[Feature][MTP]Support MTP for rl-model (#4009)

* qk norm for speculate decode C16

* support mtp in v1_scheduler mode

* support mtp rope_3d

* support mtp features

* add unit test && del some log

---------

Co-authored-by: yuanxiaolan <yuanxiaolan01@baidu.com>
Co-authored-by: xiaoxiaohehe001 <hiteezsf@163.com>

fastdeploy/engine/sched/resource_manager_v1.py

@@ -84,10 +84,14 @@ class ResourceManagerV1(ResourceManager):
         return len(request.block_tables) * self.config.cache_config.block_size
 
     def get_new_block_nums(self, request: Request, num_new_tokens: int):
-        return (
+        block_num = (
             request.num_computed_tokens + num_new_tokens + self.config.cache_config.block_size - 1
         ) // self.config.cache_config.block_size - len(request.block_tables)
 
+        if self.config.speculative_config.method is not None:
+            block_num = min(block_num + 1, self.config.cache_config.max_block_num_per_seq)
+        return block_num
+
     def _prepare_prefill_task(self, request, new_token_num):
         request.prefill_start_index = request.num_computed_tokens
         request.prefill_end_index = request.num_computed_tokens + new_token_num