blob: 4bdf6ae717f5033b0500a92d3b72e3e0809e6b55 [file] [log] [blame] [view]
This directory should be the only one in IREE that pulls a dependency on `tf`
dialect and related dialects (with the exception of a small selection of "safe"
XLA IR).
# Tools
## Development Tools
* `iree-tf-opt` : MLIR Opt tool with TensorFlow and IREE passes/dialects
linked in
* `iree-tf-translate` : Equivalent to `mlir-tf-translate` tool in TensorFlow,
with IREE passes/dialects linked in
## Production Tools
### iree-import-tf
`iree-import-tf` provides a single entry-point for compiling TensorFlow saved
models to "IREE Input Dialects" that can be fed to `iree-translate` or
`iree-opt` and operated on further.
#### Usage
```shell
iree-import-tf /path/to/saved_model_v2
# Optional args: --tf-savedmodel-exported-names=subset,of,exported,names
iree-import-tf /path/to/saved_model_v1 --tf-import-type=savedmodel_v1
# Optional args:
# --tf-savedmodel-exported-names=subset,of,exported,names
# --tf-savedmodel-tags=serving
```
# Testing
```shell
$ bazel test :saved_model_adopt_exports
```
This will capture the output and pass it through FileCheck and report pass/fail,
along with a hopefully informative description of what failed.
## Debugging failures
During development, it can be useful to just see the raw output directly.
To see the raw output of the MLIR import and conversion process:
```shell
$ bazel run :saved_model_adopt_exports -- --disable_filecheck
```
Look for the `RUN_TEST: <test_name>` and `FINISH_TEST: <test_name>` lines to
narrow in on the test that interests you.