Skip to content
Other

Plant Leaf Disease Detection Using CNN

единая CNN для диагностики листьев на edge

2 min readupdated

Это image-classification модель от Shubham Kumar (Shubham12864) для определения болезней листьев растений. В основе — ConvNeXt-V2 backbone с GRN и единый end-to-end пайплайн без внешнего детектора. Доступна в PyTorch, ONNX FP32 и edge-INT8 ONNX; размер FP32 — около 107.8–107.9 MB, INT8 — 27.7 MB.

Ключевые особенности

  • Единая дифференцируемая CNN вместо связки YOLO-кроппинга и отдельного классификатора.
  • Differentiable Spatial Saliency Gating подавляет до 90% нефолийного фона — почвы, рук, горшков, неба — без внешнего bounding-box детектора.
  • Иерархическая таксономическая обусловленность: прогноз болезни зависит от вероятностей семейства культур; биологическая согласованность — 93.83%.
  • Split Conformal Prediction Safety Guard даёт гарантированные наборы предсказаний с эмпирическим покрытием 96.61% при 95% confidence.
  • Три формата: PyTorch .pt, ONNX FP32 и edge-INT8 ONNX; INT8 уменьшает размер на 74.3%.
  • Два выхода: семейство культур (k=14, 91.04% Acc) и состояние/патология (k=38, 85.59% Acc).

Пример

import numpy as np
from PIL import Image
import onnxruntime as ort
session = ort.InferenceSession("unified_gated_plant_cnn_int8.onnx", providers=["CPUExecutionProvider"])
img = Image.open("leaf_sample.jpg").convert("RGB").resize((224, 224))
arr = np.array(img, dtype=np.float32) / 255.0
mean = np.array([0.485, 0.456, 0.406], dtype=np.float32)
std = np.array([0.229, 0.224, 0.225], dtype=np.float32)
ПодробнееБенчмарк на CPU, 50 итераций: PyTorch FP32 — 73.7 ms и 13.6 FPS; ONNX FP32 — 95.7 ms и 10.5 FPS; INT8 — 293.7 ms и 3.4 FPS.
Размер: 107.8 MB (PyTorch FP32), 107.9 MB (ONNX FP32), 27.7 MB (INT8), сокращение на 74.3%.
Численное соответствие PyTorch FP32 и ONNX FP32: максимальная абсолютная ошибка < 3.58×10⁻⁶.
Ограничение: INT8-вариант заметно медленнее на CPU, несмотря на меньший размер.
Discuss Telegram 1

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.1 GB · ready files0.6 GB
  • FP8 / INT8recommendedweights 0.1 GB0.6 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 · BF16 — 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.

    # Shubham12864/plant-leaf-disease-detection-using-cnn
    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 Shubham12864/plant-leaf-disease-detection-using-cnn --local-dir ~/models/plant-leaf-disease-detection-using-cnn
    # 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 plant-leaf-disease-detection-using-cnn (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

Shubham12864/plant-leaf-disease-detection-using-cnn
Author
Shubham12864
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

01

Ask AI about this model

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

Comments