Update Nanbeige Chat Model

config.json

@@ -1,12 +1,12 @@
 {
+  "_name_or_path": "//code/chenzongchao/SFT/DeepSpeed-Chat/SFT_train_v1/nanbeige-16b-cool-cl-sft-0226-bs320/epoch_1step_999/",
   "architectures": [
-    "NanbeigeForCausalLM"
+    "LlamaForCausalLM"
   ],
-  "auto_map": {
-    "AutoConfig": "configuration_nanbeige.NanbeigeConfig",
-    "AutoModelForCausalLM": "modeling_nanbeige.NanbeigeForCausalLM"
-  },
+  "attention_bias": false,
+  "attention_dropout": 0.0,
   "bos_token_id": 1,
+  "dropout": 0.0,
   "eos_token_id": 2,
   "hidden_act": "silu",
   "hidden_size": 5120,
@@ -14,14 +14,18 @@
   "intermediate_size": 13824,
   "max_length": 4096,
   "max_position_embeddings": 4096,
-  "model_type": "nanbeige",
+  "model_type": "llama",
   "num_attention_heads": 40,
   "num_hidden_layers": 48,
+  "num_key_value_heads": 40,
   "pad_token_id": 0,
+  "pretraining_tp": 1,
   "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.28.1",
+  "transformers_version": "4.35.0",
   "use_cache": true,
-  "vocab_size": 59136
+  "vocab_size": 59392
 }