

This is an MLX-converted version of the Google Gemma 3 27B IT model, supporting image-text-to-text tasks.

This is an image-to-text conversion model capable of processing both image and text inputs to generate corresponding text outputs.

This is a weighted/importance matrix quantized version of the Minthy/ToriiGate-v0.4-7B model, offering multiple quantization options to suit different needs.

Gemma is a lightweight cutting-edge open-source multimodal model series launched by Google, built on the technology used to create Gemini models, supporting text and image inputs to generate text outputs.

A Korean vision-language model based on Qwen2-VL-7B-Instruct, supporting image-to-text tasks

A BitsAndBytes NF4 quantized version based on Aria-sequential_mlp, suitable for image-to-text tasks with approximately 15.5 GB VRAM requirement.

Doubutsu is a lightweight vision-language model series, specifically designed for customized scenario fine-tuning.

This is a spatially-aware vision-language model capable of recognizing spatial relationships between objects in images and generating descriptive text.

Pix2Struct is a vision-language understanding model pretrained for image-to-text conversion tasks, specifically optimized for high-resolution infographic visual question answering.

Pix2Struct is a visual question answering model fine-tuned for OCR-VQA tasks, capable of parsing textual content in images and answering questions

Pix2Struct is an image encoder-text decoder model trained on image-text pairs for multitasking, specifically fine-tuned for chart question answering tasks

GIT is a vision-language model based on a Transformer decoder, trained with dual conditioning on CLIP image tokens and text tokens, specifically optimized for TextVQA tasks.

GIT is a Transformer decoder based on CLIP image tokens and text tokens, trained on large-scale image-text pairs, suitable for tasks like visual question answering.

GIT is a Transformer-based vision-language model capable of converting images into textual descriptions, specifically fine-tuned for TextVQA tasks.