Update deepseek-moe-16b-base/bert4torch_config.json

deepseek-moe-16b-base/bert4torch_config.json

@@ -1,4 +1,6 @@
 {
+  "model": "deepseek",
+  "template": "pretrained_text_continuation",
   "attention_bias": false,
   "attention_dropout": 0.0,
   "aux_loss_alpha": 0.001,
@@ -10,7 +12,6 @@
   "initializer_range": 0.02,
   "intermediate_size": 10944,
   "max_position_embeddings": 4096,
-  "model": "deepseek",
   "moe_intermediate_size": 1408,
   "moe_layer_freq": 1,
   "n_routed_experts": 64,
@@ -32,5 +33,5 @@
   "skip_init": true,
   "segment_vocab_size": 0,
   "rope_rank": "updown",
-  "generation_config": {"tokenizer_config": {"add_special_tokens": false, "skip_special_tokens": true}, "max_length": 4096, "eos_token_id": 100001}
+  "generation_config": {"tokenizer_config":  {"skip_special_tokens": true, "add_special_tokens": false}, "eos_token_id": [100001]}
 }