Record X Elite MiniCPM speech head inference

2026-09-24 08:18:41 +08:00
1 parent e5aaf3186c
commit afd3f8c716
@@ -6,6 +6,6 @@ A local HX370 x86-64 LiteRT CPU control of the identical SA/RB3 TFLite binary ma

 The [talker decode-step attempt](evidence/qwen_tts_talker_qualcomm/README.md) pins a separate 28-layer source and synthetic cache fixture. Historical FP16 QNN context-binary inference executed on S24, S25, SA8775P and X Elite, but all four changed the CPU top-1 speech token with 8.57–9.43% logits relative L2. These artifacts are numerically unqualified on the fixture. RB3 rejected the floating-point-input artifact. Explicit X Elite ONNX CPU and SA8775P TFLite CPU fallback compiles passed, with same-fixture inference pending. No complete talker loop or TTS stream passed.

-The [MiniCPM-o speech-head study](evidence/minicpmo_tts_qualcomm/README.md) recovered the real 20-layer ONNX source and a matching retained fixture. S24 and S25 FP16 context-binary outputs preserved the ONNX CPU top-1 speech token with 0.346%/0.360% logits relative L2 on one step. FP16 QNN DLC compiles passed for SA8775P, X Elite and RB3. SA8775P NPU-requested inference preserved CPU top-1 token 1867 with 0.3568% logits relative L2; placement profiling remains pending, as do X Elite and RB3 inferences. The complete multimodal model remains unqualified on those devices.
+The [MiniCPM-o speech-head study](evidence/minicpmo_tts_qualcomm/README.md) recovered the real 20-layer ONNX source and a matching retained fixture. S24 and S25 FP16 context-binary outputs preserved the ONNX CPU top-1 speech token with 0.346%/0.360% logits relative L2 on one step. FP16 QNN DLC compiles passed for SA8775P, X Elite and RB3. SA8775P NPU-requested inference preserved CPU top-1 token 1867 with 0.3568% logits relative L2; X Elite same-fixture inference also preserved token 1867 with 0.3170% logits relative L2. SA and X Elite placement profiles and RB3 inference remain pending. The complete multimodal model remains unqualified on those devices.

 The [InternVLA Cosmos S25 attempt](evidence/internvla_cosmos_qualcomm/README.md) repackages the real policy's image encoder from its parity-tested ONNX export. FLOAT16 QNN DLC inference returned a latent within 0.0524% relative L2 of the source on one pattern. Injecting it into the real policy on local CPU changed synthetic actions by 0.742% relative L2, without a task tolerance. NPU placement profiling was pending; device-local action-policy support remains unqualified.
@@ -6,6 +6,6 @@ The [S25 historical inference](s25/inference_report.json) used dataset `d7gwxe3y

 The historical [S24](s24/compile_report.json) and [S25](s25/compile_report.json) FP16 QNN context-binary compiles and inference jobs succeeded. Against the **same original ONNX CPU fixture**, [S24](s24/audit_report.json) logits relative L2 was **0.3459%** and max K/V relative L2 **1.235%**; [S25](s25/audit_report.json) logits relative L2 was **0.3605%** and max K/V relative L2 **1.311%**. Both preserved the CPU top-1 speech token **1867** on this one synthetic step. The raw [S24](s24/device_output.npz) and [S25](s25/device_output.npz) arrays are retained. [S24](s24/profile_submission.json) and [S25](s25/profile_submission.json) placement profiles were submitted and pending. These values remain numerical component evidence, without a generated-audio quality threshold or actual NPU placement proof.

-New FP16 QNN DLC [SA8775P](sa8775p/compile_report.json), [Snapdragon X Elite](xelite/compile_report.json) and [RB3 Gen 2](rb3/compile_report.json) compiles passed from the same source. The SA8775P [NPU-requested inference](sa8775p/inference_report.json) returned all 41 finite tensors on the retained fixture. Its [audit](sa8775p/audit_report.json) found **0.3568% logits relative L2** versus ONNX CPU, preserved CPU top-1 speech token **1867**, and measured maximum K/V relative L2 **1.225%**. The [placement profile](sa8775p/profile_submission.json) was submitted and remains pending; the inference request alone does not prove actual NPU node assignment. Same-fixture NPU-requested inference jobs for [X Elite](xelite/inference_submission.json) and [RB3](rb3/inference_submission.json) also remain pending. The [collector](../../../../experiments/collect_aihub_component.py) retrieves terminal jobs, and the [auditor](../../../../experiments/audit_minicpmo_tts_qualcomm.py) pins hashes and checks all 41 output tensors.
+New FP16 QNN DLC [SA8775P](sa8775p/compile_report.json), [Snapdragon X Elite](xelite/compile_report.json) and [RB3 Gen 2](rb3/compile_report.json) compiles passed from the same source. The SA8775P [NPU-requested inference](sa8775p/inference_report.json) returned all 41 finite tensors on the retained fixture. Its [audit](sa8775p/audit_report.json) found **0.3568% logits relative L2** versus ONNX CPU, preserved CPU top-1 speech token **1867**, and measured maximum K/V relative L2 **1.225%**. The [placement profile](sa8775p/profile_submission.json) was submitted and remains pending; the inference request alone does not prove actual NPU node assignment. The X Elite [NPU-requested inference](xelite/inference_report.json) also returned all 41 finite tensors: its [audit](xelite/audit_report.json) found **0.3170% logits relative L2**, preserved CPU top-1 token **1867**, and measured maximum K/V relative L2 **1.221%**. Its [placement profile](xelite/profile_submission.json) was submitted and remains pending. [RB3 same-fixture inference](rb3/inference_submission.json) remains pending. The [collector](../../../../experiments/collect_aihub_component.py) retrieves terminal jobs, and the [auditor](../../../../experiments/audit_minicpmo_tts_qualcomm.py) pins hashes and checks all 41 output tensors.

 Next: audit those device outputs and placement, then qualify the entire model with real multimodal inputs, stateful speech generation, task/audio quality, memory admission, cancellation and sustained power/thermal behavior. AI Hub is the test facility, not a deployment dependency.
@@ -0,0 +1,304 @@
+{
+  "scope": "one fixed-cache 20-layer MiniCPM-o speech-head decode step; no complete multimodal stream",
+  "status": "component_inference_numeric_measured",
+  "device": {
+    "name": "Snapdragon X Elite CRD",
+    "os": "11",
+    "attributes": [
+      "os:windows",
+      "format:compute",
+      "framework:onnx",
+      "vendor:qualcomm",
+      "abi:aarch64-windows",
+      "chipset:qualcomm-snapdragon-x-elite",
+      "chipset:sc8380xp",
+      "hexagon:v73",
+      "soc-model:60",
+      "htp-supports-fp16:true",
+      "htp-supports-weight-sharing:true",
+      "framework:qnn"
+    ]
+  },
+  "compile_job_id": "jgzlm4mk5",
+  "inference_job_id": "jgolr4dxg",
+  "source_model_id": "mn4gjjpvn",
+  "target_model_id": "mmd04pj3n",
+  "input_dataset_id": "d7gwxe3y2",
+  "requested_run_options": "--compute_unit npu",
+  "files": {
+    "source_onnx": {
+      "bytes": 776341465,
+      "sha256": "78cf64804f11ad269ee4180da61588ccc5eefe2942773227345499756e4cc229"
+    },
+    "fixture": {
+      "bytes": 29247541,
+      "sha256": "65919ea1f351feec33c1415a3e3db76d60bc9c2beb3d49a72af75dcb418e83b4"
+    },
+    "cpu_reference": {
+      "bytes": 148471,
+      "sha256": "902adcceadbe761d938ef2ade51cf953c8917ea1abfaee33a10c84e300c1bd96"
+    },
+    "device_output": {
+      "bytes": 105640,
+      "sha256": "5ec274be28d3a5436d0d2986db1368c0731732dd8db3f2da3d85642a79140c12"
+    }
+  },
+  "tensor_comparison": {
+    "logits": {
+      "relative_l2": 0.0031700028839813,
+      "max_abs_error": 0.01659095287322998,
+      "source_rms": 1.4974441755558239,
+      "device_rms": 1.4976756779801723,
+      "cpu_top1": 1867,
+      "device_top1": 1867,
+      "top1_matches": true
+    },
+    "k_new_0": {
+      "relative_l2": 0.0005147656431265523,
+      "max_abs_error": 0.003742694854736328,
+      "source_rms": 1.307108108274551,
+      "device_rms": 1.307038088812209
+    },
+    "v_new_0": {
+      "relative_l2": 0.00035345147473398163,
+      "max_abs_error": 0.0007248520851135254,
+      "source_rms": 0.5083994128636046,
+      "device_rms": 0.5084141364451746
+    },
+    "k_new_1": {
+      "relative_l2": 0.0005523109095696163,
+      "max_abs_error": 0.0020580291748046875,
+      "source_rms": 0.7542700462780494,
+      "device_rms": 0.7542605237718504
+    },
+    "v_new_1": {
+      "relative_l2": 0.0010212503665159528,
+      "max_abs_error": 0.00030508264899253845,
+      "source_rms": 0.0662130257652199,
+      "device_rms": 0.06622190429742661
+    },
+    "k_new_2": {
+      "relative_l2": 0.0012660332300085497,
+      "max_abs_error": 0.0056171417236328125,
+      "source_rms": 0.9817805619626084,
+      "device_rms": 0.9819165272792516
+    },
+    "v_new_2": {
+      "relative_l2": 0.0013941202877916915,
+      "max_abs_error": 0.0017390847206115723,
+      "source_rms": 0.24438175712240315,
+      "device_rms": 0.24439989486728894
+    },
+    "k_new_3": {
+      "relative_l2": 0.004104747307563847,
+      "max_abs_error": 0.013773024082183838,
+      "source_rms": 0.9312003114105484,
+      "device_rms": 0.9323887799652595
+    },
+    "v_new_3": {
+      "relative_l2": 0.004388581847016711,
+      "max_abs_error": 0.006326019763946533,
+      "source_rms": 0.27179250329538757,
+      "device_rms": 0.27220384489136223
+    },
+    "k_new_4": {
+      "relative_l2": 0.007779545096767518,
+      "max_abs_error": 0.039246320724487305,
+      "source_rms": 1.2078244698126999,
+      "device_rms": 1.2115520989954713
+    },
+    "v_new_4": {
+      "relative_l2": 0.008414148179106594,
+      "max_abs_error": 0.01820969581604004,
+      "source_rms": 0.4929021962159226,
+      "device_rms": 0.4943089809183415
+    },
+    "k_new_5": {
+      "relative_l2": 0.007452034487665665,
+      "max_abs_error": 0.04445651173591614,
+      "source_rms": 1.0960093751879882,
+      "device_rms": 1.0966943330718524
+    },
+    "v_new_5": {
+      "relative_l2": 0.012211649189410728,
+      "max_abs_error": 0.027927041053771973,
+      "source_rms": 0.49114616760529134,
+      "device_rms": 0.48967054012047084
+    },
+    "k_new_6": {
+      "relative_l2": 0.0061292171562801,
+      "max_abs_error": 0.03535480797290802,
+      "source_rms": 1.1726536779527166,
+      "device_rms": 1.173115323923179
+    },
+    "v_new_6": {
+      "relative_l2": 0.00944127599594071,
+      "max_abs_error": 0.016520775854587555,
+      "source_rms": 0.5369044199454948,
+      "device_rms": 0.5363747360693308
+    },
+    "k_new_7": {
+      "relative_l2": 0.003890039826671964,
+      "max_abs_error": 0.021133124828338623,
+      "source_rms": 1.6917150856043028,
+      "device_rms": 1.6937682079827496
+    },
+    "v_new_7": {
+      "relative_l2": 0.006906994746270466,
+      "max_abs_error": 0.02008109539747238,
+      "source_rms": 0.8116460286615261,
+      "device_rms": 0.8107319955904663
+    },
+    "k_new_8": {
+      "relative_l2": 0.0047713333145434685,
+      "max_abs_error": 0.031733155250549316,
+      "source_rms": 1.5021910590457033,
+      "device_rms": 1.5033379291990285
+    },
+    "v_new_8": {
+      "relative_l2": 0.0066740158237721245,
+      "max_abs_error": 0.027550816535949707,
+      "source_rms": 0.8807572129778053,
+      "device_rms": 0.8793833990886489
+    },
+    "k_new_9": {
+      "relative_l2": 0.004665944378169018,
+      "max_abs_error": 0.01780569553375244,
+      "source_rms": 1.1054146201792987,
+      "device_rms": 1.1057316668308173
+    },
+    "v_new_9": {
+      "relative_l2": 0.007561300460793002,
+      "max_abs_error": 0.01777637004852295,
+      "source_rms": 0.6580213796976341,
+      "device_rms": 0.6569723354399966
+    },
+    "k_new_10": {
+      "relative_l2": 0.004428435955964545,
+      "max_abs_error": 0.0177919864654541,
+      "source_rms": 1.325937063058518,
+      "device_rms": 1.3268511996216796
+    },
+    "v_new_10": {
+      "relative_l2": 0.0066985124074587276,
+      "max_abs_error": 0.021573245525360107,
+      "source_rms": 0.848351970479369,
+      "device_rms": 0.8468234992699584
+    },
+    "k_new_11": {
+      "relative_l2": 0.003429064933197882,
+      "max_abs_error": 0.008336663246154785,
+      "source_rms": 0.6808336872062942,
+      "device_rms": 0.6804917001316813
+    },
+    "v_new_11": {
+      "relative_l2": 0.006156563379302956,
+      "max_abs_error": 0.008510656654834747,
+      "source_rms": 0.3676426367961426,
+      "device_rms": 0.3672159888143714
+    },
+    "k_new_12": {
+      "relative_l2": 0.004129233348527031,
+      "max_abs_error": 0.008901238441467285,
+      "source_rms": 0.6128882051701765,
+      "device_rms": 0.6122828966759465
+    },
+    "v_new_12": {
+      "relative_l2": 0.005815610169090192,
+      "max_abs_error": 0.007399111986160278,
+      "source_rms": 0.34029547623567724,
+      "device_rms": 0.33994744996075893
+    },
+    "k_new_13": {
+      "relative_l2": 0.003986479035569083,
+      "max_abs_error": 0.00892484188079834,
+      "source_rms": 0.613775761616868,
+      "device_rms": 0.6134809238432836
+    },
+    "v_new_13": {
+      "relative_l2": 0.006311382994234031,
+      "max_abs_error": 0.008226275444030762,
+      "source_rms": 0.36907442918730654,
+      "device_rms": 0.36850471315926514
+    },
+    "k_new_14": {
+      "relative_l2": 0.003975218959210037,
+      "max_abs_error": 0.009820103645324707,
+      "source_rms": 0.5980841827735426,
+      "device_rms": 0.5977833545317282
+    },
+    "v_new_14": {
+      "relative_l2": 0.005947601968032226,
+      "max_abs_error": 0.006577491760253906,
+      "source_rms": 0.3506656009170703,
+      "device_rms": 0.3503015513144008
+    },
+    "k_new_15": {
+      "relative_l2": 0.0035972324999042045,
+      "max_abs_error": 0.008877158164978027,
+      "source_rms": 0.622821559168189,
+      "device_rms": 0.6224176726307988
+    },
+    "v_new_15": {
+      "relative_l2": 0.005320288804772653,
+      "max_abs_error": 0.0070031434297561646,
+      "source_rms": 0.41823950764909973,
+      "device_rms": 0.4178766892221755
+    },
+    "k_new_16": {
+      "relative_l2": 0.0031916991317839987,
+      "max_abs_error": 0.006972968578338623,
+      "source_rms": 0.6754924715823191,
+      "device_rms": 0.6753580010409884
+    },
+    "v_new_16": {
+      "relative_l2": 0.005726146021455772,
+      "max_abs_error": 0.009386241436004639,
+      "source_rms": 0.4659773620123202,
+      "device_rms": 0.4656833192747271
+    },
+    "k_new_17": {
+      "relative_l2": 0.003304237761951359,
+      "max_abs_error": 0.007736682891845703,
+      "source_rms": 0.6354120615871959,
+      "device_rms": 0.6351539196274629
+    },
+    "v_new_17": {
+      "relative_l2": 0.005790382158371946,
+      "max_abs_error": 0.011991262435913086,
+      "source_rms": 0.5384818316759848,
+      "device_rms": 0.5380287115508268
+    },
+    "k_new_18": {
+      "relative_l2": 0.0029809011639930372,
+      "max_abs_error": 0.007440060377120972,
+      "source_rms": 0.7283659936841244,
+      "device_rms": 0.7283979149704595
+    },
+    "v_new_18": {
+      "relative_l2": 0.00410675170171758,
+      "max_abs_error": 0.0086001455783844,
+      "source_rms": 0.6382820941879574,
+      "device_rms": 0.6381469978008664
+    },
+    "k_new_19": {
+      "relative_l2": 0.0036517005773021058,
+      "max_abs_error": 0.007315650582313538,
+      "source_rms": 0.5790109825658509,
+      "device_rms": 0.5790702382536065
+    },
+    "v_new_19": {
+      "relative_l2": 0.004105625819220487,
+      "max_abs_error": 0.009264230728149414,
+      "source_rms": 0.7220926350353908,
+      "device_rms": 0.7220696027280022
+    }
+  },
+  "cache_max_relative_l2": 0.012211649189410728,
+  "limits": [
+    "The historical ONNX export does not attest the exact source checkpoint revision.",
+    "One retained decode fixture cannot establish speech-token or generated-audio quality.",
+    "Actual NPU placement requires a device profile; a request or QNN artifact alone is insufficient.",
+    "Thinker, audio/vision encoders, vocoder, state loop and complete Omni request were not tested."
+  ]
+}
@@ -0,0 +1,32 @@
+{
+  "timestamp_utc": "2026-09-24T00:17:22.418440+00:00",
+  "qai_hub_version": "0.55.0",
+  "facility": "Qualcomm AI Hub Workbench",
+  "job_id": "jgolr4dxg",
+  "url": "https://workbench.aihub.qualcomm.com/jobs/jgolr4dxg/",
+  "status": "SUCCESS",
+  "status_message": "",
+  "options": "--compute_unit npu",
+  "hub_version": "aihub-2026.09.11.0",
+  "device": {
+    "name": "Snapdragon X Elite CRD",
+    "os": "11",
+    "attributes": [
+      "os:windows",
+      "format:compute",
+      "framework:onnx",
+      "vendor:qualcomm",
+      "abi:aarch64-windows",
+      "chipset:qualcomm-snapdragon-x-elite",
+      "chipset:sc8380xp",
+      "hexagon:v73",
+      "soc-model:60",
+      "htp-supports-fp16:true",
+      "htp-supports-weight-sharing:true",
+      "framework:qnn"
+    ]
+  },
+  "model_id": "mmd04pj3n",
+  "input_dataset_id": "d7gwxe3y2",
+  "outputs": "device_output.npz"
+}
@@ -0,0 +1,13 @@
+{
+  "timestamp_utc": "2026-09-24T00:17:47.758312+00:00",
+  "facility": "Qualcomm AI Hub Workbench",
+  "qai_hub_version": "0.55.0",
+  "source_compile_job_id": "jgzlm4mk5",
+  "target_model_id": "mmd04pj3n",
+  "device": "Snapdragon X Elite CRD",
+  "options": "--compute_unit npu",
+  "kind": "profile",
+  "status": "SUBMITTED",
+  "job_id": "jp1nj6qkg",
+  "url": "https://workbench.aihub.qualcomm.com/jobs/jp1nj6qkg/"
+}
@@ -20,7 +20,7 @@ A separate [Qwen3-TTS talker decode-step attempt](../../../e2e_expansion_2026092

 A [Galaxy S25 InternVLA Cosmos attempt](../../../e2e_expansion_20260924/evidence/internvla_cosmos_qualcomm/README.md) packaged the real Place_Markpen image encoder into one ONNX file after two CPU parity fixtures. FLOAT16 QNN DLC inference returned a latent within 0.0524% relative L2 of the source on one pattern; device placement profile was pending. Injecting that latent into the real policy on local CPU changed synthetic actions by 0.742% relative L2, without a task tolerance. This advances component and action-sensitivity evidence, not device-local full-policy qualification.

-A [MiniCPM-o speech-head study](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md) recovered the historical 20-layer source and matching S24/S25 fixtures. Both FP16 context-binary outputs preserved the ONNX CPU top-1 speech token on this one decode step with logits relative L2 of 0.346%/0.360%. New FP16 QNN DLC compiles passed for SA8775P, X Elite and RB3. SA8775P same-fixture NPU-requested inference returned finite outputs, preserved CPU top-1 token 1867 and had 0.3568% logits relative L2; actual NPU placement profiling is pending. X Elite and RB3 inference remains pending. These are speech-head component results, not multimodal MiniCPM-o requests.
+A [MiniCPM-o speech-head study](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md) recovered the historical 20-layer source and matching S24/S25 fixtures. Both FP16 context-binary outputs preserved the ONNX CPU top-1 speech token on this one decode step with logits relative L2 of 0.346%/0.360%. New FP16 QNN DLC compiles passed for SA8775P, X Elite and RB3. SA8775P same-fixture NPU-requested inference returned finite outputs, preserved CPU top-1 token 1867 and had 0.3568% logits relative L2; actual NPU placement profiling is pending. X Elite same-fixture inference also preserved token 1867 with 0.3170% logits relative L2; its placement profile and RB3 inference remain pending. These are speech-head component results, not multimodal MiniCPM-o requests.

 ## Configuration × model

@@ -89,7 +89,7 @@ NOT E2E means no complete device/model workload was verified. “Scoped E2E” i
 | PC Snapdragon X Elite CRD / AI Hub | Spark-X2.5 | COMPONENT_NPU_INFERENCE_PROFILED; NOT_E2E_PROFILEABLE | A real Spark-X2.5-1.7B one-layer W8A16 QDQ attention artifact, compiled as QNN DLC, ran on the exact hosted Snapdragon X Elite CRD (Windows 11 ARM64). Inference returned finite hidden/new-K/new-V tensors bitwise identical to the prior S24/S25 NPU outputs on the same fixture; the profile attributed all 73 execution-detail rows to NPU. Across 100 component samples, nearest-rank p50/p95 was 1.433/1.606 ms; the 1.346 ms estimated-time field is the minimum. Hidden/K/V relative L2 versus original FP32 ONNX CPU was 1.351%/0.377%/0.650%, without a next-token quality tolerance. There is no complete 28-layer prefill/decode/KV-ring/sampling loop, resident Omni controller or complete-request timing. | Qualify a complete X Elite resident artifact/controller with state, token quality, memory admission/cancel and sustained power; [X Elite NPU raw evidence](../../../e2e_expansion_20260923/evidence/spark_xelite_full_attention/README.md). |
 | PC Snapdragon X Elite CRD / AI Hub | Qwen3-TTS 0.6B CustomVoice | COMPONENT_ONNX_CPU_VOCODER_PARITY_PROFILED; DIRECTML_RUNTIME_FAILED; TALKER_FP16_NUMERIC_FAIL; QNN_NPU_VOCODER_NUMERIC_MEASURED; PLACEMENT_PENDING; NOT_E2E_PROFILEABLE | The fresh ONNX vocoder inference on the exact Windows 11 ARM64 X Elite CRD requested CPU and returned an unsaturated waveform on the pinned fixture, matching local ONNX Runtime CPU with 8.09e-7 relative L2 and 121.84 dB SNR. Its 94-sample component profile had nearest-rank p50/p95 6.384/6.644 s and all 634 execution-detail rows on CPU; 5.917 s estimated time is the sample minimum. This is one component, without a listening-quality gate. Corrected DirectML inference and profile both failed at provider initialization with Windows error 80070057. A separate FP16 QNN DLC vocoder inference returned an unsaturated waveform at 1.138% relative L2 / 38.88 dB SNR versus local ONNX CPU on the same fixture; NPU node-placement profile is pending and no listening-quality tolerance is set. The 28-layer FP16 QNN talker step executed, yet changed CPU top-1 token 80 to 1720 with 8.57% logits relative L2 on a synthetic cache fixture; a separate ONNX CPU talker compile passed, with inference pending. The older TFLite vocoder target was rejected on Windows ARM; no complete stream exists. | Audit QNN vocoder placement and CPU talker control, investigate DirectML failure, then qualify complete local streaming, memory and sustained behavior; [vocoder](../../../e2e_expansion_20260924/evidence/qwen_tts_vocoder_qualcomm/README.md) and [talker](../../../e2e_expansion_20260924/evidence/qwen_tts_talker_qualcomm/README.md) evidence. |
 | PC Snapdragon X Elite CRD / AI Hub | Qwen3.8-27B | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. |
-| PC Snapdragon X Elite CRD / AI Hub | MiniCPM-o 4.5 | COMPONENT_SPEECH_HEAD_QNN_COMPILE_PASS; INFERENCE_PENDING; NOT_E2E_PROFILEABLE | The real 20-layer MiniCPM-o speech-head source compiled to FP16 QNN DLC for the exact X Elite CRD. Same-fixture NPU-requested inference was pending; no numerical output or placement profile exists. Thinker/encoders/vocoder and a device-local stateful Omni backend remain unqualified. | Audit the [X Elite speech-head inference](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md), then validate the complete multimodal pipeline, state, quality, memory and sustained behavior. |
+| PC Snapdragon X Elite CRD / AI Hub | MiniCPM-o 4.5 | COMPONENT_SPEECH_HEAD_NUMERIC_MEASURED; PLACEMENT_PENDING; NOT_E2E_PROFILEABLE | The real 20-layer MiniCPM-o speech-head source compiled to FP16 QNN DLC for the exact X Elite CRD. Same-fixture NPU-requested inference returned 41 finite tensors, preserved ONNX CPU top-1 speech token 1867, and had 0.3170% logits relative L2 and 1.221% maximum K/V relative L2. Actual NPU node assignment and timing await its placement profile. Thinker/encoders/vocoder and a device-local stateful Omni backend remain unqualified. | Audit the [X Elite speech-head placement profile](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md), then validate the complete multimodal pipeline, state, quality, memory and sustained behavior. |
 | PC Snapdragon X Elite CRD / AI Hub | InternVLA-A1 | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. |
 | Mobile Galaxy S25 / Snapdragon 8 Elite for Galaxy | Spark-X2.5 | NOT_E2E_PROFILEABLE | Fresh 2026-09-22 W8A16 full-attention component inference/profile passed; full1024 denotes one cache bucket. No full prefill/decode/KV/ring/sampling loop. | Implement and validate local 128-token generation, 512/1024 transitions, cancellation and resident memory. |
 | Mobile Galaxy S25 / Snapdragon 8 Elite for Galaxy | Qwen3-TTS 0.6B CustomVoice | TALKER_FP16_NUMERIC_FAIL; NOT_E2E_PROFILEABLE | Fresh FP16 predictor inference passed; its one output exactly matches the same compiled artifact/dataset reference. The real 28-layer FP16 QNN talker step executed on the synthetic cache fixture, but changed CPU top-1 speech token 80 to 412 with 8.99% logits relative L2 and maximum new-K/V relative L2 11.71%. This tested talker artifact is numerically unqualified; no complete co-resident NPU/GPU stream exists. | Improve and revalidate the [talker artifact](../../../e2e_expansion_20260924/evidence/qwen_tts_talker_qualcomm/README.md), then qualify the device-local predictor/vocoder pipeline, output quality, recovery and 30-minute thermal behavior. |
@@ -21,7 +21,7 @@ The identical SA/RB3 TFLite vocoder binary also matched original ONNX CPU within

 A separate [28-layer talker decode-step attempt](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/qwen_tts_talker_qualcomm/README.md) retains a synthetic fixed-cache fixture and CPU reference. Historical FP16 QNN context-binary inference executed on S24, S25, SA8775P and X Elite, but all four changed the CPU top-1 speech token with 8.57–9.43% logits relative L2. RB3 rejected that floating-point-input artifact. Explicit X Elite ONNX CPU and SA8775P TFLite CPU fallback compiles passed, with same-fixture inference pending. None advances M2 to numerical talker or full-stream qualification yet.

-For M4, a [MiniCPM-o 4.5 speech-head study](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md) recovered the historical 20-layer ONNX source and matching S24/S25 fixtures. Both FP16 context-binary outputs preserved the ONNX CPU top-1 speech token on one decode step, with logits relative L2 0.346%/0.360%. New FP16 QNN DLC compiles passed for SA8775P, X Elite and RB3. SA8775P same-fixture NPU-requested inference preserved CPU top-1 speech token 1867 with 0.3568% logits relative L2 and 1.225% maximum K/V relative L2; placement profiling is pending. X Elite and RB3 inferences remain pending. This is component C numerical evidence only, not thinker/encoder/vocoder or complete multimodal support.
+For M4, a [MiniCPM-o 4.5 speech-head study](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md) recovered the historical 20-layer ONNX source and matching S24/S25 fixtures. Both FP16 context-binary outputs preserved the ONNX CPU top-1 speech token on one decode step, with logits relative L2 0.346%/0.360%. New FP16 QNN DLC compiles passed for SA8775P, X Elite and RB3. SA8775P same-fixture NPU-requested inference preserved CPU top-1 speech token 1867 with 0.3568% logits relative L2 and 1.225% maximum K/V relative L2; placement profiling is pending. X Elite same-fixture inference also preserved token 1867 with 0.3170% logits relative L2; both placement profiles and RB3 inference remain pending. This is component C numerical evidence only, not thinker/encoder/vocoder or complete multimodal support.

 For M4, a [Galaxy S25 Cosmos encoder attempt](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/internvla_cosmos_qualcomm/README.md) repackages the real Place_Markpen FP32 ONNX export after retaining two CPU parity fixtures. FLOAT16 QNN DLC inference returned a latent within 0.0524% relative L2 of the source on one synthetic pattern. Injecting that output into the real policy on local CPU changed synthetic actions by 0.742% relative L2, without a task tolerance. NPU placement profile was pending, and the policy did not run on S25. This is component/action-sensitivity evidence, not device-local M4 qualification.