Skip to content

Rename quantization option from --model-name to --model-id#984

Merged
zhenchaoni merged 2 commits into
mainfrom
private/zhenni/update_quant_cli_name
Jun 26, 2026
Merged

Rename quantization option from --model-name to --model-id#984
zhenchaoni merged 2 commits into
mainfrom
private/zhenni/update_quant_cli_name

Conversation

@zhenchaoni

Copy link
Copy Markdown
Member

Fix #563

This is part of consistency work. Because "model ID" is the correct term for a hugging face model identifier like microsoft/resnet-50. So I make the user facing options consistent.

In this PR, I

  • renamed winml quantize option from --model-name to --model-id
  • renamed the config json's quantize->model_name property to model_id
  • renamed the recipes

Caveat: This changes the user facing option name. I think it is fine since we are still in early releases.

About the consistency of our commands. Currently, all our commands follow the practice:

  • -m or --model can be used for either onnx model path or huggingface model id
  • --model-id is used for huggingface model id only, and only when the -m option is used for onnx model. There are cases where our command requires both an onnx path and model ID. For this case, the model ID goes to --model-id option.

@zhenchaoni zhenchaoni requested a review from a team as a code owner June 26, 2026 05:29
Comment thread src/winml/modelkit/commands/quantize.py Outdated
@zhenchaoni zhenchaoni merged commit e492ec7 into main Jun 26, 2026
9 checks passed
@zhenchaoni zhenchaoni deleted the private/zhenni/update_quant_cli_name branch June 26, 2026 07:33
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[cross-cutting] [P2] --model semantic drift across 10 commands

3 participants