System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow.js): no
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): macOS 15.6.1 (24G90)
• Mobile device (e.g. iPhone 8, Pixel 2, Samsung Galaxy) if the issue happens on mobile device: n/a
• TensorFlow.js installed from (npm or script link): pip
• TensorFlow.js version (use command below): tensorflowjs 4.22.0
• Browser version: n/a
• Tensorflow.js Converter Version:

tensorflowjs 4.22.0 Dependency versions: keras 2.19.0 tensorflow 2.19.0 

Describe the current behavior

When running tensorflowjs_converter with the --quantize_uint8 flag, the CLI command incorrectly says that its missing an output_path. The error disappears if the quantize_uint8 flag is not passed. See below for repro.

Describe the expected behavior

That error should not appear, since in the --help documentation, it states that --quantize_uint8 can be passed with no other parameter:

 --quantize_uint8 [QUANTIZE_UINT8] Comma separated list of node names to apply 1-byte affine quantization. You can also use wildcard symbol (*) to apply quantization to multiple nodes (e.g., conv/*/weights). When the flag is provided without any nodes the default behavior will match all nodes. 

i.e. QUANTIZE_UINT8 is optional.

Standalone code to reproduce the issue

Running:

tensorflowjs_converter --input_format=keras --output_format=tfjs_layers_model --quantize_uint8 /path/to/model.h5 /path/to/output

I get the following error:

Traceback (most recent call last): File ".../venv/bin/tensorflowjs_converter", line 8, in <module> sys.exit(pip_main()) ^^^^^^^^^^ File ".../venv/lib/python3.12/site-packages/tensorflowjs/converters/converter.py", line 959, in pip_main main([' '.join(sys.argv[1:])]) File ".../venv/lib/python3.12/site-packages/tensorflowjs/converters/converter.py", line 963, in main convert(argv[0].split(' ')) File ".../venv/lib/python3.12/site-packages/tensorflowjs/converters/converter.py", line 897, in convert raise ValueError( ValueError: Missing output_path argument. For usage, use the --help flag. 

However, if I remove the --quantize_uint8 flag, this error does not appear and the model is correctly converted.