Add ONNX exports (7 stages, 1.4GB) for Jetson Nano inference

README.md

@@ -8,17 +8,60 @@ tags:
   - imu
   - heart-rate
   - edge-deployment
+  - onnx
+  - jetson-nano
 license: mit
 ---
 
-# SISA-RoutineGuard v5b
+# SISA-RoutineGuard v5b — ONNX Edition
 
-**노인 일상 패턴 이상 감지 시스템** (Galaxy Watch + Jetson Orin Nano)
+**노인 일상 패턴 이상 감지** (Galaxy Watch + Jetson Orin Nano Super)  
+**ONNX exports** for fast Jetson inference (TensorRT compatible).
 
-4-tier hierarchical anomaly detector with **SISA backbone** (SSM-Informed Softmax Attention).
-Train/val split + early stopping 적용한 production ckpt.
+## 📦 ONNX Models (Jetson deploy)
 
-## 모델 사이즈 — **444.66M params** (목표 250-370M 초과)
+| File | Size | Stage | Input | Output |
+|---|---:|---|---|---|
+| **patch_encoder.onnx** | 0.9 MB | 1 (Phone) | acc [6, 250, 3] | tokens [6, 256] |
+| **minute_encoder.onnx** | 192.5 MB | 2 | patch_tokens [B, 6, 256] | minute_embed [B, 768] |
+| **feature_adapter.onnx** | 2.7 MB | 2 | feature [B, 12] | embed [B, 768] |
+| **hourslot_encoder.onnx** | 570.9 MB | 3 | minute_embeds [B, 60, 768] | slot [B, 1024], slot_minutes [B, 60, 1024] |
+| **history_encoder.onnx** | 567.7 MB | 4 | slot_embeds [B, 90, 1024] + meta | history_embeds [B, 90, 1024] |
+| **vitalguard.onnx** | 48.2 MB | 6 | vital_features [B, 60, 5] + hrv | hr_residual_z + trend + context |
+| **output_heads.onnx** | 6.3 MB | — | cls_pooled [B, 1024] | anomaly + reason + confidence |
+
+> **QueryRefiner ONNX** 는 cross-attention shape 복잡으로 미지원 (PyTorch ckpt만).
+
+**Total ONNX: ~1.4 GB**
+
+## 🚀 Jetson Inference Example
+
+```python
+import onnxruntime as ort
+import numpy as np
+
+# TensorRT EP (Jetson에서 자동 가속)
+providers = [
+    'TensorrtExecutionProvider',  # Jetson Orin Nano TensorRT
+    'CUDAExecutionProvider',       # fallback CUDA
+    'CPUExecutionProvider',        # last resort
+]
+
+# 1. Phone 측 PatchEncoder (Phone ONNX)
+phone_sess = ort.InferenceSession("patch_encoder.onnx", providers=['CPUExecutionProvider'])
+patches = phone_sess.run(None, {"acc": acc_array})  # [6, 250, 3] → [6, 256]
+
+# 2. Jetson Pipeline
+me_sess = ort.InferenceSession("minute_encoder.onnx", providers=providers)
+minute = me_sess.run(None, {"patch_tokens": patches.reshape(1, 6, 256)})[0]
+
+hs_sess = ort.InferenceSession("hourslot_encoder.onnx", providers=providers)
+slot, slot_minutes = hs_sess.run(None, {"minute_embeds": minute_batch})  # [B, 60, 768]
+
+# ... history, query, heads chain
+```
+
+## 🎯 모델 사이즈 — **444.66M params**
 
 | Component | Params | Deploy |
 |---|---:|---|
@@ -32,9 +75,9 @@ Train/val split + early stopping 적용한 production ckpt.
 | OutputHeads | 1.58M | Jetson |
 | **Total** | **444.66M** | — |
 
-## 성능 (OOD: HAR-70+ 노인 70-95세, 학습 X)
+## 📊 성능 (OOD: HAR-70+ 노인 70-95세, 학습 X)
 
-| 시나리오 | Mean | AUC | Reason acc |
+| 시나리오 | Score | AUC | Reason |
 |---|---:|---:|---:|
 | Normal | 0.0004 | — | — |
 | walk_missing | 0.9995 | **1.000** | 0% |
@@ -43,75 +86,30 @@ Train/val split + early stopping 적용한 production ckpt.
 | activity_drop | 0.7145 | 0.950 | 28% |
 | **Overall** | — | **0.987** | — |
 
-## Train/Val Split + Early Stopping
+## 🔧 Train/Val Split + Early Stopping
 
-- Train/Val: 80/20 split (seed=42)
-- Train samples: 3,277 / Val: 819
+- Train/Val: 80/20 (3,277 / 819 samples)
 - **Best epoch = 1, val_loss = 0.2814** (saved)
 - Early stop at epoch 13 (patience=10)
+- Overfitting 방지 검증됨
 
-## 학습 데이터
-
-| Dataset | Subjects | Duration | Stage |
-|---|---:|---|---|
-| CAPTURE-24 | 151명 | 24h wrist 100Hz | 1, 2 |
-| ArWISE V3 | 10명 | 9일, 76일 raw | 1, 3, 4, 5 |
-| PPG-DaLiA | 15명 | 2.5h wrist | 6 |
-| WESAD | 15명 | 1.7h wrist+chest | 6 |
-| MHEALTH | 10명 | 53m 23ch | 6 |
-| **HAR-70+** | **18명 70-95세** | **테스트만 (OOD)** | — |
+## 📁 PyTorch checkpoints (학습 reproducibility용)
 
-## Inference
-
-```python
-import torch
-from src.models.full_model import SISARoutineGuard
-
-model = SISARoutineGuard().cuda().eval()
-state = torch.load("stage5_full.pt", map_location="cpu")["model"]
-model.load_state_dict(state, strict=False)
-vg = torch.load("stage6_vitalguard.pt", map_location="cpu")["model"]
-model.vitalguard.load_state_dict(vg, strict=False)
-
-# forward_replay (90 history × 60 min + 3 today × 60 min)
-out = model.forward_replay(history_features_norm, today_features_norm,
-                            day_offset, slot_pos, day_type,
-                            history_mask, today_mask)
-# 448 ms / batch=2 on RTX 4090
-```
-
-## ONNX (Phone deploy)
-```python
-import onnxruntime as ort
-sess = ort.InferenceSession("patch_encoder.onnx", providers=["CPUExecutionProvider"])
-out = sess.run(None, {"acc": acc_array})  # [6, 250, 3] → [6, 256]
-```
-
-## Files
-| File | Size | Stage |
+| File | Size | Purpose |
 |---|---:|---|
-| stage1_patch.pt | 879 KB | 1 |
-| stage2_minute.pt | 192 MB | 2 |
-| stage2_adapter.pt | 2.7 MB | 2 |
-| stage3_hourslot.pt | 570 MB | 3 |
-| stage4_history.pt | 567 MB | 4 |
-| stage4_refiner.pt | 390 MB | 4 |
-| **stage5_full.pt** | **1.78 GB** | **5 (val-split best)** |
-| stage6_vitalguard.pt | 48 MB | 6 |
-| patch_encoder.onnx | 880 KB | Phone |
+| stage1_patch.pt | 879 KB | PatchEncoder (also ONNX) |
+| stage2_minute.pt | 192 MB | MinuteEncoder (also ONNX) |
+| stage2_adapter.pt | 2.7 MB | FeatureAdapter (also ONNX) |
+| stage3_hourslot.pt | 570 MB | HourSlotEncoder (also ONNX) |
+| stage4_history.pt | 567 MB | HistoryEncoder (also ONNX) |
+| stage4_refiner.pt | 390 MB | QueryRefiner (PyTorch only) |
+| stage5_full.pt | 1.78 GB | Full model (val-split best) |
+| stage6_vitalguard.pt | 48 MB | VitalGuard (also ONNX) |
 | normalizer.pkl | 405 B | Feature normalizer |
 
-## 한계
-
-1. 합성 anomaly만 학습 — 진짜 노인 이상 (낙상, 치매) 미검증
-2. Reason multi-class 1~2개만 정확 (routine_time_shift 100%, 나머지 28% 이하)
-3. VitalGuard ground-truth HR 평가 미실시
-
-## 코드
-https://github.com/tlstngud/sisa-routineguard (private)
-- PRESENTATION.md 상세 발표 자료 포함
+## 학습 데이터
+CAPTURE-24 (151) + ArWISE V3 (10명/76일) + PPG-DaLiA (15) + WESAD (15) + MHEALTH (10).
+**HAR-70+ (18명 70-95세)는 OOD 평가에만 사용.**
 
-## Reference
-- Plan v1.4 (강원대 SUNRISE 연구실 캡스톤)
-- CAPTURE-24: Walmsley 2021 (DOI 10.5287/bodleian:NGx0JOMP5)
-- ArWISE V3: CASAS / Diane Cook (Zenodo 15803341)
+## Code
+https://github.com/tlstngud/sisa-routineguard (PRESENTATION.md 포함)

onnx/feature_adapter.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e04ac945f372bd85de13c354284775abda293ccad5d3b91724032f4645ac58d9
+size 2655392

onnx/history_encoder.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:841a56105c6b7f5d13853569210557fdd54eda1472d24a824a137abc6ae06ea0
+size 567704068

onnx/hourslot_encoder.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:423c1a986934013c735bdcc52085ee9f5991dfc74a6648bd31193785de40583d
+size 570882636

onnx/minute_encoder.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d829aa866bcc5b098fb53b55245254f11ac982f8a18d0824630f4ff4aee70b81
+size 192494083

onnx/output_heads.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a0fa889488acb6182adc8a477a5e411d9d241e568d1591fdbac28ee052610b8
+size 6315889

onnx/patch_encoder.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27a63803c361a19bfc991ec55d843c057972af26e3cdb52cd7ea776e63318b20
+size 880614

onnx/vitalguard.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de948c8c0e3808b632b77129677f4805b0a21529fd7207d732edc1c231468b85
+size 48198043