Dashboard →
Training Monitor · v1.0.0

Arc Vigil —
instability detected
before your loss curve moves.

Real-time training stability monitoring. Identifies which layer is failing. Intervenes automatically. 100% detection, 0% false positives, 90% auto-recovery across 30 seeds and 6 architectures.

Get started free → GitHub PyPI
$pip install arc-vigil

Quickstart

One call before your forward pass.

Drop into your existing training loop. No changes to your model, optimizer, or data pipeline.

Usage
from arc_vigil import BendexMonitor, BendexConfig, BendexIntervention

monitor      = BendexMonitor(model, config=BendexConfig())
intervention = BendexIntervention(model, optimizer)

for step, batch in enumerate(dataloader):
    event = monitor.observe(step)        # detects instability
    intervention.step(event, step)         # intervenes if needed

    loss = model(batch)
    loss.backward()
    intervention.apply_grad_clip()
    optimizer.step()
    optimizer.zero_grad()

What it does

Detect. Attribute. Intervene.

Three things no other training monitor does simultaneously.

78
Steps early warning

Early detection

Mean 78-step lead time before instability shows up in any standard metric. When Arc Vigil fires, your loss curve hasn't moved yet.

Yes
Module attribution

Knows what's failing

Identifies the specific module that deviated first. No other tool tells you where the problem is — they only tell you something is wrong.

90%
Auto-recovery rate

Automatic intervention

Fires a three-phase intervention automatically when a trigger fires. No other tool intervenes — they only alert.


Benchmark

30 seeds. 6 architectures.

Validated against every standard instability detection method.

MethodDetectionFalse PositivesRecoveryAttributes ModuleAuto-Intervenes
Arc Vigil100%0%90%YesYes
Loss spike100%80%0%NoNo
Gradient norm90%50%0%NoNo
Patience100%50%0%NoNo

30-seed benchmark across 6 architectures: MLP, CNN, ViT, DistilBERT, GPT-2, ResNet-50 + LR spike stress test.


Pricing

Free. No credit card.

Arc Vigil is fully open source. Full access on PyPI and GitHub with no limits or sign-up required.

Arc Vigil
$0
Forever · no credit card
Get started →