Files
rgriffith e843528d8b Phase 6: ELO Tracking & Model Tournament (#70)
- Create ml/eval/elo_tracker.py:
  - EloTracker class with standard ELO formula (K=32, start=1200)
  - EloResult dataclass for tournament output
  - EloHistory class for JSON persistence (data/elo_history.json)
  - run_elo_tournament(): round-robin ELO-rated matches using
    existing run_match() from tournament.py
  - run_generation_tournament(): discovers gen-*/best_policy.pt
    checkpoints and runs cross-generation ELO tournament
  - CLI entry point: python -m ml.eval.elo_tracker

- Add ELO API endpoints to routes/ml.py:
  - POST /api/ml/elo/tournament — start background ELO tournament
  - GET  /api/ml/elo/tournament/status — poll progress
  - GET  /api/ml/elo/history — load ELO history from disk

- Integrate ELO convergence in distillation_loop.py:
  - New config fields: elo_convergence_enabled (default False),
    elo_convergence_threshold (20.0), elo_tournament_episodes (30)
  - New field: GenerationRecord.elo_rating
  - _run_elo_mini_tournament() runs after quality gate pass
  - _check_convergence() checks ELO delta when enabled

- Add ELO UI to distillation page:
  - ELO Ratings panel with canvas multi-line chart showing per-model
    ELO progression across generations
  - Run Tournament button with progress indicator
  - ELO leaderboard table with rank, rating, and visual bars
  - Responsive chart resize handling

Closes #70
2026-03-22 00:31:02 +00:00

210 lines
5.6 KiB
CSS

/* ══════════════════════════════════════════════════════════════
Commander AI Lab — Distillation Loop Dashboard
Page-specific styles only; shared tokens & components in global.css
══════════════════════════════════════════════════════════════ */
/* ── Legacy variable aliases (used by distillation.js inline styles) ── */
:root {
--lab-bg: var(--bg);
--lab-surface: var(--surface-1);
--lab-border: var(--border);
--lab-accent: var(--accent);
--lab-accent-dim: var(--accent-dim);
--lab-text: var(--text);
--lab-text-dim: var(--text-secondary);
--lab-success: var(--success);
--lab-success-dim:var(--success-dim);
--lab-warning: var(--warning);
--lab-warning-dim:var(--warning-dim);
--lab-error: var(--error);
--lab-error-dim: var(--error-dim);
--lab-purple: var(--purple);
--lab-purple-dim: var(--purple-dim);
}
/* ── Status Cards (extend global .stat-card) ───────────────── */
.status-grid {
grid-template-columns: repeat(4, 1fr);
}
.status-card {
text-align: center;
}
.status-card.active {
border-color: var(--accent);
}
.status-card.success {
border-color: var(--success);
}
.status-card.error-state {
border-color: var(--error);
}
.card-icon {
font-size: 24px;
margin-bottom: var(--sp-2);
}
/* ── Panel heading with bottom border ──────────────────────── */
.panel > .panel-header {
padding-bottom: var(--sp-3);
border-bottom: 1px solid var(--border);
}
/* ── Training Controls Layout ──────────────────────────────── */
.train-controls {
padding-top: var(--sp-4);
}
.param-grid {
margin-bottom: var(--sp-3);
}
.distill-bottom-params {
grid-template-columns: repeat(2, 1fr);
}
.distill-actions {
display: flex;
gap: var(--sp-2);
margin-top: var(--sp-2);
}
.btn-icon {
font-size: 14px;
}
/* ── Progress Sections ─────────────────────────────────────── */
.train-progress {
margin-top: var(--sp-5);
padding-top: var(--sp-4);
border-top: 1px solid var(--border);
}
.train-progress.hidden {
display: none;
}
/* ── Phase Badges ──────────────────────────────────────────── */
.phase-badge {
display: inline-block;
padding: 3px 10px;
border-radius: var(--radius-md);
font-size: 11px;
font-weight: 600;
text-transform: uppercase;
letter-spacing: 0.05em;
}
.phase-badge.running { background: var(--accent-dim); color: var(--accent); }
.phase-badge.starting { background: var(--accent-dim); color: var(--accent); }
.phase-badge.done { background: var(--success-dim); color: var(--success); }
.phase-badge.error { background: var(--error-dim); color: var(--error); }
.phase-badge.idle { background: var(--surface-3); color: var(--text-secondary); }
#progress-message {
font-size: 13px;
color: var(--text);
}
/* ── Progress Bar overrides ────────────────────────────────── */
.progress-bar-container {
margin-bottom: var(--sp-2);
}
.progress-bar {
transition: width 0.5s var(--ease);
}
.progress-bar.done {
background: var(--success);
}
.progress-bar.error {
background: var(--error);
}
/* ── Chart Containers ──────────────────────────────────────── */
.chart-container {
position: relative;
min-height: 200px;
}
.chart-container canvas {
width: 100%;
}
.chart-container .empty-state {
position: absolute;
inset: 0;
display: flex;
align-items: center;
justify-content: center;
padding: var(--sp-6);
}
/* ── History Table ─────────────────────────────────────────── */
.distill-history-table {
font-size: 13px;
}
.distill-history-table .tag {
display: inline-flex;
align-items: center;
padding: 2px 8px;
font-size: 11px;
font-weight: 600;
border-radius: var(--radius-sm);
white-space: nowrap;
text-transform: capitalize;
}
/* ── Responsive ────────────────────────────────────────────── */
@media (max-width: 768px) {
.status-grid {
grid-template-columns: repeat(2, 1fr);
}
.param-grid {
grid-template-columns: repeat(2, 1fr);
}
.distill-bottom-params {
grid-template-columns: 1fr;
}
}
@media (max-width: 480px) {
.status-grid {
grid-template-columns: 1fr;
}
.param-grid {
grid-template-columns: 1fr;
}
}
/* ── ELO Panel ─────────────────────────────────────────── */
.elo-panel-header {
display: flex;
align-items: center;
justify-content: space-between;
}
.elo-table {
margin-top: var(--sp-3);
}
.elo-table td {
vertical-align: middle;
}
/* Refresh button spin animation */
@keyframes spin {
from { transform: rotate(0deg); }
to { transform: rotate(360deg); }
}