reports/bench-v003-transformer-fp32.json

{
  "checkpoint": "models/v003-transformer",
  "quantization": "FP32",
  "platform": "Windows-11-10.0.26200-SP0",
  "torch_version": "2.14.0+cpu",
  "threads": 2,
  "load_ms": 68.4641000116244,
  "disk_bytes": 657592,
  "artifact_sha256": "151628911cf54a87523499156aa01b6c3fae935855c1d9274544542ba33d8c02",
  "observed_process_rss_bytes": 373432320,
  "memory_scope": "Observed Python RSS including training-library imports and evaluation tensors, not browser or exact peak",
  "end_to_end_policy_ms": {
    "median": 3.338349924888462,
    "p95": 5.411299993284047
  },
  "neural_forward_ms": {
    "median": 2.02550005633384,
    "p95": 3.1382000306621194
  },
  "feature_encoding_ms": {
    "median": 1.3081500073894858,
    "p95": 2.363199950195849
  },
  "python_cpu_ms": {
    "median": 0.0,
    "p95": 31.25
  },
  "evaluation": {
    "validation": {
      "samples": 480,
      "action_accuracy": 1.0,
      "target_accuracy": 1.0,
      "joint_step_accuracy": 1.0,
      "action_ece": 0.0,
      "target_ece": 0.0006519494054373354,
      "candidate_recall": 1.0
    },
    "test": {
      "samples": 480,
      "action_accuracy": 1.0,
      "target_accuracy": 1.0,
      "joint_step_accuracy": 1.0,
      "action_ece": 0.0,
      "target_ece": 0.0007661207773708156,
      "candidate_recall": 1.0
    },
    "novel_wording": {
      "samples": 480,
      "action_accuracy": 0.49166667461395264,
      "target_accuracy": 0.49791666865348816,
      "joint_step_accuracy": 0.4270833432674408,
      "action_ece": 0.506439236109145,
      "target_ece": 0.47587777813896537,
      "candidate_recall": 1.0
    }
  },
  "target_vps_validated": false,
  "scope": "Synthetic single-step benchmark. Two torch threads, not two-vCPU CPU affinity or target EPYC."
}