# LocateAnything-3B by NVIDIA: Open-Weight Model Source: https://savrn.com/models/locateanything-3b Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- ## Runs On What it takes to serve LocateAnything-3B (3.8B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them. | Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits | | --- | --- | --- | --- | --- | --- | | 16-bit | 7.7 GB | 9.2 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 8-bit | 3.8 GB | 4.6 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 4-bit | 1.9 GB | 2.3 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the [SAVRN Index](https://savrn.com/ai-index/pricing/gpus), read Oct 24, 2026. [LocateAnything-3B on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/locateanything-3b/gpus) ## SAVRN's Notes on LocateAnything-3B LocateAnything is NVIDIA's visual grounding model: describe something in words and it finds it in the image, with boxes, dense detections or points. At 3.8B parameters it runs on a single mid-range GPU, and it has drawn more than 3,000 likes since March 2026. The license is listed as other, so read NVIDIA's terms before commercial use. For robotics, inspection and document work where you need a location rather than a caption, it is one of the more capable open options. SAVRN Research, 2026-09-18 ## Model Card LocateAnything is a vision-language model for fast and high-quality visual grounding, enabling precise object localization, dense detection, and point-based localization across diverse domains in both Enterprise Intelligence and Physical AI. The model adopts a generalist design, supporting tasks such as referring expression grounding, multi-object detection, GUI element grounding, and text localization, with strong performance in complex and cluttered scenes. Its core innovation, Parallel Box Decoding (PBD), predicts complete bounding box coordinates in a single parallel step rather than autoregressive token-by-token decoding, improving efficiency while preserving geometric consistency.… Excerpt from the card by NVIDIA, licensed other. ## Configuration Architecture LocateAnythingForConditionalGeneration Context length (tokens) 32,768 Layers 36 Hidden size 2,048 Feed-forward size 11,008 Attention heads 16 Key/value heads 2 Vocabulary size 152,681 Sliding window (tokens) 32,768 RoPE base 1e+06 Stored precision bfloat16 Model type locateanything ## Identity and Version Repository nvidia/LocateAnything-3B Publisher NVIDIA Task Image and text to text Modality Image and text Library transformers Parameters 3.8B parameters Languages en Revision c32291ca5e996f5a7a485845b4f57a233936bba0 First published 2026-03-02 Last updated 2026-06-12 ## Files and Weights 47 files, 7.8 GB in total. The weights are 3 files totalling 7.7 GB in bin, safetensors. Weights3 files · 7.7 GB Configuration26 files · 1.3 MB Tokenizer3 files · 5.3 MB Documentation4 files · 36.7 KB Other10 files · 128.0 MB Repository1 file · 2.0 KB Every file | File | Type | Size | SHA-256 | | --- | --- | --- | --- | | model-00001-of-00002.safetensors | Weights | 5.0 GB | 923cfc10fed1 | | model-00002-of-00002.safetensors | Weights | 2.7 GB | 3459ba101f40 | | training_args.bin | Weights | 7.3 KB | 995b5f0a2fe7 | | added_tokens.json | Configuration | 19.8 KB | — | | all_results.json | Configuration | 247 B | — | | batch_infer.py | Configuration | 5.3 KB | — | | batch_utils/__init__.py | Configuration | 375 B | — | | batch_utils/engine_hybrid.py | Configuration | 52.4 KB | — | | batch_utils/hybrid_runtime.py | Configuration | 81.1 KB | — | | chat_template.json | Configuration | 1.0 KB | — | | config.json | Configuration | 2.4 KB | — | | configuration_locateanything.py | Configuration | 5.6 KB | — | | configuration_qwen2.py | Configuration | 7.0 KB | — | | generate_utils.py | Configuration | 18.7 KB | — | | generation_config.json | Configuration | 121 B | — | | image_processing_locateanything.py | Configuration | 4.9 KB | — | | kernel_utils/__init__.py | Configuration | 187 B | — | | kernel_utils/range_attention.py | Configuration | 14.4 KB | — | | mask_magi_utils.py | Configuration | 3.8 KB | — | | mask_sdpa_utils.py | Configuration | 9.2 KB | — | | model.safetensors.index.json | Configuration | 70.7 KB | — | | modeling_locateanything.py | Configuration | 22.2 KB | — | | modeling_qwen2.py | Configuration | 78.5 KB | — | | modeling_vit.py | Configuration | 21.3 KB | — | | preprocessor_config.json | Configuration | 480 B | — | | processing_locateanything.py | Configuration | 28.4 KB | — | | processor_config.json | Configuration | 493 B | — | | special_tokens_map.json | Configuration | 13.9 KB | — | | trainer_state.json | Configuration | 806.4 KB | — | | LICENSE | Documentation | 4.1 KB | — | | README.md | Documentation | 27.2 KB | — | | batch_utils/README.md | Documentation | 2.1 KB | — | | kernel_utils/README.md | Documentation | 3.4 KB | — | | assets/coco_lvis.png | Other | 235.3 KB | 8b75aa2122f9 | | assets/decoding_demo.mp4 | Other | 25.0 MB | 3f4083674bac | | assets/demo.mp4 | Other | 101.6 MB | 56e5c216d267 | | assets/dense_object_detection.png | Other | 161.3 KB | 5a656f3ec1d1 | | assets/layout_ocr.png | Other | 103.9 KB | ecccc9ebcdf4 | | assets/pointing.png | Other | 93.0 KB | — | | assets/referring.png | Other | 150.5 KB | 1a814ab0d925 | | assets/sspro.png | Other | 162.9 KB | 485bb3f9f49f | | assets/teaser.jpg | Other | 580.4 KB | b5af7f67594c | | pyproject.toml | Other | 392 B | — | | .gitattributes | Repository | 2.0 KB | — | | merges.txt | Tokenizer | 1.7 MB | — | | tokenizer_config.json | Tokenizer | 195.3 KB | — | | vocab.json | Tokenizer | 3.4 MB | — | ## License and Download License other Access Open weights, no gate Download size 7.7 GB [Download from NVIDIA](https://huggingface.co/nvidia/LocateAnything-3B) Released by NVIDIA through its official repository on Hugging Face. [Read the license](https://huggingface.co/nvidia/LocateAnything-3B/blob/main/LICENSE). ## Built From - Derived from [Qwen/Qwen2.5-3B-Instruct](https://savrn.com/models/qwen2-5-3b-instruct) - Described by arXiv:1405.0312 - Described by arXiv:1908.03195 - Described by arXiv:2109.10852 - Described by [arXiv:2303.05499](https://savrn.com/papers/grounding-dino-marrying-dino-with-grounded-pre-training-for-open-set-object-detection) - Described by arXiv:2504.07491 - Described by arXiv:2504.07981 - Described by arXiv:2510.12798 - Described by arXiv:2605.27365 ## Memory Requirements | Precision | Weights in memory | | --- | --- | | As published | 7.7 GB | | 16-bit | 7.7 GB | | 8-bit | 3.8 GB | | 4-bit | 1.9 GB | Weights only, from the published parameter count; the key-value cache and runtime add to this. ## Built on This Model - Quantized from[locate-anything.cpp-gguf](https://savrn.com/models/locate-anything-cpp-gguf) - Derived from[locate-anything.cpp-gguf](https://savrn.com/models/locate-anything-cpp-gguf) - Quantized from[RK182X-VLM-LocateAnything-3B](https://savrn.com/models/rk182x-vlm-locateanything-3b) - Derived from[RK182X-VLM-LocateAnything-3B](https://savrn.com/models/rk182x-vlm-locateanything-3b) ## Compare LocateAnything-3B - [LocateAnything-3B vs Qwen2.5-VL-3B-Instruct](https://savrn.com/models/compare/locateanything-3b-vs-qwen2-5-vl-3b-instruct) ## Questions About LocateAnything-3B ### How much GPU memory does LocateAnything-3B need? About 9.2 GB at 16-bit and 2.3 GB at 4-bit: the weights (3.8B parameters) plus a working margin. A long context needs more. ### What is the cheapest GPU to run LocateAnything-3B on? At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists. ### What license is LocateAnything-3B released under? other, as its publisher declares it. Read the license text before commercial use. ### What is LocateAnything-3B's context length? 32,768 tokens, from the maximum position embeddings in its published configuration. ## Similar Models Model · Image and text to text ### [Qwen2.5-VL-3B-Instruct](https://savrn.com/models/qwen2-5-vl-3b-instruct) [Qwen](https://savrn.com/model-publishers/qwen) licensename: qwen-research licenselink: https://huggingface.co/Qwen/Qwen2.5-VL-3B-Instruct/blob/main/LICENSE pipelinetag: image-text-to-text - multimodal libraryname: transformers In the past five months since Qwen2-VL’s release, numerous developers have built new models on the Qwen2-VL vision-language models, providing us with valuable feedback. During this period, we focused on building more useful vision-language models. Today, we are excited to introduce the latest addition to the Qwen family: Qwen2.5-VL. Understanding long videos and capturing events: Qwen2.5-VL can comprehend videos of over 1 hour, and this time it has a new ability of cpaturing event by pinpointing the relevant video… Open weights 3.8B parameters 128,000 tokens transformers [View model](https://savrn.com/models/qwen2-5-vl-3b-instruct) Model · Image and text to text ### [Nanonets-OCR2-3B](https://savrn.com/models/nanonets-ocr2-3b) [Nanonets](https://savrn.com/model-publishers/nanonets) Nanonets-OCR2: A model for transforming documents into structured markdown with intelligent content recognition and semantic tagging Nanonets-OCR2 by Nanonets is a family of powerful, state-of-the-art image-to-markdown OCR models that go far beyond traditional text extraction. It transforms documents into structured markdown with intelligent content recognition and semantic tagging, making it ideal for downstream processing by Large Language Models (LLMs). Nanonets-OCR2 is packed with features designed to handle complex documents with ease: 1. Start the vLLM server. Check out Docstrange for more details. 1. Increasing the image resolution will improve model's performance. 2. For complex… Open weights 3.8B parameters 128,000 tokens transformers [View model](https://savrn.com/models/nanonets-ocr2-3b) Model · Image and text to text ### [swiftvln-satnav-3b-1ep-f32s4-overlap0-pf-h8-pool-s2-noembed](https://savrn.com/models/swiftvln-satnav-3b-1ep-f32s4-overlap0-pf-h8-pool-s2-noembed) [Jiajun Jiang](https://savrn.com/model-publishers/eku127) SwiftVLN: training and evaluation code for these checkpoints. - SatNav: satellite-image navigation environments, datasets, and evaluation tools. This checkpoint is designed for SwiftVLN on SatNav, the continuous-state vision-and-language navigation benchmark over satellite imagery. It predicts short navigation-action sequences from an instruction, the current RGB window, and sampled visual memory. - swiftvln-satnav: SwiftVLN trained and evaluated on SatNav - 3b: Qwen2.5-VL 3B backbone - 1ep: trained for one epoch - f32s4: uses a 32-frame window and predicts four actions - overlap0: uses non-overlapping training windows - pf-h8: uses per-frame memory with up to eight history frames… Open weights 3.8B parameters 128,000 tokens transformers [View model](https://savrn.com/models/swiftvln-satnav-3b-1ep-f32s4-overlap0-pf-h8-pool-s2-noembed) Model · Image and text to text ### [blip2-opt-2.7b](https://savrn.com/models/blip2-opt-2-7b) [Salesforce AI Research](https://savrn.com/model-publishers/salesforce) BLIP-2 model, leveraging OPT-2.7b (a large language model with 2.7 billion parameters). It was introduced in the paper BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models by Li et al. and first released in this repository. Disclaimer: The team releasing BLIP-2 did not write a model card for this model so this model card has been written by the Hugging Face team. BLIP-2 consists of 3 models: a CLIP-like image encoder, a Querying Transformer (Q-Former) and a large language model. The authors initialize the weights of the image encoder and large language model from pre-trained checkpoints and keep them frozen while training the Querying… Open weights mit 3.7B parameters transformers [View model](https://savrn.com/models/blip2-opt-2-7b) Model · Image and text to text ### [blip2-flan-t5-xl](https://savrn.com/models/blip2-flan-t5-xl) [Salesforce AI Research](https://savrn.com/model-publishers/salesforce) BLIP-2 model, leveraging Flan T5-xl (a large language model). It was introduced in the paper BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models by Li et al. and first released in this repository. Disclaimer: The team releasing BLIP-2 did not write a model card for this model so this model card has been written by the Hugging Face team. BLIP-2 consists of 3 models: a CLIP-like image encoder, a Querying Transformer (Q-Former) and a large language model. The authors initialize the weights of the image encoder and large language model from pre-trained checkpoints and keep them frozen while training the Querying Transformer, which is a… Open weights mit 3.9B parameters transformers [View model](https://savrn.com/models/blip2-flan-t5-xl) Model · Image and text to text ### [Rex-Omni](https://savrn.com/models/rex-omni) [IDEA-Research](https://savrn.com/model-publishers/idea-research) This model is Rex-Omni, a 3B-parameter Multimodal Large Language Model (MLLM) presented in the paper "Detect Anything via Next Point Prediction". It is compatible with the Hugging Face transformers library and is licensed under the IDEA License 1.0. src="https://img.shields.io/badge/RexOmni-Website-BADFDB?style=flat-square&logo=deno&logoColor=violet&color=BADFDB" alt="RexThinker Website" src="https://img.shields.io/badge/RexOmni-Paper-Red%25red?logo=arxiv&logoColor=red&color=yellow" alt="RexThinker Paper on arXiv" src="https://img.shields.io/badge/RexOmni-Weight-orange?logo=huggingface&logoColor=yellow" alt="RexThinker weight on Hugging Face"… Open weights other 4.1B parameters 128,000 tokens transformers [View model](https://savrn.com/models/rex-omni) ## NVIDIA [All models and datasets](https://savrn.com/model-publishers/nvidia) ## Versions - [c32291ca5e99](https://savrn.com/models/locateanything-3b/versions/c32291ca5e99) · current 2026-09-18 ## Explore More - [All image and text to text models](https://savrn.com/models/tasks/image-and-text-to-text) - [Model comparisons](https://savrn.com/models/comparisons) - [The model directory](https://savrn.com/models) - [Open model prices by host](https://savrn.com/ai-index/pricing/open-models) ## Source - Repository metadata, read 2026-10-20. - [NVIDIA website](https://build.nvidia.com/models) - [Hugging Face record](https://huggingface.co/nvidia/LocateAnything-3B) - [How the hub is built](https://savrn.com/model-hub/methodology)