Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -10,32 +10,39 @@ tags:
|
|
| 10 |
---
|
| 11 |
# twinlitenet — ExecuTorch XNNPACK
|
| 12 |
|
| 13 |
-
`twinlitenet_xnnpack_fp32.pte` (1.8 MB, fp32, XNNPACK-delegated)
|
| 14 |
-
|
| 15 |
- **Source**: chequanghuy/TwinLiteNet (pretrained/best.pth)
|
| 16 |
- **License**: MIT
|
| 17 |
- **Input**: [[1, 3, 360, 640]] — RGB 0-1, 360x640
|
| 18 |
- **Output**: drivable area [1,2,360,640] + lane line [1,2,360,640]
|
| 19 |
|
| 20 |
-
##
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
|
| 22 |
-
|
|
|
|
|
|
|
|
|
|
| 23 |
|
| 24 |
| output | shape | max_abs_diff | corr |
|
| 25 |
|--------|-------|--------------|------|
|
| 26 |
-
| 0 | [1, 2, 360, 640] | 1.
|
| 27 |
-
| 1 | [1, 2, 360, 640] | 1.
|
| 28 |
|
| 29 |
-
|
| 30 |
-
ExecuTorch 33.7 ms vs torch eager 23.3 ms.
|
| 31 |
|
| 32 |
## Conversion
|
| 33 |
|
| 34 |
torch.export -> to_edge_transform_and_lower(XnnpackPartitioner) -> .pte
|
| 35 |
(conversion scripts: [executorch-models](https://github.com/john-rocky/executorch-models))
|
| 36 |
-
|
| 37 |
-
**Notes**: PReLU is excluded from XNNPACK delegation (XNNPACK PReLU segfaults at execute on macOS arm64 in executorch 1.4.0; minimal repro: a lone nn.PReLU(1)). PReLU runs on the portable kernel instead; outputs are bit-identical to the stock model.
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
---
|
| 41 |
-
Part of [executorch-models](https://github.com/john-rocky/executorch-models) — a verified .pte zoo for ExecuTorch. Conversion scripts and all models are indexed there.
|
|
|
|
| 10 |
---
|
| 11 |
# twinlitenet — ExecuTorch XNNPACK
|
| 12 |
|
|
|
|
|
|
|
| 13 |
- **Source**: chequanghuy/TwinLiteNet (pretrained/best.pth)
|
| 14 |
- **License**: MIT
|
| 15 |
- **Input**: [[1, 3, 360, 640]] — RGB 0-1, 360x640
|
| 16 |
- **Output**: drivable area [1,2,360,640] + lane line [1,2,360,640]
|
| 17 |
|
| 18 |
+
## Variants
|
| 19 |
+
|
| 20 |
+
All variants take and return fp32 tensors — swap the `.pte` file, keep your app code.
|
| 21 |
+
|
| 22 |
+
| precision | file | size (MB) | parity vs fp32 eager (worst corr) | Mac median (ms)* |
|
| 23 |
+
|-----------|------|-----------|------------------------------------|------------------|
|
| 24 |
+
| fp32 | `twinlitenet_xnnpack_fp32.pte` | 1.8 | 1.000000 | 32.5 |
|
| 25 |
+
|
| 26 |
+
\*Mac arm64, single process, median of 10 — a reference point for relative cost
|
| 27 |
+
only, not a device number (torch eager fp32 on the same machine: 23.9 ms).
|
| 28 |
+
|
| 29 |
+
### Precisions that did not earn a slot
|
| 30 |
+
|
| 31 |
+
- **fp16 is not shipped**: it comes out at 100% of the fp32 file (1.8 MB vs 1.8 MB), so it buys nothing. XNNPACK serializes convolution weights as fp32 no matter what dtype the graph carries, so on a conv-heavy model fp16 saves no disk and only adds cast operations. Reach for int8 here, not fp16.
|
| 32 |
|
| 33 |
+
## Verification (executorch 1.4.0, torch 2.13.0)
|
| 34 |
+
|
| 35 |
+
Parity is measured against the fp32 eager model on real image input; `corr` is
|
| 36 |
+
the correlation over all elements of each output tensor.
|
| 37 |
|
| 38 |
| output | shape | max_abs_diff | corr |
|
| 39 |
|--------|-------|--------------|------|
|
| 40 |
+
| 0 | [1, 2, 360, 640] | 1.240e-05 | 1.000000 |
|
| 41 |
+
| 1 | [1, 2, 360, 640] | 1.287e-05 | 1.000000 |
|
| 42 |
|
| 43 |
+
XNNPACK delegate coverage (fp32): 79.7% (177/222 ops); ops left on the portable kernels: `aten.gt.Scalar` x20, `aten.where.self` x20, `aten.avg_pool2d.default` x3, `aten.max.dim` x1, `aten.expand_copy.default` x1
|
|
|
|
| 44 |
|
| 45 |
## Conversion
|
| 46 |
|
| 47 |
torch.export -> to_edge_transform_and_lower(XnnpackPartitioner) -> .pte
|
| 48 |
(conversion scripts: [executorch-models](https://github.com/john-rocky/executorch-models))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|