Carlosmg23/fruit-ripeness-12clases
Updated โข 106 โข 1
How to use Carlosmg23/fruit-ripeness-final with Keras:
# !pip install -U keras tensorflow huggingface_hub
# Keras needs TensorFlow installed to read "hf://" paths, so the tensorflow backend is selected here;
# "jax" and "torch" also work for computation once TensorFlow is installed.
import os
os.environ["KERAS_BACKEND"] = "tensorflow"
import keras
model = keras.saving.load_model("hf://Carlosmg23/fruit-ripeness-final")
Modelo final de la Actividad 9 del proyecto Sistema inteligente para estimar el estado de madurez y deterioro de frutas.
SparseCategoricalCrossentropy no lo soporta en esta version de Keras.| Modelo | Exactitud | Exactitud balanceada | F1 macro | QWK | Top-3 |
|---|---|---|---|---|---|
| Mayoria | 0.4131 | 0.0833 | 0.0487 | 0.0000 | 0.4713 |
| Random Forest | 0.6694 | 0.5147 | 0.5217 | 0.6546 | 0.8812 |
| Criterio 03 | 0.5958 | 0.5720 | 0.5335 | 0.5532 | nan |
| MobileNetV2 | 0.6621 | 0.6334 | 0.5940 | 0.6899 | 0.9038 |
from huggingface_hub import hf_hub_download
import tensorflow as tf
ruta = hf_hub_download(
repo_id="Carlosmg23/fruit-ripeness-final",
filename="modelo_final.keras",
)
modelo = tf.keras.models.load_model(ruta)
Para movil, usar modelo_final.tflite con TensorFlow Lite.
Rotten es deterioro, no una etapa de maduracion.CC BY 4.0. Dataset derivado de Fruit Ripeness de Muhammad's Workspace (Roboflow Universe, version 3).