Add official MicroDuck walking ONNX with source and local validation

BEST_alpha_walking.json

@@ -0,0 +1,7 @@
+{
+  "source": "Official Pollen Robotics pretrained walking policy; downloaded, not trained locally",
+  "source_url": "https://huggingface.co/spaces/pollen-robotics/microduck-simulator/resolve/183f99a40bd7308da3e848de961ed32bb02624a5/app/public/policies/BEST_alpha_walking.onnx",
+  "upstream_revision": "183f99a40bd7308da3e848de961ed32bb02624a5",
+  "tags": ["official", "walk"],
+  "note": "Local CPU inference test only. See validation_report.json. No source checkpoint parity, physics rollout, or hardware test performed."
+}

BEST_alpha_walking.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e36332d383997d51401897734cd3e79cf5038406feddb18b4d57ecfb141daa6c
+size 793705

README.md

@@ -0,0 +1,56 @@
+---
+tags:
+- onnx
+- robotics
+- reinforcement-learning
+- microduck
+---
+
+# MicroDuck walking policy (official pretrained copy)
+
+This repository contains `BEST_alpha_walking.onnx`, an unchanged copy of the
+pretrained walking policy distributed by **Pollen Robotics**. It was downloaded
+and smoke-tested locally; it was **not trained or fine-tuned by this uploader**.
+
+## Source
+
+- Original project: [pollen-robotics/microduck-simulator](https://huggingface.co/spaces/pollen-robotics/microduck-simulator)
+- Pinned revision: `183f99a40bd7308da3e848de961ed32bb02624a5`
+- [Original model file](https://huggingface.co/spaces/pollen-robotics/microduck-simulator/blob/183f99a40bd7308da3e848de961ed32bb02624a5/app/public/policies/BEST_alpha_walking.onnx)
+- SHA-256: `e36332d383997d51401897734cd3e79cf5038406feddb18b4d57ecfb141daa6c`
+
+No new license is granted by this copy. Consult the original project and its
+authors for applicable model usage and redistribution terms.
+
+## Interface and local validation
+
+- Input: `obs`, float32, shape `[1, 61]`.
+- Output: `actions`, float32, shape `[1, 14]`.
+- Observation normalization is included in the graph.
+- ONNX graph checker passed.
+- All 512 wide-distribution random input samples produced finite outputs.
+- Local CPU inference averaged approximately 0.027 ms per call over 1000 calls.
+  Timing is specific to the test machine.
+
+See `validation_report.json` for recorded results. No source-checkpoint numerical
+parity comparison, physics rollout, or real-hardware test was performed.
+
+## Minimal inference example
+
+Install `numpy` and `onnxruntime`, download the ONNX file, then run:
+
+```python
+import numpy as np
+import onnxruntime as ort
+
+session = ort.InferenceSession(
+    "BEST_alpha_walking.onnx", providers=["CPUExecutionProvider"]
+)
+obs = np.zeros((1, 61), dtype=np.float32)
+actions = session.run(["actions"], {"obs": obs})[0]
+print(actions.shape)  # (1, 14)
+```
+
+Zero observations here are synthetic smoke-test data. A robot rollout needs the
+upstream observation layout, control loop, and action processing; these raw
+outputs are not direct hardware commands.

validation_report.json

@@ -0,0 +1,44 @@
+{
+  "tested_at": "2026-09-21T09:00:51.513512+00:00",
+  "model": "BEST_alpha_walking.onnx",
+  "size_bytes": 793705,
+  "sha256": "e36332d383997d51401897734cd3e79cf5038406feddb18b4d57ecfb141daa6c",
+  "onnx_version": "1.16.0",
+  "onnxruntime_version": "1.17.1",
+  "providers": [
+    "CPUExecutionProvider"
+  ],
+  "input_shape": [
+    1,
+    61
+  ],
+  "output_shape": [
+    1,
+    14
+  ],
+  "onnx_checker": "passed",
+  "wide_random_samples": 512,
+  "wide_random_test": "passed",
+  "zero_observation_actions": [
+    [
+      -0.27203676104545593,
+      0.33868545293807983,
+      0.29769596457481384,
+      -0.11047147959470749,
+      0.07768982648849487,
+      0.07041269540786743,
+      0.016396326944231987,
+      -0.5535610318183899,
+      0.26069194078445435,
+      -0.07948483526706696,
+      0.05550389364361763,
+      0.26101064682006836,
+      -0.20797322690486908,
+      -0.1478530466556549
+    ]
+  ],
+  "mean_inference_ms_1000_calls": 0.026622299919836223,
+  "checkpoint_parity": "not tested: source checkpoint unavailable",
+  "physics_rollout": "not tested",
+  "hardware_test": "not tested"
+}