Better timeouts

[jlamypoirier]

✨ Description

Fix: #122, #76

Set a custom timeout for slow operations (data setup, checkpoint save/load/export) using the well-hidden add_ephemeral method. Using a single "long timeout" for simplicity, the alternative would be to have 4-5 separate timeouts, which would be more powerful and actually simpler to implement but also more annoying to configure.

🔍 Type of change

Select all that apply:

• 🐛 Bug fix (non-breaking change that addresses a specific issue)
• 🚀 New feature (non-breaking change that adds functionality)
• ⚠️ Breaking change (a change that could affect existing functionality)
• 📈 Performance improvement/optimization (improves speed, memory usage, or efficiency)
• 🛠️ Code refactor (non-functional changes that improve code readability, structure, etc.)
• 📦 Dependency bump (updates dependencies, including Dockerfile or package changes)
• 📝 Documentation change (updates documentation, including new content or typo fixes)
• 🔧 Infrastructure/Build change (affects build process, CI/CD, or dependencies)

[tscholak]

why can't this be moved to a module in the tests?

[jlamypoirier]

Because it has to be run in a separate process (for distributed) which doesn't import the test file.

[tscholak]

I see, so you're calling the fast-llm CLI in a subprocess in the tests.