bug: Could not able to run models since 172

[louis-jan]

Cortex version

172-176

Describe the Bug

Model run is broken at 172; I was unable to start HF model models.
It worked fine before.

https://huggingface.co/bartowski/Llama-3.2-1B-Instruct-GGUF/resolve/main/Llama-3.2-1B-Instruct-Q8_0.gguf

Steps to Reproduce

1. Update to version 172 or the latest version 176.
2. Clean.
3. Pull a HF model
4. Run
5. See errors.

Screenshots / Logs

What is your OS?

• MacOS
• Windows
• Linux

What engine are you running?

• cortex.llamacpp (default)
• cortex.tensorrt-llm (Nvidia GPUs)
• cortex.onnx (NPUs, DirectML)

[nguyenhoangthuan99]

This issue has been resolved by this pr #1507

[gabrielle-ong]

@nguyenhoangthuan99 / @vansangpfiev for v183,

• I'm getting the same error as Louis' first picture
• For huggingface models only (cortexso models are ok)
• cortex run <Model ID> is successful but cortex run <model alias> throws the uncaught exception YAML::BadSubscript

cortex-cli.log
cortex.log

[vansangpfiev]

Hi @gabrielle-ong, we will deprecate model_alias soon #1410 (comment). So can we mark this bug as won't fix?

[gabrielle-ong]

True agree, thanks @vansangpfiev
This will be handled as part of #1512 next week

• deprecating model_alias
• using regex for cortex run, so this error should not occur