feat: training pipeline for all 10 ML ensemble models

- Add standalone training binaries: TGGN, KAN, xLSTM, Diffusion (DBN data)
- Update Dockerfile.training: 6 → 16 binaries (all 10 models + hyperopt + baseline)
- Expand train.sh: 4 → 10 models, fix registry URL and GPU pool nodeSelector
- Add GPU overlay manifests for trading-service and ml-training-service
- Create training data PVC and upload pod manifests
- Expand web-gateway model validation: 4 → 10 types (training + tune routes)
- Extend dashboard: 10 model cards grouped by category (RL/Temporal/Graph/Generative)
- Add training image build job to Gitea CI workflow
- Update GPU taint controller to exclude inference pool from tainting
- Fix job-template nodeSelector: gpu → gpu-training

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

web-dashboard/src/components/ml/EnsemblePanel.tsx

@@ -4,7 +4,7 @@ interface Props {
   predictions?: MlPrediction[];
 }
 
-const MODELS = ['DQN', 'PPO', 'TFT', 'Mamba2'];
+const MODELS = ['DQN', 'PPO', 'TFT', 'Mamba2', 'xLSTM', 'TGGN', 'TLOB', 'LNN', 'KAN', 'Diffusion'];
 
 export function EnsemblePanel({ predictions = [] }: Props) {
   // Count votes

web-dashboard/src/pages/MLDashboard.tsx

@@ -13,7 +13,12 @@ interface RegimeData {
   duration?: string;
 }
 
-const MODELS = ['DQN', 'PPO', 'TFT', 'Mamba2'] as const;
+const MODEL_GROUPS = [
+  { label: 'Reinforcement Learning', models: ['DQN', 'PPO'] },
+  { label: 'Temporal', models: ['TFT', 'Mamba2', 'xLSTM'] },
+  { label: 'Graph / Structure', models: ['TGGN', 'TLOB', 'LNN'] },
+  { label: 'Generative', models: ['KAN', 'Diffusion'] },
+] as const;
 
 export function MLDashboard() {
   const predictions = useApiQuery<MlPrediction[]>(
@@ -46,25 +51,34 @@ export function MLDashboard() {
         </div>
       )}
 
-      {/* Model Cards */}
+      {/* Model Cards grouped by category */}
       <ComponentErrorBoundary name="Model Cards">
-        <div className="grid grid-cols-1 sm:grid-cols-2 lg:grid-cols-4 gap-4">
-          {MODELS.map((model) => {
-            const pred = preds.find((p) => p.model === model);
-            return (
-              <div
-                key={model}
-                className="rounded border border-[var(--color-border)] bg-[var(--color-bg-card)]"
-              >
-                <ModelCard
-                  model={model}
-                  signal={pred?.signal}
-                  confidence={pred?.confidence}
-                  predictedReturn={pred?.predicted_return}
-                />
+        <div className="space-y-4">
+          {MODEL_GROUPS.map((group) => (
+            <div key={group.label}>
+              <h3 className="text-xs font-semibold uppercase tracking-wider text-[var(--color-text-secondary)] mb-2">
+                {group.label}
+              </h3>
+              <div className="grid grid-cols-1 sm:grid-cols-2 xl:grid-cols-5 gap-4">
+                {group.models.map((model) => {
+                  const pred = preds.find((p) => p.model === model);
+                  return (
+                    <div
+                      key={model}
+                      className="rounded border border-[var(--color-border)] bg-[var(--color-bg-card)]"
+                    >
+                      <ModelCard
+                        model={model}
+                        signal={pred?.signal}
+                        confidence={pred?.confidence}
+                        predictedReturn={pred?.predicted_return}
+                      />
+                    </div>
+                  );
+                })}
               </div>
-            );
-          })}
+            </div>
+          ))}
         </div>
       </ComponentErrorBoundary>