Enlarging max_position_embeddings to support long-context generation for the specific inference engine. (#8)

- Enlarging max_position_embeddings to support long-context generation for the specific inference engine. (d806df7d7c1007e08dbd7e161f3c3243e4f46b80)

config.json

@@ -26,7 +26,7 @@
     "hidden_size": 4096,
     "initializer_range": 0.02,
     "intermediate_size": 8192,
-    "max_position_embeddings": 40960,
+    "max_position_embeddings": 65536,
     "max_window_layers": 94,
     "mlp_only_layers": [],
     "model_type": "qwen3_moe",