{
  "artifacts": [
    {
      "file": "phocinae_largha_150m_L512_fp16fc_int8emb.tflite",
      "sha256": "33454add13a442cf43ee84c61555b412dff31ab1744385dc86a2c57477926926",
      "size_mb": 186.684
    }
  ],
  "benchmarks": [],
  "browser": {
    "backends": [
      {
        "backend": "wasm_xnnpack",
        "date": "2026-10-11",
        "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": 976.315,
        "loads": true,
        "max_rel_diff": null,
        "output_match": null,
        "provenance": "measured",
        "runs": true
      },
      {
        "backend": "webgpu_mldrift",
        "date": "2026-10-11",
        "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": 24.563,
        "loads": true,
        "max_rel_diff": 0.00027210438317542537,
        "output_match": true,
        "provenance": "measured",
        "runs": true
      }
    ],
    "demo_url": null,
    "sweep_source": "data/sweep/2.5.3/2026-10-11/phocinae-largha-150m__phocinae_largha_150m_l512_fp16fc_int8emb.json"
  },
  "conversion": {
    "command": "conversion/ scripts of the Hub repo (export.py per window, quantize.py for the fp16 FC weights and the table forms, tables.py for the host fp16 table); REPRODUCE.md of the Hub repo",
    "quantization": "int8 EMBEDDING_LOOKUP table inside the graph (input_ids form) + fp16 FULLY_CONNECTED weights read through DEQUANTIZE; the int8 table costs max |Δp| 0.0066 to 0.0107 against the reference (card 'Conversion notes'); the L256 graph has 1,804 operators, among them 1 EMBEDDING_LOOKUP and 99 FULLY_CONNECTED",
    "tool": "litert-torch",
    "tool_version": "0.10.0 (ai-edge-litert 2.2.0 for the desktop checks)"
  },
  "cross_runtime": [],
  "delegation": {
    "backend": "gpu_mldrift",
    "blocking_ops": [
      "DEQUANTIZE",
      "SQUARED_DIFFERENCE",
      "RESHAPE",
      "EMBEDDING_LOOKUP"
    ],
    "coverage_ops_pct": 91.6,
    "lint_report_version": "1.1",
    "litert_version": "2.2.0",
    "matched_provenance_counts": {
      "measured": 1752,
      "unmatched": 52
    },
    "partitions": 150
  },
  "device": {
    "records": [
      {
        "device": "galaxy-s26",
        "run": {
          "accelerator": "cpu_xnnpack",
          "context_length": null,
          "date": "2026-10-10",
          "decode_tokens_per_s": null,
          "delegated_ops": 1802,
          "env": {
            "device": "Galaxy S26 (SM-S942Q)",
            "machine_label": "galaxy-s26-phocinae-gate",
            "os_build": "Android 36 build BP4A.251205.006.S942QOPS1AZH9",
            "runtime": "litert",
            "runtime_version": "2.2.0",
            "soc": "Qualcomm SM8850",
            "vendor_sdk": null
          },
          "error": null,
          "evidence": [
            "Debug gate app com.phocinae.gate (android/gate/, Kotlin CompiledModel 2.2.0), one cold process per run, fixture ids + markers pushed as JSON, the host token-table lookup (embeds form) outside the measured interval; measured interval = input writes + run + output readback per row.",
            "parity vs the publisher's fp32 PyTorch (phocinae-server 0.1.5 Engine; scripts/device_compare.py): same answer on 1951/1951 non-near-tie test rows and 45/49 near ties (reference top-2 gap <= 0.02), max |dp| 0.0107, mean 6.1e-04 over 2605 rows (test + reversed-order + control); bar (argmax 100 % on non-near-tie rows, max |dp| <= 0.02, mean <= 0.002): PASS.",
            "output_match = element rule |out - ref| <= max(1e-5, 1e-3 * |ref|) over 7100 option logits of the 2000 test rows against the PyTorch logits: 5665 violations (fp16 weights move logits by up to 0.0340; the bar is on probabilities).",
            "latency_p50_ms = the rested run (thermal status 0, CPU / GPU caps off, 45 s rest): a 323-token row, median of 20 calls after 5 warm-ups = 96.55 ms (min 90.38, max 100.98); the 2605-row parity run's warm median was 219.29 ms (a long run caps the GPU to 461 MHz and the CPU to ~1.5 GHz by its end).",
            "conditions: screen on (gate activity over the keyguard), USB powered, airplane mode off, one process per run; compile 296 ms (NPU: JIT, cold cache on the first run of a graph).",
            "10-10 16:52:00.889  9338  9375 I litert  : [compiled_model.cc:1250] 0 compiler plugins were applied successfully:",
            "10-10 16:52:00.991  9338  9375 I tflite  : Replacing 1802 out of 1804 node(s) with delegate (TfLiteXNNPackDelegate) node, yielding 2 partitions for subgraph 0 (main).",
            "10-10 16:52:01.132  9338  9375 I litert  : [compiled_model.cc:1250] 0 compiler plugins were applied successfully:",
            "10-10 16:52:01.132  9338  9375 I tflite  : Replacing 4 out of 4 node(s) with delegate (TfLiteXNNPackDelegate) node, yielding 1 partitions for subgraph 0 (main)."
          ],
          "failure_class": null,
          "full_delegation": false,
          "latency_p50_ms": 96.553698,
          "loads": true,
          "max_abs_diff": 0.034007370471954346,
          "max_rel_diff": 6.092772586094159,
          "metrics": {
            "all_rows": 2605,
            "argmax_equal_official": 1951,
            "bar_pass": 1,
            "element_rule_elements": 7100,
            "element_rule_violations": 5665,
            "first_call_ms": 150.845782,
            "fixtures": 2000,
            "load_compile_ms": 296.086354,
            "max_abs_dprob_vs_official": 0.010686568120104711,
            "mean_abs_dprob_vs_official": 0.0006108502442360601,
            "near_tie_argmax_equal": 45,
            "near_tie_rows": 49,
            "non_near_tie_rows": 1951,
            "parity_leg_warm_max_ms": 260.974167,
            "parity_leg_warm_median_ms": 219.285417,
            "parity_leg_warm_min_ms": 132.646458,
            "partitions": 2,
            "rested_leg_max_ms": 100.982812,
            "rested_leg_median_ms": 96.553698,
            "rested_leg_min_ms": 90.377709,
            "rested_leg_n": 20,
            "rested_leg_row_len": 323,
            "rested_leg_warmups": 5,
            "rows_dprob_gt_1e2": 1,
            "threads": 4
          },
          "output_match": false,
          "peak_mem_mb": null,
          "prefill_tokens_per_s": null,
          "provenance": "measured",
          "runs": true,
          "signature": "CPU XNNPACK 4 threads; int8 EMBEDDING_LOOKUP table in the graph (input_ids); debug gate app com.phocinae.gate",
          "total_ops": 1804,
          "ttft_ms": null
        },
        "source": "data/device_runs/2.2.0/2026-10-10/phocinae-largha-150m__phocinae_largha_150m_l512_fp16fc_int8emb__galaxy-s26.json"
      },
      {
        "device": "galaxy-s26",
        "run": {
          "accelerator": "gpu_mldrift",
          "context_length": null,
          "date": "2026-10-10",
          "decode_tokens_per_s": null,
          "delegated_ops": 1804,
          "env": {
            "device": "Galaxy S26 (SM-S942Q)",
            "machine_label": "galaxy-s26-phocinae-gate",
            "os_build": "Android 36 build BP4A.251205.006.S942QOPS1AZH9",
            "runtime": "litert",
            "runtime_version": "2.2.0",
            "soc": "Qualcomm SM8850",
            "vendor_sdk": null
          },
          "error": null,
          "evidence": [
            "Debug gate app com.phocinae.gate (android/gate/, Kotlin CompiledModel 2.2.0), one cold process per run, fixture ids + markers pushed as JSON, the host token-table lookup (embeds form) outside the measured interval; measured interval = input writes + run + output readback per row.",
            "parity vs the publisher's fp32 PyTorch (phocinae-server 0.1.5 Engine; scripts/device_compare.py): same answer on 1951/1951 non-near-tie test rows and 45/49 near ties (reference top-2 gap <= 0.02), max |dp| 0.0107, mean 6.1e-04 over 2605 rows (test + reversed-order + control); bar (argmax 100 % on non-near-tie rows, max |dp| <= 0.02, mean <= 0.002): PASS.",
            "output_match = element rule |out - ref| <= max(1e-5, 1e-3 * |ref|) over 7100 option logits of the 2000 test rows against the PyTorch logits: 5665 violations (fp16 weights move logits by up to 0.0340; the bar is on probabilities).",
            "latency_p50_ms = the rested run (thermal status 0, CPU / GPU caps off, 45 s rest): a 323-token row, median of 20 calls after 5 warm-ups = 62.73 ms (min 62.45, max 64.11); the 2605-row parity run's warm median was 132.34 ms (a long run caps the GPU to 461 MHz and the CPU to ~1.5 GHz by its end).",
            "conditions: screen on (gate activity over the keyguard), USB powered, airplane mode off, one process per run; compile 3099 ms (NPU: JIT, cold cache on the first run of a graph).",
            "10-10 16:35:00.060  3275  3298 I litert  : [compiled_model.cc:1250] 0 compiler plugins were applied successfully:",
            "10-10 16:35:00.172  3275  3298 I tflite  : Replacing 1804 out of 1804 node(s) with delegate (LITERT_CL) node, yielding 1 partitions for subgraph 0 (main).",
            "10-10 16:35:03.113  3275  3298 I litert  : [compiled_model.cc:1250] 0 compiler plugins were applied successfully:",
            "10-10 16:35:03.114  3275  3298 I tflite  : Replacing 4 out of 4 node(s) with delegate (TfLiteXNNPackDelegate) node, yielding 1 partitions for subgraph 0 (main)."
          ],
          "failure_class": null,
          "full_delegation": true,
          "latency_p50_ms": 62.727917,
          "loads": true,
          "max_abs_diff": 0.03400486707687378,
          "max_rel_diff": 6.092605710549259,
          "metrics": {
            "all_rows": 2605,
            "argmax_equal_official": 1951,
            "bar_pass": 1,
            "element_rule_elements": 7100,
            "element_rule_violations": 5665,
            "first_call_ms": 146.071406,
            "fixtures": 2000,
            "load_compile_ms": 3098.91104,
            "max_abs_dprob_vs_official": 0.010685852174722843,
            "mean_abs_dprob_vs_official": 0.0006108472523848311,
            "near_tie_argmax_equal": 45,
            "near_tie_rows": 49,
            "non_near_tie_rows": 1951,
            "parity_leg_warm_max_ms": 146.495573,
            "parity_leg_warm_median_ms": 132.335833,
            "parity_leg_warm_min_ms": 95.551146,
            "partitions": 1,
            "rested_leg_max_ms": 64.108541,
            "rested_leg_median_ms": 62.727917,
            "rested_leg_min_ms": 62.454896,
            "rested_leg_n": 20,
            "rested_leg_row_len": 323,
            "rested_leg_warmups": 5,
            "rows_dprob_gt_1e2": 1
          },
          "output_match": false,
          "peak_mem_mb": null,
          "prefill_tokens_per_s": null,
          "provenance": "measured",
          "runs": true,
          "signature": "GpuOptions precision=FP32; int8 EMBEDDING_LOOKUP table in the graph (input_ids); debug gate app com.phocinae.gate",
          "total_ops": 1804,
          "ttft_ms": null
        },
        "source": "data/device_runs/2.2.0/2026-10-10/phocinae-largha-150m__phocinae_largha_150m_l512_fp16fc_int8emb__galaxy-s26.json"
      }
    ]
  },
  "model": {
    "family": "phocinae-largha-150m",
    "id": "phocinae-largha-150m__phocinae_largha_150m_l512_fp16fc_int8emb",
    "license": "apache-2.0",
    "source_url": "https://huggingface.co/litert-community/Phocinae-Largha-150M-LiteRT",
    "task": "text-classification"
  },
  "pitfalls": [
    "Every graph computes 21 LayerNorms on x·2⁻⁶ with the epsilon scaled to match: bit-identical in fp32; without it the sums of squares inside these norms leave the float16 range, so fp16 arithmetic and the NPU need it (card 'Conversion notes').",
    "Inputs: `input_ids` int32 [1,L] (int8emb) or `inputs_embeds` float32 [1,L,384] (embeds), `attention_mask` float32 [1,L], `qtype_onehot` float32 [1,3] in the order choice, score, noul; outputs `token_logits` float32 [1,L] and `pooled_cls` float32 [1,384]. Gather token_logits at each option's <mask> marker, divide by the question type's temperature from rl_agent_config.json (choice 0.8660, score 0.8081, noul 0.6625) and softmax; the act head takes pooled_cls + four features of the pre-temperature softmax and returns act_logits (card 'Raw CompiledModel contract').",
    "GPU FP16 (float16 arithmetic) passes the bar with less margin: it flipped 1 to 2 of the 20 near ties at L256 and its max |Δp| reaches 0.0188 at L512 (Mac Metal default, embeds); GPU FP32, FP16_WITH_FP32_ACCUM and the NPU keep max |Δp| at or below 0.0137 (card 'Limits').",
    "Dynamic int8 FULLY_CONNECTED fails the bar (max |Δp| 0.0835) and does not ship; the GPU delegate refuses a float16 table inside the graph, so the float16 table ships for host lookup instead (card 'Limits', 'Conversion notes').",
    "A graph computes every position of its window, so a smaller window runs faster; the host cuts the state to fit the window while the publisher's runtime cuts at 512 (card 'Conversion notes', 'Limits').",
    "The checks measure agreement with the publisher's fp32 runtime (phocinae-server) on the public typed-decisions test split, not accuracy against gold labels; measured on one Apple M4 Max and one Galaxy S26 only, not through the hfmodels-android SDK (card 'Checks', 'Limits').",
    "Of the 2,000 test decisions the state is cut on 111 at L512 (card 'Limits').",
    "The int8emb form was not run on the NPU at L384 or L512; at L256 it ran with the int8 lookup on the CPU, in a second partition (card 'Limits')."
  ],
  "schema_version": "1.2"
}
