sparkring lane: drop kernel-config (mutually exclusive with their --no-enable-flashinfer-autotune); note engine's data->weights mm-encoder fallback

recipes/mimo26-flash-mopd-4x-sparkring/recipe.toml

@@ -76,7 +76,17 @@ args = [
     "--no-enable-flashinfer-autotune",
     "--load-format", "b12x",
     "--kv-cache-dtype", "bfloat16",
-    "--kernel-config", """
+    # NOTE: our nightly's `--kernel-config {enable_flashinfer_autotune:false}`
+    # is mutually exclusive with their `--no-enable-flashinfer-autotune`
+    # (ValueError at engine build) — their qualified shape wins, kernel-config
+    # is dropped for this lane.
+    # Their qualified multimodal shape: full Omni + data-parallel encoder
+    # (this fork auto-falls back to weights for MiMo: "model does not support
+    # --mm-encoder-tp-mode data"), per-prompt media limits, video kwargs,
+    # pinned processor config.
+    "--decode-context-parallel-size", "1",
+    "--enable-prompt-tokens-details",
+    "--limit-mm-per-prompt", """
 '{"enable_cutedsl_warmup":false,"enable_flashinfer_autotune":false}'""",
     # Their qualified multimodal shape: full Omni + data-parallel encoder,
     # per-prompt media limits, video kwargs, pinned processor config.