M-LSD-tiny LiteRT fp16 (fully-GPU, Pixel 8a corr 0.997, 2ms, 1.4MB)

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+samples/sample.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,49 @@
+---
+license: apache-2.0
+library_name: LiteRT
+pipeline_tag: image-segmentation
+tags: [litert, tflite, on-device, android, gpu, line-segment-detection, mlsd, wireframe, mobilenetv2]
+base_model: navervision/mlsd
+---
+
+# M-LSD-tiny — LiteRT (on-device line segment detection, fully-GPU)
+
+[M-LSD](https://github.com/navervision/mlsd) (NAVER, AAAI 2022) light-weight real-time **line segment
+detection**, converted to **LiteRT** and running **fully on the `CompiledModel` GPU** (ML Drift) on Android.
+Detects straight line segments — building edges, document borders, wireframes, room layout. The **tiny**
+variant (MobileNetV2 backbone, 0.62M params) is **1.4 MB** in fp16.
+
+![M-LSD — input | detected line segments (on-device LiteRT GPU)](samples/sample.png)
+
+## On-device (Pixel 8a, Tensor G3 — verified)
+
+| | |
+|---|---|
+| nodes on GPU | **99 / 99** LITERT_CL (full residency) |
+| inference | **~2 ms** (512×512) |
+| size | **1.4 MB** (fp16) |
+| accuracy | device-vs-PyTorch corr **0.997** (127 vs 128 lines decoded) |
+
+```
+image[1,4,512,512] (RGB + ones channel, scaled to [-1,1]) →[GPU: MobileNetV2 U-Net]→ tpMap[1,9,256,256]
+```
+
+The output is a "TP map": channel 0 = line-center heatmap, channels 1–4 = start/end displacement. The decode
+(sigmoid + 3×3 NMS over centers, displacement → endpoints, ×2) runs on the host.
+
+## How it converts (litert-torch)
+
+Pure CNN encoder-decoder. A single re-authoring: the decoder's `F.interpolate(bilinear, align_corners=True)`
+→ **`align_corners=False`** (the Mali delegate bans `align_corners=True` + half-pixel). MobileNetV2 has no
+max-pool (strided convs → no `PADV2`), and the upsample is `RESIZE_BILINEAR`, not a transposed conv → fully
+GPU-clean. Result: banned ops NONE, all tensors ≤4D, tflite-vs-torch corr **1.0**, device-vs-torch corr **0.997**.
+
+## Preprocessing & decode
+
+Resize to 512×512, append a 4th channel of ones, scale `(x/127.5) - 1`, NCHW. Decode: sigmoid the center map,
+3×3 max NMS, threshold (0.10), displacement → endpoints, filter by length, ×2 to 512-space.
+
+## License
+
+[Apache-2.0](https://github.com/navervision/mlsd/blob/main/LICENSE). Upstream:
+[navervision/mlsd](https://github.com/navervision/mlsd); PyTorch port [lhwcv/mlsd_pytorch](https://github.com/lhwcv/mlsd_pytorch).

mlsd_fp16.tflite

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5da930ae4a0dacd7725b0126a1d1481eaac3b3f537000f7f981461d1b1305f2
+size 1389952