Compatibility registry
Which models run on which devices, on which accelerator, how fast, and with which fix. Every node benchmarks the models it runs and reports the result to its hub: works, works with tweaks (and the tweak), or fails (and why). The hub's router uses the same data.
11 results for 7 models on 2 devices (2 SoCs). So far they come from a Pixel 7 Pro and a TCL Google TV.
All results
Latency is the runner time on the accelerator (p50; no queueing or model load) unless the note says otherwise. Filters are kept in the URL, so a filtered view can be shared.
11 of 11 results
| clip-l14-336 image-embedding ~1450 ms per image (two views: letterbox + centre crop); fp16 weights, 582 MB; 1.65 s round trip from the Mac | Google Pixel 7 Pro Google Tensor G2 | Google Tensor G2 | GPU litert-gpu | Works | 1.45 s |
| efficientnet-lite0-int8 image-classification invoke time | TCL Smart TV Pro MediaTek MT9689 | MediaTek MT9689 | NPU nnapi:mtk-mdla_shim | Works | 6 ms |
| efficientnet-lite2-int8 image-classification invoke time | TCL Smart TV Pro MediaTek MT9689 | MediaTek MT9689 | NPU nnapi:mtk-mdla_shim | Works | 9.4 ms |
| mobilenet-v2-int8 image-classification | Google Pixel 7 Pro Google Tensor G2 | Google Tensor G2 | CPU cpu | Works | 7 ms |
| mobilenet-v2-int8 image-classification | Google Pixel 7 Pro Google Tensor G2 | Google Tensor G2 | NPU nnapi | Works | 0.78 ms |
| mobilenet-v2-int8 image-classification 4.4–5.0 ms invoke | TCL Smart TV Pro MediaTek MT9689 | MediaTek MT9689 | NPU nnapi:mtk-mdla_shim | Works | 4.4 ms p95 5 ms |
| efficientnet-lite4-int8 image-classification Tweak: Load the model from a memory-mapped buffer, not a path on shared storage: SELinux blocks the NN HAL (mtk_hal_neuralnetworks) from reading FUSE files, and NNAPI then silently falls back to its CPU reference path (~9 s per image). | TCL Smart TV Pro MediaTek MT9689 | MediaTek MT9689 | NPU nnapi:mtk-mdla_shim | Works with tweaks | 22 ms |
| mobilenet-v2-fp32 image-classification Tweak: Enable allow_fp16 on the NNAPI delegate (setAllowFp16(true)). | TCL Smart TV Pro MediaTek MT9689 | MediaTek MT9689 | NPU nnapi:mtk-mdla_shim | Works with tweaks | 11.2 ms |
| clip-vit-b16 image-embedding fp32, 330 MB: out of memory while loading; the TV rebooted. | TCL Smart TV Pro MediaTek MT9689 | MediaTek MT9689 | Unknown n/a | Fails | – |
| clip-vit-b16 image-embedding The GPU delegate does not support the graph (unsupported CONCATENATION). | TCL Smart TV Pro MediaTek MT9689 | MediaTek MT9689 | GPU gpu (Mali-G52 delegate) | Fails | – |
| mobilenet-v2-fp32 image-classification The MDLA does not run fp32 graphs; without allow_fp16 NNAPI rejects the model for the NPU. | TCL Smart TV Pro MediaTek MT9689 | MediaTek MT9689 | NPU nnapi:mtk-mdla_shim | Fails | – |
Model × device matrix
Best result per pair, from the 2026-11-18 snapshot.
| Model | TCL Smart TV Pro | Google Pixel 7 Pro |
|---|---|---|
| clip-l14-336 | not tested | Works1.45 s |
| clip-vit-b16 | Fails– | not tested |
| efficientnet-lite0-int8 | Works6 ms | not tested |
| efficientnet-lite2-int8 | Works9.4 ms | not tested |
| efficientnet-lite4-int8 | Works with tweaks22 ms | not tested |
| mobilenet-v2-fp32 | Works with tweaks11.2 ms | not tested |
| mobilenet-v2-int8 | Works4.4 ms | Works0.78 ms |
How results get here
After a model's first five runs on a node, the node sends p50/p95 runner times and a status to its hub: works (loaded on the manifest's first-choice accelerator), works with tweaks (it fell back; the note says why) or fails (the load error). The owner's measured TV and Pixel results seed the registry. Public submissions from other hubs are planned.