prithivMLmods/SpatialBlock-7B-reason-GGUF

SpatialBlock-7B-reason-GGUF

SpatialBlock-7B-reason is an open-source multimodal model released on Hugging Face by rsoohyun under the Apache-2.0 license, developed to advance spatial reasoning capabilities in Large Vision-Language Models (LVLMs). Based on the Qwen/Qwen2.5-VL-7B-Instruct architecture and compatible with Hugging Face's transformers library under the image-text-to-text pipeline, the model is fine-tuned on the synthetic SpatialBlock-15k dataset as introduced in the research paper SpatialBlock: Enhancing Spatial Intelligence in LVLMs via Synthetic Block-Stacking Problem. It is designed to tackle complex spatial intelligence benchmarks by reasoning through and directly predicting answers for challenging tasks such as 3D-to-2D projections, viewpoint transformations, and structural combinations, with further implementation details, benchmarks, and companion checkpoints hosted on its project GitHub repository.

Model Files

llama.cpp

LLM inference in C/C++ — https://github.com/ggml-org/llama.cpp