Function resolve_attn_implementation

clis/moss_tts_app.py:147–171

(requested: str, device: torch.device, dtype: torch.dtype)

def resolve_attn_implementation(requested: str, device: torch.device, dtype: torch.dtype) -> str | None:
    requested_norm = (requested or "").strip().lower()

    if requested_norm in {"none"}:
        return None

    if requested_norm not in {"", "auto"}:
        return requested

    # Prefer FlashAttention 2 when package + device conditions are met.
    if (
        device.type == "cuda"
        and importlib.util.find_spec("flash_attn") is not None
        and dtype in {torch.float16, torch.bfloat16}
    ):
        major, _ = torch.cuda.get_device_capability(device)
        if major >= 8:
            return "flash_attention_2"

    # CUDA fallback: use PyTorch SDPA kernels.
    if device.type == "cuda":
        return "sdpa"

    # CPU fallback.
    return "eager"


def detect_text_language(text: str) -> str: