Background
The model operates on some input of defined shape and returns an output of defined shape. There are two models we care about: the diffusion model which de-noises images, and a decoder model, which takes in the diffusion models output and returns a full grid. Both of these are models with different shapes
Description
Find and record, for the core model and for the decoder separately:
- Input shape: height, width, and channel count
- Output shape: the same, for every output the model returns
- Value range : what band the numbers are expected to be in
- Precision: float32 or float16
Then create an object that can hold these properties and to act as the component boundary between model input / output
Background
The model operates on some input of defined shape and returns an output of defined shape. There are two models we care about: the diffusion model which de-noises images, and a decoder model, which takes in the diffusion models output and returns a full grid. Both of these are models with different shapes
Description
Find and record, for the core model and for the decoder separately:
Then create an object that can hold these properties and to act as the component boundary between model input / output