Compiled BPU models for RDK S100/S100P

.gitattributes

@@ -33,3 +33,5 @@ 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
+las2_m_crop_nashm.hbm filter=lfs diff=lfs merge=lfs -text
+las2_m_int16_nashm.hbm filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,86 @@
+---
+license: mit
+library_name: bcdl
+tags:
+- rdk-s100
+- rdk-s100p
+- d-robotics
+- bpu
+- hbm
+- depth-estimation
+- stereo
+---
+
+# Lite Any Stereo V2 for RDK S100/S100P — zero-shot stereo disparity
+
+Compiled BPU models (`.hbm`) for the **D-Robotics RDK S100 / S100P**, ready to
+load — no ONNX export, no calibration, no `hb_compile`. Built and measured with
+[**BCDL**](https://github.com/ruisv/bcdl), a C++17 inference and media library
+for the RDK S-series with Python bindings.
+
+Upstream: [Lite Any Stereo V2](https://github.com/Tinker-Twins/LiteAnyStereo) (M)
+
+> [!TIP]
+> **Redistributable, including commercially.** The licence chain was checked on
+> the code, the pretrained weights it started from, and the data it was trained
+> on — all three, because a permissive repository badge does not by itself say
+> anything about the weights. See [Licence](#licence).
+
+## Files
+
+| file | what it is |
+|---|---|
+| `las2_m_crop_nashm.hbm` | centre-crop-fit calibration, 640x480 — 38.8 MB |
+| `las2_m_int16_nashm.hbm` | resize-fit calibration, 640x480 — 39.3 MB |
+
+## Measured on an S100P
+
+| stage | latency | throughput |
+|---|---|---|
+| either build | 13.70 ms | 73 FPS |
+
+`hrt_model_exec perf`, one thread, minimum of three runs, on a board first gated
+against its own previously recorded numbers. **BPU time only** — CPU
+pre/post-processing is on top and is listed per task in BCDL's
+[benchmark results](https://github.com/ruisv/bcdl/blob/main/benchmarks/RESULTS.md).
+
+## Use it
+
+```bash
+conda install -c https://mirrors.ruis.ai/conda -c conda-forge bcdl
+```
+
+```python
+import bcdl
+engine = bcdl.Engine("las2_m_crop_nashm.hbm")
+print(engine.input_shape(0), engine.output_shape(0))
+```
+
+Each task has a decoder in BCDL that turns those raw outputs into boxes,
+keypoints, masks, disparity or text — see the
+[Python API](https://github.com/ruisv/bcdl/blob/main/docs/API.md)
+([中文](https://github.com/ruisv/bcdl/blob/main/docs/API.zh.md)).
+
+## What to know before deploying
+
+**Two calibration modes, and you must match the one you letterbox with.** `_crop`
+was calibrated on centre-cropped pairs and `_int16` on resize-fitted ones; feed a
+build the geometry it was not calibrated for and the disparity degrades quietly
+rather than failing.
+
+All-2D and feed-forward — no 3-D cost volume — which is exactly why it runs at
+73 FPS here. Cost-volume stereo of comparable quality took ~40 s/frame on this
+part and was abandoned.
+
+## Licence
+
+MIT on the code and the weights. One thing worth knowing rather than a term passed down: the release was distilled from FoundationStereo, whose NVIDIA licence is research-only. That is upstream's compliance question, but look at it before a commercial build.
+
+**BCDL itself is Apache-2.0 and is unrelated to these terms** — it is a
+general-purpose runtime that loads any `.hbm`. The licence above constrains
+*these weights and this compiled artefact*.
+
+The conversion recipe — ONNX export, calibration, `hb_compile` config and the
+acceptance numbers — is public in
+[**bcdl-model-zoo**](https://github.com/ruisv/bcdl-model-zoo), so this build can
+be reproduced or retargeted rather than taken on trust.

las2_m_crop_nashm.hbm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:323f567365f70ab1ea59e42364560293a8b8a2eb057df52e02fa8d67d1b7d06d
+size 40700800

las2_m_int16_nashm.hbm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78caecfa9d31a954dbd77c45fd2acd06e98fa5efe779e2a064634ba4e01e7908
+size 41197264