Input Guide

Supported model formats for `torq-compile`

The input type tells the compiler what kind of intermediate representation (IR) it’s processing, such as TOSA or Torch dialects, and selects the appropriate frontend pipeline.

Note: By default, torq-compile expects TOSA MLIR input, so you do not need to specify --iree-input-type=tosa-torq unless you want to override the default.

When compiling a model, specify the input type with --iree-input-type.

For TOSA MLIR: --iree-input-type=tosa-torq
For Torch MLIR: --iree-input-type=torch-torq

Choose the appropriate value based on your model’s origin.

For more options, run:

$ torq-compile --help

The Torq specific options can be retrieved by running:

$ torq-compile --help | grep torq

Input Structure for `iree-run-module`

This section explains how to provide inputs to the iree-run-module tool when running models. You’ll need to match the expected shape, data type, and order of the model’s inputs. Mismatches in input specifications (e.g., wrong dtype or shape) will result in execution errors. Inputs can be provided directly as literals or loaded from files like .npz or .bin.

Dtype: Match the compiled model (e.g., i8, i16, f32). Mismatched dtypes will fail.
Shapes: Use the exact input shapes expected by the model.

Literal Inputs

Pattern:

--input="<shape>x<dtype>=<value_or_list>"

Example:

$ iree-run-module --device=torq --module=model.vmfb --input="1x64x64x3xi8=0"

List syntax (comma-separated) for tensors:

$ iree-run-module --device=torq --module=model.vmfb --input="1x3xi8=[[1,2,3]]"

Multiple Inputs

Order matters: pass inputs in the same order as the model’s signature.

Example with two inputs (e.g., image tensor + scale):

$ iree-run-module --device=torq --module=model.vmfb \
  --input="1x224x224x3xi32=0" \
  --input="1xi16=1"

Feeding Input from Files

NPY (numpy arrays): Numpy npy files from numpy.save.

$ iree-run-module --device=torq --module=model.vmfb --input=@input.npy

Raw binary: Raw binary files can be read to provide buffer contents.
```
$ iree-run-module --device=torq --module=model.vmfb \
  --input="1x224x224x3xf16=@input.bin"
```
(Suppose input.bin holds float16 data for shape 1x224x224x3)

Converting Images to .npy or .bin

You can easily convert a PNG or JPEG image to .npy or .bin format using the provided tools.
Just run:

$ image_to_tensor.py --input=image.jpg -o=input.npy --format=npy

or

$ image_to_tensor.py --input=image.png -o=input.bin --format=rgb

The script will automatically detect the image type and save the tensor in the desired format.
Make sure to match the output shape and dtype with your model’s requirements.
You can also specify the resize image size using the --size option (e.g., --size=224x224).
for more options, run:
```
$ image_to_tensor.py --help
```

For more information, run:

  $ iree-run-module --help