Skip to main content

Python client for coach

Project description

Coach Python SDK

Coach is an end-to-end Image Recognition platform, we provide the tooling to do effective data collection, training, and on-device parsing of Image Recognition models.

See for more information!


Install and update using pip:

pip install coach-ml


Coach can be initialized 2 different ways. If you are only using the offline model parsing capabilities and already have a model package on disk, you can initialize like so:

coach = CoachClient()

# We already had the `flowers` model on disk, no need to authenticate:
result = coach.get_model('flowers').predict('rose.jpg')

However, in order to download your trained models, you must authenticate with your API key:

coach = CoachClient().login('myapikey')

# Now that we're authenticated, we can cache our models for future use:

# Evaluate with our cached model:
result = coach.get_model('flowers').predict('rose.jpg')

API Breakdown


Optional is_debug, if True, additional logs will be displayed

login(apiKey) -> CoachClient
Authenticates with Coach service and allows for model caching. Accepts API Key as its only parameter. Returns its own instance.

cache_model(model_name, path='.', skip_match=False, model_type='frozen') Downloads model from Coach service to disk. Specify the name of the model, and the path to store it. This will create a new directory in the specified path and store any model related documents there. By default, if a model already exists with the same version, in the same path, caching will be skipped. Set skip_match to False to override this behaviour. model_type can be one of: frozen, unity, mobile, and can be useful if you're interested in caching a specific version of your model.

get_model(path='.') -> CoachModel Loads model into memory. Specify the path of the cached models directory. Returns a CoachModel

get_model_remote(model_name, path='.') -> CoachModel Downloads and loads model into memory. Specify the path of the cached models directory. Returns a CoachModel


__init__(graph, labels, base_module)
Initializes a new instance of CoachModel, accepts a loaded tf.Graph(), array of labels, and the base_module the graph was trained off of.

predict(image, input_name="input", output_name="output") -> dict
Specify the directory of an image file or the image as a byte array. Parses the specified image into memory and runs it through the loaded model. Returns a dict of its predictions in order of confidence. If you have a pretrained frozen graph with different Tensor input/output names, you can specify them with input_name and output_name respectfully.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for coach-ml, version 0.22
Filename, size File type Python version Upload date Hashes
Filename, size coach_ml-0.22-py3-none-any.whl (5.8 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size coach-ml-0.22.tar.gz (4.7 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page