{
  "artifacts": [
    {
      "file": "6df59f7a42d94008dfb39940e27fa82425c8e9a7b8a3ac6b2c8ef1e7d705ff68.tflite",
      "sha256": "3ea55a43a61b8ca57b2663a9e0d33f92cde669fe0bbd1fcbf5daa1570a6cd54e",
      "size_mb": 8.818
    }
  ],
  "benchmarks": [],
  "browser": {
    "backends": [
      {
        "backend": "wasm_xnnpack",
        "date": "2026-08-11",
        "env": {
          "browser": "chromium",
          "browser_version": "151.0.7922.34",
          "headless": true,
          "jspi": true,
          "litertjs_core_version": "2.5.3",
          "machine_label": "mac-studio-m4-max",
          "os": "macOS",
          "os_version": "27.0.0",
          "webgpu_adapter": {
            "architecture": "metal-3",
            "description": "",
            "device": "",
            "vendor": "apple"
          }
        },
        "full_delegation": null,
        "latency_p50_ms": 309.91,
        "loads": true,
        "max_rel_diff": null,
        "output_match": null,
        "provenance": "measured",
        "runs": true
      },
      {
        "backend": "webgpu_mldrift",
        "date": "2026-08-11",
        "env": {
          "browser": "chromium",
          "browser_version": "151.0.7922.34",
          "headless": true,
          "jspi": true,
          "litertjs_core_version": "2.5.3",
          "machine_label": "mac-studio-m4-max",
          "os": "macOS",
          "os_version": "27.0.0",
          "webgpu_adapter": {
            "architecture": "metal-3",
            "description": "",
            "device": "",
            "vendor": "apple"
          }
        },
        "full_delegation": true,
        "latency_p50_ms": 3.365,
        "loads": true,
        "max_rel_diff": 0.07765237020316026,
        "output_match": true,
        "provenance": "measured",
        "runs": true
      }
    ],
    "demo_url": null,
    "sweep_source": "data/sweep/2.5.3/2026-08-11/wav2vec2-keyword-spotting__w2v2_frontend_fp16.json"
  },
  "conversion": {
    "command": "python wav2vec2-kws/scripts/build_w2v2_split.py",
    "quantization": "fp16",
    "tool": "litert-torch",
    "tool_version": "0.10.0 (editable dev checkout 115a136 + local patches)"
  },
  "cross_runtime": [],
  "delegation": null,
  "device": {
    "records": [
      {
        "device": "galaxy-s26",
        "run": {
          "accelerator": "gpu_mldrift",
          "context_length": null,
          "date": "2026-08-25",
          "decode_tokens_per_s": null,
          "delegated_ops": null,
          "env": {
            "device": "Galaxy S26 (SM-S942Q)",
            "machine_label": "galaxy-s26-npubench",
            "os_build": "Android 16",
            "runtime": "litert",
            "runtime_version": "2.2.0",
            "soc": "Qualcomm SM8850",
            "vendor_sdk": null
          },
          "error": null,
          "evidence": [
            "npubench sweep row: N=50 median, 1 backend = 1 process, accepted only with thermal NONE->NONE; NPU rows additionally required qnn_partition delegate evidence in logcat (a stock model asked for on the NPU can silently land on XNNPACK and report a CPU number)",
            "mode=jit thermal=NONE->NONE headroom=0.8->0.8 attempt=0"
          ],
          "failure_class": null,
          "full_delegation": null,
          "latency_p50_ms": 5.285,
          "loads": true,
          "max_abs_diff": null,
          "max_rel_diff": null,
          "metrics": {
            "iterations": 50,
            "latency_max_ms": 5.538,
            "latency_min_ms": 5.05,
            "load_ms": 656.4
          },
          "output_match": null,
          "peak_mem_mb": null,
          "prefill_tokens_per_s": null,
          "provenance": "measured",
          "runs": true,
          "total_ops": null,
          "ttft_ms": null
        },
        "source": "data/device_runs/2.2.0/2026-08-25/wav2vec2-keyword-spotting__w2v2_frontend_fp16__galaxy-s26.json"
      },
      {
        "device": "galaxy-s26",
        "run": {
          "accelerator": "npu_qnn",
          "context_length": null,
          "date": "2026-08-25",
          "decode_tokens_per_s": null,
          "delegated_ops": null,
          "env": {
            "device": "Galaxy S26 (SM-S942Q)",
            "machine_label": "galaxy-s26-npubench",
            "os_build": "Android 16",
            "runtime": "litert",
            "runtime_version": "2.2.0",
            "soc": "Qualcomm SM8850",
            "vendor_sdk": "QAIRT (Hexagon, JIT on-device)"
          },
          "error": "LiteRtException: Failed to invoke the compiled model",
          "evidence": [
            "LiteRtException: Failed to invoke the compiled model"
          ],
          "failure_class": "invoke_failed",
          "full_delegation": null,
          "latency_p50_ms": null,
          "loads": true,
          "max_abs_diff": null,
          "max_rel_diff": null,
          "output_match": null,
          "peak_mem_mb": null,
          "prefill_tokens_per_s": null,
          "provenance": "measured",
          "runs": false,
          "total_ops": null,
          "ttft_ms": null
        },
        "source": "data/device_runs/2.2.0/2026-08-25/wav2vec2-keyword-spotting__w2v2_frontend_fp16__galaxy-s26.json"
      },
      {
        "device": "raspberry-pi-5",
        "run": {
          "accelerator": "cpu_xnnpack",
          "context_length": null,
          "date": "2026-08-31",
          "decode_tokens_per_s": null,
          "delegated_ops": null,
          "env": {
            "device": "Raspberry Pi 5 Model B Rev 1.1",
            "machine_label": "raspberry-pi-5",
            "os_build": "Linux-6.18.34+rpt-rpi-2712-aarch64-with-glibc2.41",
            "runtime": "litert",
            "runtime_version": "2.2.0.dev20260804",
            "soc": null,
            "vendor_sdk": null
          },
          "error": null,
          "evidence": [
            "pi5 sweep row: LiteRT benchmark_model, CPU/XNNPACK at --num_threads=4, 3 invocations per file of 10 warm-up + 50 timed runs (the tool caps a phase at 150 s, so very slow graphs run fewer); latency = median of the three per-invocation medians over the timed phase; a row counts as measured only when every invocation exited 0 with XNNPACK engaged and vcgencmd get_throttled 0x0 before and after",
            "versions: ai-edge-litert-nightly=2.2.0.dev20260804, benchmark_model_sha256=babc9275addd8612caf0842294cb5b3f301c3cce46e6c5979cd2803f215f6bee, cpu=Raspberry Pi 5 Model B Rev 1.1, litert-cli-nightly=0.2.0.dev20260805, platform=Linux-6.18.34+rpt-rpi-2712-aarch64-with-glibc2.41, python=3.13.5",
            "invocation 0: exit=0 wall_s=6.5 temp 51.0->63.1C throttled=0x0 xnnpack=True median_us=106192.0 runs=50 footprint_peak_mb=137.5",
            "invocation 1: exit=0 wall_s=6.5 temp 51.6->63.1C throttled=0x0 xnnpack=True median_us=106258.0 runs=50 footprint_peak_mb=137.5",
            "invocation 2: exit=0 wall_s=6.5 temp 50.5->61.5C throttled=0x0 xnnpack=True median_us=106245.0 runs=50 footprint_peak_mb=137.5"
          ],
          "failure_class": null,
          "full_delegation": null,
          "latency_p50_ms": 106.245,
          "loads": true,
          "max_abs_diff": null,
          "max_rel_diff": null,
          "metrics": {
            "init_ms": 58.64,
            "invocations": 3,
            "iterations": 150,
            "latency_max_ms": 109.173,
            "latency_min_ms": 105.706,
            "threads": 4,
            "warmup_runs_per_invocation": 10
          },
          "output_match": null,
          "peak_mem_mb": 137.5,
          "prefill_tokens_per_s": null,
          "provenance": "measured",
          "runs": true,
          "total_ops": null,
          "ttft_ms": null
        },
        "source": "data/device_runs/2.2.0.dev20260804/2026-08-31/wav2vec2-keyword-spotting__w2v2_frontend_fp16__raspberry-pi-5.json"
      }
    ]
  },
  "model": {
    "family": "wav2vec2",
    "id": "wav2vec2-keyword-spotting__w2v2_frontend_fp16",
    "license": "apache-2.0",
    "source_url": "https://huggingface.co/litert-community/wav2vec2-keyword-spotting",
    "task": "audio-classification"
  },
  "pitfalls": [
    "The model ships as two graphs (frontend + head) because the full 1008-node graph exceeds the Mali shader-compile limit; both halves are needed for a prediction (zoo README L1195)."
  ],
  "schema_version": "1.2"
}