Update config.json so that it can be run by llm serving engine

config.json

@@ -4,11 +4,11 @@
   ],
   "attention_bias": false,
   "attention_dropout": 0.0,
-  "attn_logit_softcapping": 50.0,
+  "attn_logit_softcapping": null,
   "bos_token_id": 2,
   "cache_implementation": "hybrid",
   "eos_token_id": 1,
-  "final_logit_softcapping": 30.0,
+  "final_logit_softcapping": null,
   "head_dim": 256,
   "hidden_act": "gelu_pytorch_tanh",
   "hidden_activation": "gelu_pytorch_tanh",