sgl-project/sglang · error · RuntimeError

Failed to get model info: {str(e)}

Error message

Failed to get model info: {str(e)}

What it means

get_model_info queries the serving backend's /models endpoint (base_url with /v1 stripped) and wraps any requests exception (connection error, timeout, non-2xx status) in a RuntimeError with the underlying message. It signals the SGLang image server is unreachable or erroring.

Source

Thrown at python/sglang/multimodal_gen/apps/ComfyUI_SGLDiffusion/core/server_api.py:61

        Get information about the model served by this server.

        Returns:
            Dictionary containing model information including:
            - model_path: Path to the model
            - task_type: Type of task (e.g., "T2V", "I2I")
            - pipeline_name: Name of the pipeline
            - num_gpus: Number of GPUs
            - dit_precision: DiT model precision
            - vae_precision: VAE model precision
        """
        try:
            # Remove /v1 from base_url for /models endpoint
            models_url = self.base_url.removesuffix("/v1") + "/models"
            response = requests.get(models_url, headers=self.headers, timeout=30)
            response.raise_for_status()
            return response.json()
        except requests.exceptions.RequestException as e:
            raise RuntimeError(f"Failed to get model info: {str(e)}")

    def generate_image(
        self,
        prompt: str,
        image_path: Optional[str] = None,
        mask_path: Optional[str] = None,
        size: Optional[str] = None,
        width: Optional[int] = None,
        height: Optional[int] = None,
        n: int = 1,
        negative_prompt: Optional[str] = None,
        guidance_scale: Optional[float] = None,
        num_inference_steps: Optional[int] = None,
        seed: Optional[int] = None,
        enable_teacache: bool = False,
        response_format: str = "b64_json",
        quality: Optional[str] = "auto",
        style: Optional[str] = "vivid",

View on GitHub (pinned to 0132848349)

Solutions

1. Verify the server is up: curl the /models endpoint from the same machine
2. Correct SGLDiffusionClient base_url to match the launched server host:port
3. If it's a transient startup race, wait for server readiness and retry
4. Inspect the wrapped message for the specific HTTP status or connection error and fix accordingly

Example fix

# before
client = SGLDiffusionClient(base_url="http://localhost:8000")
info = client.get_model_info()  # wrong port

# after
client = SGLDiffusionClient(base_url="http://localhost:30000/v1")
info = client.get_model_info()

import requests
def server_ready(base_url, tries=30):
    for _ in range(tries):
        try:
            requests.get(base_url.removesuffix("/v1") + "/models", timeout=5).raise_for_status()
            return True
        except requests.exceptions.RequestException:
            time.sleep(1)
    return False
assert server_ready(client.base_url)

try:
    info = client.get_model_info()
except RuntimeError as e:
    if "Failed to get model info" in str(e):
        # server not reachable: check URL / restart server, then retry once
        raise
    raise

When it happens

Trigger: Calling get_model_info() when the SGLang diffusion server is down, base_url points to the wrong host/port, TLS/proxy issues, or the server returns 4xx/5xx so raise_for_status fires.

Common situations: Wrong base_url (missing port, left /v1 off when it should be there), server not started yet, firewall/DNS issues, or the server crashed mid-request.

Understand the failure class

Background: 'Something went wrong' / 'Request failed (500)' / 'HTTP error! status: 404' — what failed HTTP requests actually mean and how to find the real cause — this error's family across 28 libraries.

Related errors

• Failed to get server info. {error_data['error']['message']}
• Failed to edit image: {str(e)}
• Failed to generate image: {str(e)}
• Lost connection to server after {consecutive_errors} consecu
• Network error after {consecutive_errors} consecutive failure

AI-assisted analysis of sgl-project/sglang@0132848349 (2026-08-28). Data as JSON: /api/errors/bf9849762f18e152. Report an issue: GitHub.