Reset buffer size of R3

Author: gongshaotian

fastdeploy/worker/input_batch.py

@@ -189,9 +189,8 @@ def init_share_inputs(self):
         self.cu_seqlens_k = paddle.full([max_num_seqs + 1], 0, dtype="int32")
 
         # Initialize addressing buffers
-        _max_batched_tokens = self.scheduler_config.max_num_batched_tokens
-        self.position_ids_buffer = paddle.zeros([_max_batched_tokens], dtype=paddle.int32)
-        self.slot_mapping_buffer = paddle.zeros([_max_batched_tokens], dtype=paddle.int64)
+        self.position_ids_buffer = paddle.zeros([self.max_chunk_tokens], dtype=paddle.int32)
+        self.slot_mapping_buffer = paddle.zeros([self.max_chunk_tokens], dtype=paddle.int64)
 
         # Declare AttentionBackend buffers
         self.decoder_batch_ids = None