Title: "Need guidance: M2 Mac deployment & model files location"

[jonschull]

Environment

• MacOS on M2 Apple Silicon
• UI-TARS Desktop v0.0.6 (arm64)
• vLLM v0.7.2 installed
• Command Line Tools installed

Current Status

1. Successfully installed UI-TARS Desktop and vLLM
2. Attempted to start vLLM server with:

   python -m vllm.entrypoints.openai.api_server --served-model-name ui-tars --model <path>
   

3.Cannot proceed due to two blockers:
Unable to locate UI-TARS model files for download
Unclear if there are M2-specific deployment steps needed
Questions

1.Where can I download the UI-TARS model files for local deployment?
2.Are there any special considerations for M2 Mac deployment?
3.If local deployment isn't recommended for M2, what's the preferred alternative?
Additional Context

Reviewed Issue #64 about M2 deployment
Looking into Inference Endpoints as alternative