Class FakeModel

tests/test_llm_parser.py:682–689

            return list(range(6))

    class FakeModel:
        def __init__(self, device):
            self.device = device

        def generate(self, **kwargs):
            if self.device == "cuda":
                raise RuntimeError("CUDA out of memory. Tried to allocate 38.00 MiB.")
            return FakeOutputs()

    backends = {
        "cuda": {