OM Simple
Project description
Image Classification
pip install om-simple
Data Preparation
{dataset name}/
├── train/
│ ├── {class1}/
│ ├── {class2}/
│ ├── ...
└── val/
├── {class1}/
├── {class2}/
├── ...
Example Code
from om_simple.img_class_head import ImageClassification
from om_simple.img_class_model import run_train
from om_simple.tools.utils import is_blur
# Train
run_train("{dataset name}",
model_name="{model_name}", # e.g., resnet18, resnet50, vit_base_patch16_224, etc....
outdir="{output_dir}")
# Simple Classification
X = ImageClassification("epoch099.ckpt")
z = X.predict(images=["sample.jpg"])
# Blur detection
print (is_blur("sample.jpg"))
# Multi label classification
from om_simple.multi_class_model import MultiClass
X = MultiClass("model.ckpt","label.json")
z = X.predict(images=["sample.jpg"])
How to get available model_name
import timm
avail_pretrained_models = timm.list_models(pretrained=True)
print (avail_pretrained_models)
all_vit_models = timm.list_models('vit*')
print (all_vit_models)
Tensorboard
tensorboard --logdir {output_dir} --bind_all
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
om_simple-0.1.6.tar.gz
(1.4 MB
view hashes)