Accessing ObjectStore Keras Model Outside Research

Hi Chase,

It seems like you're having trouble loading your Keras model from the ObjectStore. Here are the steps you should follow to load your model:

1. First, you need to check if the model key exists in the ObjectStore. You can do this by calling the ContainsKey method with the model key. This method returns a boolean that represents if the model_key is in the ObjectStore. If the ObjectStore does not contain the model_key, you should save the model using the model_key before you proceed.

if self.ObjectStore.ContainsKey(model_key):

1. If the model key exists, you can get the file path to the saved model by calling the GetFilePath method with the model key.

file_name = self.ObjectStore.GetFilePath(model_key)

1. Once you have the file path, you can load the model by calling the load_model method with the file path.

self.model = load_model(file_name)

Please note that the file path you get from the GetFilePath method is the path where the model is stored in the ObjectStore, not the local file system. The model should be saved to the ObjectStore in the research environment before you can load it in the algorithm.

If you're still having trouble, it might be helpful to print out the file path you get from the GetFilePath method to make sure it's correct.

I hope this helps! Let me know if you have any other questions.

Sources:

• Keras Load Models
• Stable Baselines Load Models
• Writing Algorithms with Keras

Hi,

I have tried the above and I get an error stating the model is not an attribute as the above is not actually locating my model file from its key. However, I can find the model in the Organization/Object Store tabs. 

Hi Chase,

I'm sorry to hear that you're still having trouble. If you're sure that the model is in the Object Store (as you can see it in the Organization/Object Store tabs), then the issue might be with the key you're using to access it.

The key you use to retrieve the model from the Object Store should be the same as the one you used to save it. If you're using a different key or if there's a typo in the key, the ContainsKey method will return False and you won't be able to get the file path to the model.

Here's how you can check if a key exists in the Object Store and get the file path to the model:

model_key = "your_model_key"  # replace with your actual key

if qb.ObjectStore.ContainsKey(model_key):
    file_name = qb.ObjectStore.GetFilePath(model_key)
    print(f"File path: {file_name}")
else:
    print(f"Key {model_key} does not exist in the Object Store.")

If the key does not exist, you'll need to check the key you're using and make sure it's the same as the one you used to save the model.

Once you have the correct file path, you can load the model with Keras' load_model function:

from keras.models import load_model

loaded_model = load_model(file_name)

I hope this helps! Let me know if you have any other questions.

Sources:

• Keras Load Models
• Object Store

Hi Chase Joyal 

We have fixed the issue with the Object Store. Please retry and let us know if it happens again.

We apologize for the inconvenience. 

Best regards,
Alex