ruvnet/ruflo · error

dropoutRate must be between 0 and 1

Error message

dropoutRate must be between 0 and 1

What it means

Thrown by AttentionCoordinator#validateConfig (v3/@claude-flow/integration/src/attention-coordinator.ts:488) when config.dropoutRate is outside [0, 1]. Dropout is a probability, so negative values or values above 1 are rejected (both boundaries inclusive — 0 and 1 are valid).

Source

Thrown at v3/@claude-flow/integration/src/attention-coordinator.ts:488

    return {
      avgLatencyMs: 0,
      throughputTps: 0,
      memoryEfficiency: 1.0,
      cacheHitRate: 0,
      totalOperations: 0,
      speedupFactor: 1.0,
    };
  }

  private validateConfig(): void {
    if (this.config.numHeads <= 0) {
      throw new Error('numHeads must be positive');
    }
    if (this.config.headDim <= 0) {
      throw new Error('headDim must be positive');
    }
    if (this.config.dropoutRate < 0 || this.config.dropoutRate > 1) {
      throw new Error('dropoutRate must be between 0 and 1');
    }
    if (this.config.flashOptLevel < 0 || this.config.flashOptLevel > 3) {
      throw new Error('flashOptLevel must be between 0 and 3');
    }
  }

  private async prewarmCache(): Promise<void> {
    // Pre-compute common attention patterns
    // This is a no-op in the simplified implementation
  }

  /**
   * Perform attention computation
   *
   * ADR-001: For sequences longer than 512 tokens, delegates to
   * agentic-flow's native Flash Attention (approximate sparse attention;
   * speedup unverified — see docs/reviews/intelligence-system-audit-2026-05-29.md).
   */

View on GitHub (pinned to fa13ee4ad6)

Solutions

1. Express dropout as a fraction in [0, 1] — e.g. 0.1 for 10%, not 10.
2. If your config stores percent, convert before constructing: dropoutRate: percent / 100.
3. Add a config lint that clamps or rejects out-of-range probabilities at load time.

Example fix

// before
new AttentionCoordinator({ dropoutRate: 10 }); // meant 10%

// after
new AttentionCoordinator({ dropoutRate: 0.1 }); // fraction

function toProbability(v: unknown, fallback = 0.1): number {
  const n = typeof v === 'number' ? v : Number(v);
  return Number.isFinite(n) && n >= 0 && n <= 1 ? n : fallback;
}
new AttentionCoordinator({ ...cfg, dropoutRate: toProbability(cfg.dropoutRate) });

function isProbability(v: unknown): v is number {
  return typeof v === 'number' && Number.isFinite(v) && v >= 0 && v <= 1;
}

try {
  coord = new AttentionCoordinator(cfg);
} catch (e) {
  if ((e as Error).message === 'dropoutRate must be between 0 and 1') {
    coord = new AttentionCoordinator({ ...cfg, dropoutRate: 0.1 });
  } else throw e;
}

When it happens

Trigger: Passing dropoutRate: 5 (a percentage instead of a fraction), -1, or NaN-sourced values; configs shared with a library that expresses dropout in percent (0–100); values parsed from strings ('0.1' works via coercion, but '1e-' style typos yield NaN which passes the range check silently only if NaN — note NaN < 0 is false and NaN > 1 is false, so NaN slips through; the thrown case is concrete out-of-range numbers).

Common situations: Porting a config from a framework using percent dropout; hand-edited YAML with a typo like 0..1; LLM-generated config with plausible-looking but wrong values.

Understand the failure class

Background: "Invalid value" and "allowed values are" config errors: what your library rejected and how to fix it — this error's family across 41 libraries.

Related errors

• flashOptLevel must be between 0 and 3
• numHeads must be positive
• headDim must be positive
• SemanticRouter requires a dimension in config
• Flash attention not initialized

AI-assisted analysis of ruvnet/ruflo@fa13ee4ad6 (2026-08-18). Data as JSON: /api/errors/360c8436a3b825e8. Report an issue: GitHub.