kerasmlflow
Does mlflow support to save model to .h5?
I can input my Keras data (model, kappa, loss, epoch, ...) into MLflow now. My original model is H5, but the model saved in MLflow under artifacts/model/data/model/saved_model.pb is PB. Is it possible to save .h5 in MLflow?
Solution
By default mlflow save_format is tf which will allow the model to save in pb format. I made it work(changing to h5 format) using below code
mlflow.tensorflow.log_model(model, "models", keras_model_kwargs={"save_format": "h5"})
After this save format will change to h5
and the model will be saved in model.h5
Tested this in mlflow v2.3.0
- Adding visualization for every layer activation layer of Convolution Network - Keras
- KeyError: 'filename' (Pandas)
- tensorflow - how to use 16 bit precision float
- ImportError: No module named tensorflow.keras.applications
- Why is the accuracy for my Keras model always 0 when training?
- Keras model.export() fails because of NoneType shapes in model
- module 'keras.utils.generic_utils' has no attribute 'get_custom_objects' when importing segmentation_models
- Inconsistency in input shape to neural network
- How to calculate optimal batch size?
- Data Augmentation Layer in Keras Sequential Model
- How can I use a pre-trained neural network with grayscale images?
- keras lstm input shape error pleas help :(
- Visualize TFLite graph and get intermediate values of a particular node?
- How to get an Intermediate output of a keras model?
- keras.utils.get_file() throws TypeError: '<' not supported between instances of 'int' and 'NoneType'
- the get_file() function freezes my script
- Load several Images without label in keras cnn
- lstm prediction result delay phenomenon
- Merge weights of same model trained on 2 different computers using tensorflow
- Need to convert Keras model to TensorFlow.js but facing version compatibility issues between TensorFlow and Keras
- Windowing a TensorFlow dataset without losing cardinality information?
- TypeError: __init__() got an unexpected keyword argument 'name' when loading a model with Custom Layer
- Pyinstaller created exe has an AttributeError
- Python tensorflow keras error when load a json model: Could not locate class 'Sequential'
- stmemory and ltmemory in "How to build your own AlphaZero AI using Python and Keras"
- Forecast future values with LSTM in Python
- What's the difference between a Tensorflow Keras Model and Estimator?
- Difficulty with stacking MNIST and Fashion_MNIST
- Why does my pruned model have a larger file size than my initial model?
- load weights require h5py