Fix potential generator exhaustion bug in model loading file detection

[rolandtannous]

PR Description:

This resolves PR #2334

Problem

In the current code, the files variable is created as a generator expression:

files = (os.path.split(x)[-1] for x in files)

Generators can only be consumed once. After iteration, they become empty. While the current code works fine (since sum() only consumes the generator once), this creates a fragile situation that could lead to bugs if future modifications use the files variable multiple times.

Example of potential issue

# This works fine
if sum(x == "adapter_config.json" or x == "config.json" for x in files) >= 2:
    both_exist = True

# But if someone later adds this, it would fail silently:
print("Files found:", list(files))  # Would print: [] (empty!)

Solution

Convert the generator expression to a list immediately:

files = list(os.path.split(x)[-1] for x in files)

This change:

• Maintains current functionality (no behavior changes)
• Makes the code more robust against future modifications
• Prevents potential silent bugs
• Has negligible performance impact (the file list is typically small)

Testing

The change maintains identical behavior to the current implementation but makes the code more maintainable and less error-prone.