Skip to content
Other

PlumaDINO-vits16plus-r2 - classifying over 10,000 bird species

2 min readupdated

The PlumaDINO model is designed to recognize 10,911 bird species based on the AviList v2025b taxonomy. It takes a cropped image containing the entire bird as input and outputs a ranked list of likely biological species with their scientific and English names. The architecture is built on DINOv3 approaches, and it supports both PyTorch and Core ML for integration into native macOS applications. This solution is useful for ornithologists, wildlife app developers, and photography enthusiasts.

Key features

  • Recognition of 10,911 bird species using the AviList v2025b taxonomy.
  • Native Core ML support for running on Apple devices with macOS.
  • Utilization of the Transformers library and PyTorch architecture for CPU operation.
  • High accuracy on specialized datasets, including NABirds.

What you can do

  • Build a script or local application for automatic bird identification from photographs.
  • Integrate the model into a native iOS or macOS application for field observations.
  • Process a collection of wildlife photos in batches using Python.

Example

An inference example using PyTorch for a batch of images:

import torch
from transformers import AutoImageProcessor, AutoModelForImageClassification

processor = AutoImageProcessor.from_pretrained(".", trust_remote_code=True, local_files_only=True, use_fast=False)
model = AutoModelForImageClassification.from_pretrained(
    ".", trust_remote_code=True, local_files_only=True, dtype=torch.float32
).eval()

images = ["bird-a.jpg", "bird-b.jpg"]
inputs = processor(images, return_tensors="pt")
with torch.inference_mode():
    probabilities = model(**inputs).logits.softmax(dim=-1)
Read moreDetails
Benchmarks: on the NABirds dataset, Top-1 accuracy is 92.25%, and Top-5 is 98.57%
On the RareSpecies dataset, Top-1 reaches 79.08%
Limitations: weights and derivatives are available for non-commercial use only under the terms of the Meta DINOv3 license
Discuss Telegram 0 26

Will it run?

Runs in F32

Fits without quantisation.

  • F32weights 0.1 GB · ready files0.7 GB
  • BF16weights 0.1 GB0.6 GB
  • FP8 / INT8recommendedweights 0.0 GB0.5 GB
  • 4-bitweights 0.0 GB0.5 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 · F32 — from RTX 2060

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.

    # yxsi/PlumaDINO-vits16plus-r2
    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 yxsi/PlumaDINO-vits16plus-r2 --local-dir ~/models/PlumaDINO-vits16plus-r2
    # The model is in ~/models. How to run it is in the model card (link above).
  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

New posts about versions of PlumaDINO-vits16plus-r2 (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

yxsi/PlumaDINO-vits16plus-r2
Author
yxsi
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

260

Ask AI about this model

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

Comments