Tenzro Train.

Decentralized training. Rust protocol layer (tenzro-training) plus a Python reference trainer (PyTorch FSDP2 + Hivemind).

STATUS: Testnet
CRATE: tenzro-training
STABILITY: Phase 1
REFERENCE: TRAIN.md

Architecture

Two layers. The Rust crate owns the protocol: OuterGradient, Fragment, SyncRound, aggregation rules, the outer optimizer, the syncer state machine, on-chain commitments, fraud proofs, RPC, CLI. The Python trainer owns the inner loop: PyTorch FSDP2 + Hivemind + safetensors with per-modality adapters.

Aggregation

Mean, TrimmedMean, CoordinateMedian, Krum. Tier-gated: Open admits Mean only; Verified and Confidential admit all four. Nesterov SGD outer optimizer.

Witness committee

k-of-N witness committee with idempotent on-chain finalize. No-endorsement certificate carries the run forward when a quorum cannot be assembled inside grace_window_ms.

Adapters

timeseries   TimesFM-class 200M
language     Qwen 3 0.6B default, swappable via metadata
vision       timm ViT-B/16 default, swappable via metadata

CLI

tenzro train post-task
tenzro train list-runs
tenzro train get-run --id <run_id>
tenzro train enroll-trainer
tenzro train submit-gradient
tenzro train finalize-round

← All docs