fix: force line-buffered stdout in Argo training pod

Rust tracing with JSON subscriber in a container without TTY uses
fully-buffered stdout — logs only flush on buffer full or process exit.
This made H100 training appear stuck for 28+ minutes with no output.
stdbuf -oL forces line buffering so epoch logs appear immediately.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

infra/k8s/argo/compile-and-train-template.yaml

@@ -497,7 +497,7 @@ spec:
 
             echo "=== Training best: $MODEL ({{workflow.parameters.train-epochs}} epochs) ==="
 
-            ${BINARY} \
+            stdbuf -oL ${BINARY} \
               --model "$MODEL" \
               --symbol {{workflow.parameters.symbol}} \
               --tx-cost-bps {{workflow.parameters.tx-cost-bps}} \