zeromodels / owlv2-large-patch14-ensemble

huggingface.co
Total runs: 73
24-hour runs: 1
7-day runs: -185
30-day runs: -214
Model's Last Updated: August 27 2026
zero-shot-object-detection

Introduction of owlv2-large-patch14-ensemble

Model Details of owlv2-large-patch14-ensemble

See our collection for all versions of OWLv2.

Run OWLv2 with Keras 3: JAX, PyTorch, or TensorFlow

GitHub Docs Collection

zeromodels/owlv2-large-patch14-ensemble

Paper: Scaling Open-Vocabulary Object Detection (arXiv:2306.09683) · HF Papers

OWLv2 keeps OWL-ViT's dual-tower skeleton and per-patch detection head, and scales it with self-training on web image-text pairs. It adds an objectness head (a learned is-this-patch-an-object score) and pads images to a square before resizing, which matters for post-processing target sizes.

For more details on the model, please go to Google's original model card .

Pure- Keras 3 conversion of google/owlv2-large-patch14-ensemble for zeromodels . One implementation runs unmodified on TensorFlow / Torch / JAX .

This is an open-vocabulary object detection checkpoint ( Owlv2Detect ): pass free-text prompts at inference time.

✨ Quick start
import os
os.environ["KERAS_BACKEND"] = "torch"  # or "jax" / "tensorflow"

from PIL import Image
from zeromodels.models.owlv2 import (
    Owlv2Detect,
    Owlv2Processor,
    Owlv2ImageProcessor,
)

model = Owlv2Detect.from_weights("zeromodels/owlv2-large-patch14-ensemble")
processor = Owlv2Processor.from_weights("zeromodels/owlv2-large-patch14-ensemble")
image_processor = Owlv2ImageProcessor.from_weights("zeromodels/owlv2-large-patch14-ensemble")

image = Image.open("your_image.jpg").convert("RGB")
prompts = ["a photo of a mug", "a photo of a knife"]
inputs = processor(text=[prompts], images=image)
output = model(
    {
        "input_ids": inputs["input_ids"],
        "pixel_values": inputs["pixel_values"],
    }
)
results = image_processor.post_process_object_detection(
    output,
    threshold=0.1,
    target_sizes=[(image.height, image.width)],
    text_labels=[prompts],
)[0]
for score, name, box in zip(
    results["scores"], results["text_labels"], results["boxes"]
):
    print(f"{name}: {float(score):.3f} {box}")

Load any OWL-ViT / OWLv2 variant the same way with from_weights("zeromodels/<variant>") (use Owlv2Detect for this repo):

Variant Hub Family
owlvit-base-patch32 zeromodels/owlvit-base-patch32 OWL-ViT
owlvit-base-patch16 zeromodels/owlvit-base-patch16 OWL-ViT
owlvit-large-patch14 zeromodels/owlvit-large-patch14 OWL-ViT
owlv2-base-patch16 zeromodels/owlv2-base-patch16 OWLv2
owlv2-base-patch16-ensemble zeromodels/owlv2-base-patch16-ensemble OWLv2
owlv2-base-patch16-finetuned zeromodels/owlv2-base-patch16-finetuned OWLv2
owlv2-large-patch14 zeromodels/owlv2-large-patch14 OWLv2
owlv2-large-patch14-ensemble zeromodels/owlv2-large-patch14-ensemble OWLv2
owlv2-large-patch14-finetuned zeromodels/owlv2-large-patch14-finetuned OWLv2
Tips
  • Set KERAS_BACKEND before importing Keras / zeromodels.
  • Prefer Processor.from_weights(...) so image size matches the variant.
  • Open-vocab thresholds are often much lower than closed-set detectors (try 0.1 ).
  • OWLv2 pads to square before resize; pass the original (height, width) as target_sizes carefully (see the OWLv2 docs for the padding trap).
  • See OWLv2 docs and Loading Weights .
  • Community / upstream safetensors still work via the hf: prefix, e.g. Owlv2Detect.from_weights("hf:google/owlv2-large-patch14-ensemble") .
Special Thanks

A huge thank you to the Google OWLv2 authors for creating and releasing these models.

License: Apache 2.0.

Runs of zeromodels owlv2-large-patch14-ensemble on huggingface.co

73
Total runs
1
24-hour runs
-1
3-day runs
-185
7-day runs
-214
30-day runs

More Information About owlv2-large-patch14-ensemble huggingface.co Model

More owlv2-large-patch14-ensemble license Visit here:

https://choosealicense.com/licenses/apache-2.0

owlv2-large-patch14-ensemble huggingface.co

owlv2-large-patch14-ensemble huggingface.co is an AI model on huggingface.co that provides owlv2-large-patch14-ensemble's model effect (), which can be used instantly with this zeromodels owlv2-large-patch14-ensemble model. huggingface.co supports a free trial of the owlv2-large-patch14-ensemble model, and also provides paid use of the owlv2-large-patch14-ensemble. Support call owlv2-large-patch14-ensemble model through api, including Node.js, Python, http.

owlv2-large-patch14-ensemble huggingface.co Url

https://huggingface.co/zeromodels/owlv2-large-patch14-ensemble

zeromodels owlv2-large-patch14-ensemble online free

owlv2-large-patch14-ensemble huggingface.co is an online trial and call api platform, which integrates owlv2-large-patch14-ensemble's modeling effects, including api services, and provides a free online trial of owlv2-large-patch14-ensemble, you can try owlv2-large-patch14-ensemble online for free by clicking the link below.

zeromodels owlv2-large-patch14-ensemble online free url in huggingface.co:

https://huggingface.co/zeromodels/owlv2-large-patch14-ensemble

owlv2-large-patch14-ensemble install

owlv2-large-patch14-ensemble is an open source model from GitHub that offers a free installation service, and any user can find owlv2-large-patch14-ensemble on GitHub to install. At the same time, huggingface.co provides the effect of owlv2-large-patch14-ensemble install, users can directly use owlv2-large-patch14-ensemble installed effect in huggingface.co for debugging and trial. It also supports api for free installation.

owlv2-large-patch14-ensemble install url in huggingface.co:

https://huggingface.co/zeromodels/owlv2-large-patch14-ensemble

Url of owlv2-large-patch14-ensemble

owlv2-large-patch14-ensemble huggingface.co Url

Provider of owlv2-large-patch14-ensemble huggingface.co

zeromodels
ORGANIZATIONS

Other API from zeromodels