Getting started
API Reference
ModleeClient
Converter
DataMetafeatures
ImageDataMetafeatures
bench_kmeans_unsupervised()
extract_features_from_model()
get_image_features()
manipulate_x_1()
manipulate_x_2()
manipulate_x_3()
manipulate_x_4()
manipulate_x_5()
pad_image_channels()
sample_dataloader()
sample_image_channels()
sample_image_from_video()
extract_loss_functions()
get_exp_loss_for_model()
exhaust_sequence_branch()
get_code_text()
get_code_text_for_model()
save_code_text_for_model()
ImageCallback
ModleeImageModel
get_cached_vars()
get_data_snapshot()
get_model()
get_runs()
run_path_exists()
Trainer
apply_discretize_to_summary()
closest_power_of_2()
convert_to_scientific()
discretize()
get_fashion_mnist()
get_model_size()
is_cacheable()
last_run_path()
quantize()
quantize_dict()
safe_mkdir()
save_run()
save_run_as_json()
typewriter_print()
uri_to_path()
Examples
Links
Log several different losses for an experiment.
Extracts unique loss function calls from ModleeModel class definition passed as string
code_str – The Python code string to analyze.
A list of unique loss function calls found in the code.
Get a list of loss functions used in a given model. NOTE Assumption that loss function is called in __init__ or training_step
module – The model to analyze.
A list of unique loss function calls found in the model.