abhigyanpatwari/GitNexus · error

No suitable device found for embedding model

Error message

No suitable device found for embedding model

What it means

initLocalEmbedder throws this after exhausting every candidate device (e.g. webgpu/cuda/dml then cpu): each attempt failed with its own deviceError, and the last attempt's error was re-thrown before reaching this line only if it was not the final device. Reaching this throw means the devicesToTry loop somehow ended without a successful pipeline and without the last device re-throwing — an exhaustion fallback indicating no execution provider could run the embedding model.

Solutions

  1. Reinstall the runtime: `gitnexus embeddings install --force` (or a clean reinstall) to repair a corrupted onnxruntime.
  2. Run `gitnexus doctor` to check the embedding runtime and device status.
  3. Try explicitly setting GITNEXUS_EMBEDDING_DEVICE=cpu to skip GPU attempts and see the underlying CPU error.
  4. If the CPU lacks required instructions or onnxruntime cannot load at all, switch to GITNEXUS_EMBEDDING_URL HTTP embedding.
  5. Inspect the deviceError logged during the loop for the root cause of each failed device.

Example fix

// before: every device fails, root cause hidden
const embedder = await initLocalEmbedder();
// after (shell): isolate the CPU failure and repair the runtime
GITNEXUS_EMBEDDING_DEVICE=cpu node -e "...initLocalEmbedder()"  # see real deviceError
gitnexus embeddings install --force
Defensive patterns

Strategy: try-catch

Validate before calling

import { assessLocalEmbeddingRuntime } from 'gitnexus/src/core/embeddings/runtime-support.js';
if (assessLocalEmbeddingRuntime().status !== 'ready') {
  throw new Error('Embedding runtime not ready; run gitnexus doctor / embeddings install');
}

Try / catch

try {
  await initLocalEmbedder();
} catch (err) {
  if (err.message === 'No suitable device found for embedding model') {
    // reinstall runtime with --force, retry with GITNEXUS_EMBEDDING_DEVICE=cpu, or use HTTP embedding
  }
}

Prevention

When it happens

Trigger: Calling initLocalEmbedder() when pipeline creation fails on every device in devicesToTry — e.g. broken/corrupted onnxruntime-node install, GPU drivers missing while forced device was expected to fall back, or an ONNX build incompatible with the CPU.

Common situations: Damaged embedding-runtime prefix (interrupted install); containers without GPU drivers where non-CPU devices fail and CPU also fails due to a bad onnxruntime build; very old CPUs lacking required instruction sets.

Related errors

AI-assisted analysis of abhigyanpatwari/GitNexus@ac9a4e9abd (2026-09-15). Data as JSON: /api/errors/eaba5267a7c8a7fc. Report an issue: GitHub.

Appendix: source

Thrown at gitnexus/src/core/embeddings/embedding-local-init.ts:185

            const endpointHint = process.env.HF_ENDPOINT
              ? `The configured endpoint (${process.env.HF_ENDPOINT}) may be unreachable.`
              : `huggingface.co may be unreachable from your network.\n` +
                `  Set HF_ENDPOINT to a mirror and retry:\n` +
                `    HF_ENDPOINT=https://hf-mirror.com npx gitnexus analyze --embeddings\n` +
                `    (Windows: set HF_ENDPOINT=https://hf-mirror.com && npx gitnexus analyze --embeddings)`;
            throw new Error(`Failed to download embedding model: ${errMsg}\n  ${endpointHint}`);
          }
          if (isDev && (device === 'cuda' || device === 'dml')) {
            const gpuType = device === 'dml' ? 'DirectML' : 'CUDA';
            logger.info(`⚠️  ${gpuType} not available, falling back to CPU...`);
          }
          if (device === devicesToTry[devicesToTry.length - 1]) {
            throw deviceError;
          }
        }
      }

      throw new Error('No suitable device found for embedding model');
    } catch (error) {
      initPromise = null;
      embedderInstance = null;
      throw error;
    }
  })();

  return initPromise;
};

const getLocalEmbedder = (): FeatureExtractionPipeline => {
  if (!embedderInstance) {
    throw new Error('Embedder not initialized. Call initLocalEmbedder() first.');
  }
  return embedderInstance;
};

export const localEmbedBatch = async (texts: string[]): Promise<Float32Array[]> => {

View on GitHub (pinned to ac9a4e9abd)