onnx_diagnostic.tasks.text_to_image

onnx_diagnostic.tasks.text_to_image.get_inputs(model: Module, config: Any | None, batch_size: int, sequence_length: int, cache_length: int, in_channels: int, sample_size: int, cross_attention_dim: int, add_second_input: int = 1, **kwargs)

Generates inputs for task text-to-image. Example:

sample:T10s2x4x96x96[-3.7734375,4.359375:A-0.043463995395642184]
timestep:T7s=101
encoder_hidden_states:T10s2x77x1024[-6.58203125,13.0234375:A-0.16780663634440257]

onnx_diagnostic.tasks.text_to_image.random_input_kwargs(config: Any) → Tuple[Dict[str, Any], Callable]

Inputs kwargs.

If the configuration is None, the function selects typical dimensions.

onnx_diagnostic.tasks.text_to_image.reduce_model_config(config: Any) → Dict[str, Any]

Reduces a model size.