Fix bugs of `logprobs_nums` (#1548)

python/sglang/srt/managers/schedule_batch.py

@@ -748,6 +748,8 @@ class ScheduleBatch:
             self.top_logprobs_nums = [
                 self.top_logprobs_nums[i] for i in unfinished_indices
             ]
+        else:
+            self.top_logprobs_nums = None
         self.has_stream = any(req.stream for req in self.reqs)
 
         self.sampling_info.filter_batch(unfinished_indices, new_indices)
@@ -758,13 +760,11 @@ class ScheduleBatch:
         # needs to be called with pre-merged Batch.reqs.
         self.sampling_info.merge_batch(other.sampling_info)
 
-        self.reqs.extend(other.reqs)
         self.req_pool_indices = torch.concat(
             [self.req_pool_indices, other.req_pool_indices]
         )
         self.seq_lens = torch.concat([self.seq_lens, other.seq_lens])
         self.out_cache_loc = None
-        self.return_logprob = any(req.return_logprob for req in self.reqs)
         if self.return_logprob and other.return_logprob:
             self.top_logprobs_nums.extend(other.top_logprobs_nums)
         elif self.return_logprob:
@@ -772,6 +772,8 @@ class ScheduleBatch:
         elif other.return_logprob:
             self.top_logprobs_nums = [0] * len(self.reqs) + other.top_logprobs_nums
         self.has_stream = any(req.stream for req in self.reqs)
+        self.reqs.extend(other.reqs)
+        self.return_logprob = self.return_logprob or other.return_logprob
 
     def get_model_worker_batch(self):
         if self.forward_mode.is_decode():