Skip to content
Other

LocateAnything-3B

Accurate text-prompted object detection in images

2 min readupdated

This object-detection model locates specific elements in an image based on a text query. It takes an image and a text instruction, then returns the coordinates of the detected regions in a structured format. Based on a replication of a model from nvidia, it runs on the transformers library and has around 3.83 billion parameters. The tool is useful for developers and researchers who need to automatically find objects, faces, or specific details in visual media using text prompts.

Key features

  • Full compatibility with modern versions of the transformers library
  • Support for text prompts to locate arbitrary objects in an image
  • Structured coordinate output for detected regions
  • Image processing using the standard processor and model from the Hugging Face ecosystem

What you can do

  • Integrate the model into computer vision pipelines for labeling and object search
  • Automatically find specific details or parts in user-uploaded photos
  • Use it as a foundation for visual assistants and multimodal systems

Example

import argparse
from pathlib import Path

import torch
from PIL import Image
from transformers import AutoModel, AutoProcessor

repo_id = 'SD-Next/LocateAnything-3B'
processor = AutoProcessor.from_pretrained(repo_id, trust_remote_code=True)
model = AutoModel.from_pretrained(repo_id, trust_remote_code=True, torch_dtype="auto", device_map="auto").eval()

image = Image.open("image.png").convert("RGB")
messages = [{
    "role": "user",
    "content": [
        {"type": "image", "image": image},
        {"type": "text", "text": "locate hands and face"}
    ]
}]
text = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = processor(images=image, text=text, return_tensors="pt").to(model.device)

with torch.inference_mode():
    output = model.generate(**inputs, max_new_tokens=128, do_sample=False)
decoded = processor.batch_decode(output[:, inputs["input_ids"].shape[1]:], skip_special_tokens=False)
print(decoded[0])
Read moreDetails
Task: object-detection
Base model: nvidia/LocateAnything-3B
Parameters: 3.83B
Library: transformers
Discuss Telegram 0 41

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 7.1 GB · ready files8.3 GB
  • FP8 / INT8recommendedweights 3.6 GB4.4 GB
  • 4-bitweights 2.0 GB2.7 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · BF16 — from RTX 3080

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # SD-Next/LocateAnything-3B
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download SD-Next/LocateAnything-3B --local-dir ~/models/LocateAnything-3B
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to LocateAnything-3B
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

  1. Qwen2.5-3Bfamily root
  2. Qwen2.5-3B-Instruct
  3. LocateAnything-3B
  4. LocateAnything-3B

New posts about versions of Qwen2.5-3B (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 12All on Hugging Face
Fine-tunes 12All on Hugging Face
LoRAs and adapters 2All on Hugging Face
Merges 3All on Hugging Face

About the model

SD-Next/LocateAnything-3B
Author
SD-Next
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

410

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments

Other

ToonOut

ToonOut