Add llama-pull binary for downloading models from HuggingFace and Docker Registry

[Copilot]

This PR adds a new llama-pull command-line tool that provides a dedicated interface for downloading AI models from HuggingFace and Docker Registry repositories.

Features

The new tool supports downloading models from two sources:

# Download from HuggingFace
llama-pull -hf bartowski/Llama-3.2-1B-Instruct-GGUF:Q4_K_M

# Download from Docker Registry  
llama-pull -dr gemma3

Implementation

The tool leverages the existing robust download infrastructure from common/arg.cpp, ensuring:

• Consistent caching behavior with other llama.cpp tools
• Support for authentication tokens and offline mode
• Proper error handling and progress reporting
• Integration with the existing build system

Key Benefits

• Dedicated download tool: Provides a focused interface for model acquisition without requiring model loading
• Minimal implementation: Reuses existing download logic, reducing code duplication and maintenance burden
• Consistent UX: Uses the same argument patterns and caching as other llama.cpp tools
• Multiple sources: Supports both HuggingFace and Docker Registry in a single tool

Usage Examples

# Basic HuggingFace download
llama-pull -hf microsoft/DialoGPT-medium

# Docker with specific quantization
llama-pull -dr ai/gemma3:Q4_K_M

# Use cached models only
llama-pull -dr gemma3 --offline

# Private HuggingFace repo
llama-pull -hf private/model --hf-token $HF_TOKEN

Downloaded models are stored in the standard llama.cpp cache directory (~/.cache/llama.cpp/) and can be immediately used with other tools like llama-cli, llama-server, etc.

Files Added

• tools/llama-pull/llama-pull.cpp - Main implementation
• tools/llama-pull/CMakeLists.txt - Build configuration
• tools/llama-pull/README.md - Documentation
• Updated tools/CMakeLists.txt to include the new tool

The implementation is minimal (~100 lines) and integrates seamlessly with the existing codebase.

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

• huggingface.co
  • Triggering command: ./build/bin/llama-pull -hf bartowski/Llama-3.2-1B-Instruct-GGUF:Q4_K_M (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.