oobabooga/textgen · error · ImportError

--ik requires the ik_llama_cpp_binaries package. Install it

Error message

--ik requires the ik_llama_cpp_binaries package. Install it with: pip install <ik_llama_cpp_binaries wheel URL>

What it means

Error "--ik requires the ik_llama_cpp_binaries package. Install it with: pip install <ik_llama_cpp_binaries wheel URL>" thrown in oobabooga/textgen.

Source

Thrown at modules/llama_cpp_server.py:423

        """Find an available port, preferring main port + 5."""
        preferred_port = shared.args.api_port + 5
        if self._is_port_available(preferred_port):
            return preferred_port

        # Fall back to OS-assigned random port
        with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
            s.bind(('', 0))
            return s.getsockname()[1]

    def _start_server(self):
        """Start the llama.cpp server and wait until it's ready."""
        # Determine the server path
        if self.server_path is None:
            if shared.args.ik:
                try:
                    import ik_llama_cpp_binaries
                except ImportError:
                    raise ImportError("--ik requires the ik_llama_cpp_binaries package. Install it with: pip install <ik_llama_cpp_binaries wheel URL>")

                self.server_path = ik_llama_cpp_binaries.get_binary_path()
            else:
                import llama_cpp_binaries
                self.server_path = llama_cpp_binaries.get_binary_path()

        # Build the command
        cmd = [
            self.server_path,
            "--model", self.model_path,
            "--batch-size", str(shared.args.batch_size),
            "--ubatch-size", str(shared.args.ubatch_size),
            "--port", str(self.port),
            "--no-webui",
            "--flash-attn", "on",
        ]

        if shared.args.ctx_size < 0:

View on GitHub (pinned to ed888c71f2)

Solutions

1. Install the ik_llama_cpp_binaries package with: pip install <ik_llama_cpp_binaries wheel URL>.
2. Remove the --ik flag to use the standard llama.cpp loader instead.

When it happens

Trigger: Raised at startup/model-load when the --ik flag (ik_llama.cpp backend) is used but the ik_llama_cpp_binaries package is not installed. Install it with: pip install <ik_llama_cpp_binaries wheel URL>.

Common situations: See trigger scenarios.

AI-assisted analysis of oobabooga/textgen@ed888c71f2 (2026-08-15). Data as JSON: /api/errors/2ffe662b3ea23ee7. Report an issue: GitHub.