API Reference¶
This section documents the Tinify Python API.
Module Overview¶
| Module | Description |
|---|---|
tinify.models |
Compression model architectures |
tinify.entropy_models |
Entropy bottleneck and hyperprior models |
tinify.latent_codecs |
Latent space encoding/decoding |
tinify.layers |
Neural network layers (GDN, attention, etc.) |
tinify.losses |
Rate-distortion loss functions |
| Metrics | Quality and performance metrics |
tinify.datasets |
Dataset loading utilities |
tinify.transforms |
Data transformations |
tinify.ops |
Custom operations |
Top-Level Functions¶
available_entropy_coders
¶
get_entropy_coder
¶
set_entropy_coder
¶
Specifies the default entropy coder used to encode the bit-streams.
Use :mod:available_entropy_coders to list the possible values.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
entropy_coder
|
string
|
Name of the entropy coder |
required |