Train and evaluate RuView models using Wi-Fi measurements for pose estimation, retrieval, and adaptation to different environments. Follow laptop or GCloud GPU workflows, then package and publish trained models to Hugging Face.
한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.
Train RuView models — camera-free WiFlow pose (10 sensor signals, no labels), camera-supervised pose (MediaPipe + ESP32 CSI → 92.9% PCK@20, ADR-079), RuVector contrastive embeddings (AETHER, ADR-024), domain generalization (MERIDIAN, ADR-027), local SNN environment adaptation, plus GPU training on GCloud and Hugging Face publishing. Use when building, fine-tuning, evaluating, or shipping a model.
설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.
RuView trains several kinds of model. Pick the track that matches the goal; all of them run on a laptop, with an optional GPU path.
Trains 17-keypoint pose from 10 sensor signals. Fast, fully unsupervised, modest accuracy.
cd v2
# Pretrain on raw CSI (contrastive)
cargo run -p wifi-densepose-sensing-server -- --pretrain --dataset data/csi/ --pretrain-epochs 50
# Train pose head, save an RVF artifact
cargo run -p wifi-densepose-sensing-server -- --train --dataset data/mmfi/ --epochs 100 --save-rvf model.rvf
~84 s on an M4 Pro. Benchmarks: node scripts/benchmark-wiflow.js, eval: node scripts/eval-wiflow.js.
Uses a webcam + MediaPipe as ground truth, paired with ESP32 CSI. ~19 min on a laptop.
# 1. Collect paired data (camera + CSI)
python scripts/collect-ground-truth.py # MediaPipe pose landmarks
python scripts/collect-training-data.py # CSI capture, time-synced
node scripts/align-ground-truth.js # align camera ↔ CSI timestamps
# 2. Train (the camera-supervised path through the sensing-server / train crate)
cd v2
cargo run -p wifi-densepose-sensing-server -- --train --dataset data/paired/ --epochs <N> --save-rvf model.rvf
# 3. Evaluate
cd .. && node scripts/eval-wiflow.js # reports PCK@20
Requires data/pose_landmarker_lite.task (MediaPipe model). See docs/adr/ADR-079-camera-ground-truth-training.md.
CSI subcarrier amplitude/phase → embeddings for re-ID and retrieval (171K emb/s on M4 Pro). Driven by wifi-densepose-train + wifi-densepose-ruvector (RuVector v2.0.4). Spectrogram embeddings: ADR-076.
cd v2
cargo check -p wifi-densepose-train --no-default-features # sanity
cargo run -p wifi-densepose-sensing-server -- --model model.rvf --embed
cargo run -p wifi-densepose-sensing-server -- --model model.rvf --build-index env
Make a model transfer across environments without retraining. Configured through the training pipeline's domain-generalization options; see ADR-027 and wifi-densepose-train + ruview_metrics.
Spiking neural network that adapts to a new room in <30 s, on-device or on a Cognitum Seed:
node scripts/snn-csi-processor.js --port 5006
See docs/tutorials/cognitum-seed-pretraining.md, ADR-084/085 (RaBitQ similarity sensor), ADR-086 (edge novelty gate).
Project cognitum-20260110 has L4 / A100 / H100 quota.
gcloud auth login
gcloud config set project cognitum-20260110
bash scripts/gcloud-train.sh --dry-run # smoke test, synthetic data
bash scripts/gcloud-train.sh --gpu l4 --hours 2 # prototyping
bash scripts/gcloud-train.sh --gpu a100 --config scripts/training-config-sweep.json
bash scripts/gcloud-train.sh --sweep # full hyperparameter sweep
# VM is auto-deleted after training unless --keep-vm. Cost: L4 ~$0.80/hr, A100 40GB ~$3.60/hr.
Local Mac training: bash scripts/mac-mini-train.sh. Model benchmark: python scripts/benchmark-model.py.
python scripts/publish-huggingface.py # or: bash scripts/publish-huggingface.sh
Pushes the RVF artifact + card to Hugging Face. See docs/huggingface/.
| Path | Contents |
|---|---|
data/recordings/ | Raw CSI captures (*.csi.jsonl), overnight runs |
data/csi/ | CSI datasets for pretraining |
data/mmfi/ | MM-Fi dataset (ADR-015) |
data/paired/ | Camera ↔ CSI paired samples (ADR-079) |
data/ground-truth/ | MediaPipe pose landmarks |
data/pose_landmarker_lite.task | MediaPipe model file |
models/ | Trained artifacts |
Record more data: python scripts/record-csi-udp.py (UDP CSI capture from a live node).
cd v2 && cargo test --workspace --no-default-features # 1,400+ pass, 0 fail
cd .. && python archive/v1/data/proof/verify.py # VERDICT: PASS
Then hand off to ruview-verify for the witness bundle.
wifi-densepose-train, wifi-densepose-nn, wifi-densepose-ruvector, wifi-densepose-sensing-serverscripts/gcloud-train.sh, mac-mini-train.sh, benchmark-wiflow.js, eval-wiflow.js, benchmark-model.py아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.
로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.
아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.
Build advanced AgentDB workflows with database synchronization, multiple databases, and search that combines meaning with attribute filters. Covers similarity measures, diverse search results, and operational patterns for distributed AI applications.
한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.
Create AgentDB learning plugins that use recorded experience to improve agent decisions. Choose a learning method, store states, actions, and rewards, train models, and evaluate suggested actions through CLI and API workflows.
한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.
Build AI agents that retain conversation history, user preferences, and useful interaction patterns with AgentDB. The skill covers persistent storage, meaning memories survive between sessions, plus retrieval, context management, and ReasoningBank integration.
한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.
Tune AgentDB memory use and search speed through quantization, which stores vectors more compactly, search indexes, caching, and batch operations. Includes configuration examples and benchmarks to help balance accuracy, speed, and storage as datasets grow.
한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.
Build document search with AgentDB that finds related content by meaning. Store documents as numerical representations, combine similarity search with metadata filters, and retrieve relevant passages for systems that generate answers using source material.
한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.
Automate browser tasks with agent-browser, using snapshots that label page elements for the agent. Handle login flows, fill and submit forms, extract page information, and capture screenshots through a repeatable navigation and interaction workflow.
한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.