Optimize retract (#440)

examples/usage/json_logprobs.py

+# NOTE: Currently this can only be run through HTTP requests.
+import json
+from concurrent.futures import ThreadPoolExecutor
+
+from json_decode import character_regex
+
+from sglang.utils import http_request
+
+character_names = ["Hermione Granger", "Ron Weasley", "Harry Potter"]
+
+base_url = "http://localhost:30000"
+
+prompt = "is a character in Harry Potter. Please fill in the following information about this character.\n"
+
+
+def openai_api_request(name):
+    data = {
+        "model": "",
+        "prompt": name + prompt,
+        "temperature": 0,
+        "max_tokens": 128,
+        "regex": character_regex,
+        "logprobs": 3,
+    }
+    res = http_request(base_url + "/v1/completions", json=data).json()
+
+    # with open(f"json_logprobs_{name.replace(' ', '_')}_tmp.json", "w") as fout:
+    #     fout.write(json.dumps(res, indent=4))
+
+    logprobs = res["choices"][0]["logprobs"]
+    usage = res["usage"]
+    assert len(logprobs["token_logprobs"]) == len(logprobs["tokens"])
+    assert len(logprobs["token_logprobs"]) == len(logprobs["top_logprobs"])
+    assert len(logprobs["token_logprobs"]) == usage["completion_tokens"] - 1
+
+    return res
+
+
+def srt_api_request(name):
+    data = {
+        "text": name + prompt,
+        "sampling_params": {
+            "temperature": 0,
+            "max_new_tokens": 128,
+            "regex": character_regex,
+        },
+        "return_logprob": True,
+        "logprob_start_len": 0,
+        "top_logprobs_num": 3,
+        "return_text_in_logprobs": True,
+    }
+
+    res = http_request(base_url + "/generate", json=data).json()
+
+    # with open(f"json_logprobs_{name.replace(' ', '_')}_tmp.json", "w") as fout:
+    #     fout.write(json.dumps(res, indent=4))
+
+    meta_info = res["meta_info"]
+    assert len(meta_info["prefill_token_logprobs"]) == len(
+        meta_info["prefill_top_logprobs"]
+    )
+    assert len(meta_info["decode_token_logprobs"]) == len(
+        meta_info["decode_top_logprobs"]
+    )
+    assert len(meta_info["prefill_token_logprobs"]) == meta_info["prompt_tokens"]
+    assert len(meta_info["decode_token_logprobs"]) == meta_info["completion_tokens"] - 1
+
+    return res
+
+
+def pretty_print(res):
+    meta_info = res["meta_info"]
+
+    print("\n\n", "=" * 30, "Prefill", "=" * 30)
+    for i in range(len(meta_info["prefill_token_logprobs"])):
+        print(f"{str(meta_info['prefill_token_logprobs'][i][2].encode()): <20}", end="")
+        top_ks = (
+            [str(t[2].encode()) for t in meta_info["prefill_top_logprobs"][i]]
+            if meta_info["prefill_top_logprobs"][i]
+            else []
+        )
+        for top_k in top_ks:
+            print(f"{top_k: <15}", end="")
+        print()
+
+    print("\n\n", "=" * 30, "Decode", "=" * 30)
+    for i in range(len(meta_info["decode_token_logprobs"])):
+        print(f"{str(meta_info['decode_token_logprobs'][i][2].encode()): <20}", end="")
+        top_ks = [str(t[2].encode()) for t in meta_info["decode_top_logprobs"][i]]
+        for top_k in top_ks:
+            print(f"{top_k: <15}", end="")
+        print()
+
+    print(res["text"])
+
+
+if __name__ == "__main__":
+    with ThreadPoolExecutor() as executor:
+        ress = executor.map(srt_api_request, character_names)
+
+    for res in ress:
+        pretty_print(res)
+
+    openai_api_request("Hermione Granger")

python/sglang/global_config.py

@@ -28,5 +28,11 @@ class GlobalConfig:
         # Request dependency time due to network delay
         self.request_dependency_time = 0.03
 
+        # New generation token ratio estimation
+        self.base_new_token_ratio = 0.4
+        self.base_min_new_token_ratio = 0.2
+        self.new_token_ratio_decay = 0.0001
+        self.new_token_ratio_recovery = 0.05
+
 
 global_config = GlobalConfig()

python/sglang/srt/layers/logits_processor.py

@@ -50,21 +50,22 @@ class LogitsProcessor(nn.Module):
             prefill_top_logprobs, decode_top_logprobs = [], []
             pt = 0
             # NOTE: the GPU-CPU overhead can be reduced
-            extend_seq_lens_cpu = input_metadata.extend_seq_lens.cpu().numpy()
-            for i in range(len(extend_seq_lens_cpu)):
-                if extend_seq_lens_cpu[i] == 0:
+            extend_seq_lens_cpu = input_metadata.extend_seq_lens.tolist()
+            for i, extend_seq_len in enumerate(extend_seq_lens_cpu):
+                if extend_seq_len == 0:
                     prefill_top_logprobs.append([])
                     decode_top_logprobs.append([])
                     continue
                 k = input_metadata.top_logprobs_nums[i]
-                t = all_logprobs[pt : pt + extend_seq_lens_cpu[i]].topk(k)
+                t = all_logprobs[pt : pt + extend_seq_len].topk(k)
                 vs_cpu = t.values.tolist()
                 ps_cpu = t.indices.tolist()
                 prefill_top_logprobs.append(
                     [list(zip(vs_cpu[j], ps_cpu[j])) for j in range(len(vs_cpu) - 1)]
                 )
                 decode_top_logprobs.append(list(zip(vs_cpu[-1], ps_cpu[-1])))
-                pt += extend_seq_lens_cpu[i]
+                pt += extend_seq_len
+
             return prefill_top_logprobs, decode_top_logprobs
 
     def forward(self, input_ids, hidden_states, weight, input_metadata: InputMetadata):
@@ -145,7 +146,7 @@ class LogitsProcessor(nn.Module):
                 )
 
 
-if __name__ == "__main__":
+def test():
     all_logprobs = torch.tensor(
         #       s                     s                s
         [[0, 1, 2, 3], [1, 2, 3, 4], [2, 3, 4, 5], [3, 4, 5, 6], [4, 5, 6, 7]],
@@ -173,3 +174,7 @@ if __name__ == "__main__":
     print("start", start)
     print("end", end)
     print("sum_logp", sum_logp)
+
+
+if __name__ == "__main__":
+    test()

python/sglang/srt/managers/detokenizer_manager.py

@@ -51,11 +51,6 @@ class DetokenizerManager:
                 # Trim stop str
                 # TODO(lmzheng): handle the case where multiple stop strs are hit
                 for i in range(len(output_strs)):
-                    if recv_obj.hit_stop_str[i] is not None:
-                        pos = output_strs[i].find(recv_obj.hit_stop_str[i])
-                        if pos != -1:
-                            output_strs[i] = output_strs[i][:pos]
-
                     if len(output_tokens[i]) > 0:
                         first_token = self.tokenizer.convert_ids_to_tokens(
                             int(output_tokens[i][0])
@@ -65,9 +60,12 @@ class DetokenizerManager:
                         if first_token.startswith("▁"):
                             output_strs[i] = " " + output_strs[i]
 
-                    output_strs[i] = (
-                        recv_obj.output_and_jump_forward_strs[i] + output_strs[i]
-                    )
+                    output_strs[i] = recv_obj.prev_output_strs[i] + output_strs[i]
+
+                    if recv_obj.hit_stop_str[i] is not None:
+                        pos = output_strs[i].find(recv_obj.hit_stop_str[i])
+                        if pos != -1:
+                            output_strs[i] = output_strs[i][:pos]
 
                 self.send_to_tokenizer.send_pyobj(
                     BatchStrOut(

python/sglang/srt/managers/io_struct.py

@@ -106,8 +106,8 @@ class TokenizedGenerateReqInput:
 @dataclass
 class BatchTokenIDOut:
     rids: List[str]
+    prev_output_strs : List[str]
     output_tokens: List[List[int]]
-    output_and_jump_forward_strs: List[str]
     hit_stop_str: List[Optional[str]]
     skip_special_tokens: List[bool]
     spaces_between_special_tokens: List[bool]

python/sglang/srt/managers/router/infer_batch.py

@@ -36,15 +36,15 @@ class FinishReason(IntEnum):
 
 
 class Req:
-    def __init__(self, rid, input_text, input_ids):
+    def __init__(self, rid, origin_input_text, origin_input_ids):
         self.rid = rid
-        self.input_text = input_text
-        self.input_ids = input_ids
+        self.origin_input_text = origin_input_text
+        self.origin_input_ids = origin_input_ids
+        self.origin_input_ids_unpadded = origin_input_ids  # before image padding
+        self.prev_output_str = ""
+        self.prev_output_ids = []
         self.output_ids = []
-
-        # Since jump forward may retokenize the prompt with partial outputs,
-        # we maintain the original prompt length to report the correct usage.
-        self.prompt_tokens = len(input_ids)
+        self.input_ids = None  # input_ids = origin_input_ids + prev_output_ids
 
         # The number of decoded tokens for token usage report. Note that
         # this does not include the jump forward tokens.
@@ -76,15 +76,24 @@ class Req:
         self.top_logprobs_num = 0
         self.normalized_prompt_logprob = None
         self.prefill_token_logprobs = None
-        self.decode_token_logprobs = None
+        self.decode_token_logprobs = []
         self.prefill_top_logprobs = None
-        self.decode_top_logprobs = None
+        self.decode_top_logprobs = []
+        # The tokens is prefilled but need to be considered as decode tokens
+        # and should be updated for the decode logprobs
+        self.last_update_decode_tokens = 0
 
         # Constrained decoding
         self.regex_fsm = None
         self.regex_fsm_state = 0
         self.jump_forward_map = None
-        self.output_and_jump_forward_str = ""
+
+    def partial_decode(self, ids):
+        first_token = self.tokenizer.convert_ids_to_tokens(ids[0])
+        first_token = (
+            first_token.decode() if isinstance(first_token, bytes) else first_token
+        )
+        return (" " if first_token.startswith("▁") else "") + self.tokenizer.decode(ids)
 
     def max_new_tokens(self):
         return self.sampling_params.max_new_tokens
@@ -93,7 +102,10 @@ class Req:
         if self.finished:
             return
 
-        if len(self.output_ids) >= self.sampling_params.max_new_tokens:
+        if (
+            len(self.prev_output_ids) + len(self.output_ids)
+            >= self.sampling_params.max_new_tokens
+        ):
             self.finished = True
             self.finish_reason = FinishReason.LENGTH
             return
@@ -112,60 +124,66 @@ class Req:
             )
 
             for stop_str in self.sampling_params.stop_strs:
-                if stop_str in tail_str:
+                # FIXME: (minor) try incremental match in prev_output_str
+                if stop_str in tail_str or stop_str in self.prev_output_str:
                     self.finished = True
                     self.finish_reason = FinishReason.STOP_STR
                     self.hit_stop_str = stop_str
                     return
 
     def jump_forward_and_retokenize(self, jump_forward_str, next_state):
-        old_output_str = self.tokenizer.decode(self.output_ids)
         # FIXME: This logic does not really solve the problem of determining whether
         # there should be a leading space.
-        first_token = self.tokenizer.convert_ids_to_tokens(self.output_ids[0])
-        first_token = (
-            first_token.decode() if isinstance(first_token, bytes) else first_token
+        cur_output_str = self.partial_decode(self.output_ids)
+
+        # TODO(lsyin): apply re-tokenize only for decode tokens so that we do not need origin_input_text anymore
+        if self.origin_input_text is None:
+            # Recovering text can only use unpadded ids
+            self.origin_input_text = self.tokenizer.decode(
+                self.origin_input_ids_unpadded
             )
-        if first_token.startswith("▁"):
-            old_output_str = " " + old_output_str
-        if self.input_text is None:
-            # TODO(lmzheng): This can be wrong. Check with Liangsheng.
-            self.input_text = self.tokenizer.decode(self.input_ids)
-        new_input_string = (
-            self.input_text
-            + self.output_and_jump_forward_str
-            + old_output_str
+
+        all_text = (
+            self.origin_input_text
+            + self.prev_output_str
+            + cur_output_str
             + jump_forward_str
         )
-        new_input_ids = self.tokenizer.encode(new_input_string)
-        if self.pixel_values is not None:
-            # NOTE: This is a hack because the old input_ids contains the image padding
-            jump_forward_tokens_len = len(self.tokenizer.encode(jump_forward_str))
+        all_ids = self.tokenizer.encode(all_text)
+        prompt_tokens = len(self.origin_input_ids_unpadded)
+        self.origin_input_ids = all_ids[:prompt_tokens]
+        self.origin_input_ids_unpadded = self.origin_input_ids
+        # NOTE: the output ids may not strictly correspond to the output text
+        old_prev_output_ids = self.prev_output_ids
+        self.prev_output_ids = all_ids[prompt_tokens:]
+        self.prev_output_str = self.prev_output_str + cur_output_str + jump_forward_str
+        self.output_ids = []
+
+        self.regex_fsm_state = next_state
+
+        if self.return_logprob:
+            # For fast-forward part's logprobs
+            k = 0
+            for i, old_id in enumerate(old_prev_output_ids):
+                if old_id == self.prev_output_ids[i]:
+                    k = k + 1
                 else:
-            jump_forward_tokens_len = (
-                len(new_input_ids) - len(self.input_ids) - len(self.output_ids)
-            )
+                    break
+            self.decode_token_logprobs = self.decode_token_logprobs[:k]
+            self.decode_top_logprobs = self.decode_top_logprobs[:k]
+            self.logprob_start_len = prompt_tokens + k
+            self.last_update_decode_tokens = len(self.prev_output_ids) - k
 
         # print("=" * 100)
         # print(f"Catch jump forward:\n{jump_forward_str}")
         # print(self.tokenizer.convert_ids_to_tokens(self.input_ids))
         # print(self.tokenizer.convert_ids_to_tokens(new_input_ids))
 
-        self.input_ids = new_input_ids
-        self.output_ids = []
-        self.sampling_params.max_new_tokens = max(
-            self.sampling_params.max_new_tokens - jump_forward_tokens_len, 0
-        )
-        self.regex_fsm_state = next_state
-        self.output_and_jump_forward_str = (
-            self.output_and_jump_forward_str + old_output_str + jump_forward_str
-        )
-
         # print(f"Output and jump forward str:\n{self.output_and_jump_forward_str}")
         # print("*" * 100)
 
     def __repr__(self):
-        return f"rid(n={self.rid}, " f"input_ids={self.input_ids}, "
+        return f"rid(n={self.rid}, " f"input_ids={self.origin_input_ids}, "
 
 
 @dataclass
@@ -336,6 +354,7 @@ class Batch:
 
     def retract_decode(self):
         sorted_indices = [i for i in range(len(self.reqs))]
+        # TODO(lsyin): improve the priority of retraction
         sorted_indices.sort(
             key=lambda i: (len(self.reqs[i].output_ids), -len(self.reqs[i].input_ids)),
             reverse=True,
@@ -356,18 +375,27 @@ class Batch:
             ][last_uncached_pos : seq_lens_cpu[idx]]
             self.token_to_kv_pool.dec_refs(token_indices)
 
+            # release the last node
             self.tree_cache.dec_lock_ref(req.last_node)
+
+            cur_output_str = req.partial_decode(req.output_ids)
+            req.prev_output_str = req.prev_output_str + cur_output_str
+            req.prev_output_ids.extend(req.output_ids)
+
             req.prefix_indices = None
             req.last_node = None
             req.extend_input_len = 0
             req.output_ids = []
-            req.regex_fsm_state = 0
+
+            # For incremental logprobs
+            req.last_update_decode_tokens = 0
+            req.logprob_start_len = 10**9
 
         self.filter_batch(sorted_indices)
 
         return retracted_reqs
 
-    def check_for_jump_forward(self):
+    def check_for_jump_forward(self, model_runner):
         jump_forward_reqs = []
         filter_indices = [i for i in range(len(self.reqs))]
 
@@ -397,6 +425,18 @@ class Batch:
                     # jump-forward
                     req.jump_forward_and_retokenize(jump_forward_str, next_state)
 
+                    # re-applying image padding
+                    if req.pixel_values is not None:
+                        (
+                            req.origin_input_ids,
+                            req.image_offset,
+                        ) = model_runner.model.pad_input_ids(
+                            req.origin_input_ids_unpadded,
+                            req.pad_value,
+                            req.pixel_values.shape,
+                            req.image_size,
+                        )
+
                     jump_forward_reqs.append(req)
                     filter_indices.remove(i)
 

python/sglang/srt/managers/router/model_rpc.py

@@ -4,7 +4,7 @@ import multiprocessing
 import time
 import warnings
 from concurrent.futures import ThreadPoolExecutor
-from typing import Any, Dict, List, Optional, Tuple, Union
+from typing import List, Optional
 
 import rpyc
 import torch
@@ -16,6 +16,7 @@ try:
 except ImportError:
     from vllm.logger import logger as vllm_default_logger
 
+from sglang.global_config import global_config
 from sglang.srt.constrained.fsm_cache import FSMCache
 from sglang.srt.constrained.jump_forward import JumpForwardCache
 from sglang.srt.hf_transformers_utils import get_processor, get_tokenizer
@@ -106,7 +107,8 @@ class ModelRpcServer:
         set_random_seed(server_args.random_seed)
 
         # Print info
-        logger.info(f"[rank={self.tp_rank}] "
+        logger.info(
+            f"[rank={self.tp_rank}] "
             f"max_total_num_token={self.max_total_num_token}, "
             f"max_prefill_num_token={self.max_prefill_num_token}, "
             f"context_len={self.model_config.context_len}, "
@@ -151,9 +153,20 @@ class ModelRpcServer:
         self.jump_forward_cache = JumpForwardCache()
 
         # Init new token estimation
-        self.new_token_ratio = min(0.4 * server_args.schedule_conservativeness, 1.0)
-        self.min_new_token_ratio = min(0.2 * server_args.schedule_conservativeness, 1.0)
-        self.new_token_ratio_step = (0.0001, 0.05)  # (down, up)
+        assert (
+            server_args.schedule_conservativeness >= 0
+        ), "Invalid schedule_conservativeness"
+        self.new_token_ratio = min(
+            global_config.base_new_token_ratio * server_args.schedule_conservativeness,
+            1.0,
+        )
+        self.min_new_token_ratio = min(
+            global_config.base_min_new_token_ratio
+            * server_args.schedule_conservativeness,
+            1.0,
+        )
+        self.new_token_ratio_decay = global_config.new_token_ratio_decay
+        self.new_token_ratio_recovery = global_config.new_token_ratio_recovery
 
     def exposed_step(self, recv_reqs):
         if self.tp_size != 1:
@@ -256,8 +269,13 @@ class ModelRpcServer:
                 (recv_req.image_hash >> 64) % self.model_config.vocab_size,
             ]
             req.image_size = recv_req.image_size
-            req.input_ids, req.image_offset = self.model_runner.model.pad_input_ids(
-                req.input_ids, req.pad_value, req.pixel_values.shape, req.image_size
+            req.origin_input_ids, req.image_offset = (
+                self.model_runner.model.pad_input_ids(
+                    req.origin_input_ids_unpadded,
+                    req.pad_value,
+                    req.pixel_values.shape,
+                    req.image_size,
+                )
             )
         req.sampling_params = recv_req.sampling_params
         req.return_logprob = recv_req.return_logprob
@@ -275,11 +293,11 @@ class ModelRpcServer:
                 )
 
         # Truncate prompts that are too long
-        req.input_ids = req.input_ids[: self.model_config.context_len - 1]
+        req.origin_input_ids = req.origin_input_ids[: self.model_config.context_len - 1]
         req.sampling_params.max_new_tokens = min(
             req.sampling_params.max_new_tokens,
-            self.model_config.context_len - 1 - len(req.input_ids),
-            self.max_total_num_token - 128 - len(req.input_ids),
+            self.model_config.context_len - 1 - len(req.origin_input_ids),
+            self.max_total_num_token - 128 - len(req.origin_input_ids),
         )
         self.forward_queue.append(req)
 
@@ -292,6 +310,10 @@ class ModelRpcServer:
 
         # Compute matched prefix length
         for req in self.forward_queue:
+            assert (
+                len(req.output_ids) == 0
+            ), "The output ids should be empty when prefilling"
+            req.input_ids = req.origin_input_ids + req.prev_output_ids
             prefix_indices, last_node = self.tree_cache.match_prefix(req.input_ids)
             if req.return_logprob:
                 prefix_indices = prefix_indices[: req.logprob_start_len]
@@ -319,7 +341,7 @@ class ModelRpcServer:
             )
 
         for req in self.forward_queue:
-            if req.return_logprob:
+            if req.return_logprob and req.normalized_prompt_logprob is None:
                 # Need at least two tokens to compute normalized logprob
                 if req.extend_input_len < 2:
                     delta = 2 - req.extend_input_len
@@ -441,8 +463,10 @@ class ModelRpcServer:
             req.check_finished()
 
             if req.return_logprob:
+                if req.normalized_prompt_logprob is None:
                     req.normalized_prompt_logprob = normalized_prompt_logprobs[i]
 
+                if req.prefill_token_logprobs is None:
                     # If logprob_start_len > 0, then first logprob_start_len prompt tokens will be ignored.
                     req.prefill_token_logprobs = list(
                         zip(
@@ -454,15 +478,38 @@ class ModelRpcServer:
                         req.prefill_token_logprobs = [
                             (None, req.input_ids[0])
                         ] + req.prefill_token_logprobs
-                req.decode_token_logprobs = [
+
+                if req.last_update_decode_tokens != 0:
+                    req.decode_token_logprobs.extend(
+                        list(
+                            zip(
+                                prefill_token_logprobs[
+                                    pt
+                                    + req.extend_input_len
+                                    - req.last_update_decode_tokens : pt
+                                    + req.extend_input_len
+                                    - 1
+                                ],
+                                req.input_ids[-req.last_update_decode_tokens + 1 :],
+                            )
+                        )
+                    )
+
+                req.decode_token_logprobs.append(
                     (last_token_logprobs[i], next_token_ids[i])
-                ]
+                )
 
             if req.top_logprobs_num > 0:
+                if req.prefill_top_logprobs is None:
                     req.prefill_top_logprobs = prefill_top_logprobs[i]
                     if req.logprob_start_len == 0:
                         req.prefill_top_logprobs = [None] + req.prefill_top_logprobs
-                req.decode_top_logprobs = [decode_top_logprobs[i]]
+
+                if req.last_update_decode_tokens != 0:
+                    req.decode_top_logprobs.extend(
+                        prefill_top_logprobs[i][-req.last_update_decode_tokens + 1 :]
+                    )
+                req.decode_top_logprobs.append(decode_top_logprobs[i])
 
             pt += req.extend_input_len
 
@@ -484,7 +531,7 @@ class ModelRpcServer:
         # check if decode out of memory
         if not batch.check_decode_mem():
             old_ratio = self.new_token_ratio
-            self.new_token_ratio = min(old_ratio + self.new_token_ratio_step[1], 1.0)
+            self.new_token_ratio = min(old_ratio + self.new_token_ratio_recovery, 1.0)
 
             retracted_reqs = batch.retract_decode()
             logger.info(
@@ -495,26 +542,13 @@ class ModelRpcServer:
             self.forward_queue.extend(retracted_reqs)
         else:
             self.new_token_ratio = max(
-                self.new_token_ratio - self.new_token_ratio_step[0],
+                self.new_token_ratio - self.new_token_ratio_decay,
                 self.min_new_token_ratio,
             )
 
         if not self.disable_regex_jump_forward:
             # check for jump-forward
-            jump_forward_reqs = batch.check_for_jump_forward()
-
-            # check for image jump-forward
-            for req in jump_forward_reqs:
-                if req.pixel_values is not None:
-                    (
-                        req.input_ids,
-                        req.image_offset,
-                    ) = self.model_runner.model.pad_input_ids(
-                        req.input_ids,
-                        req.pad_value,
-                        req.pixel_values.shape,
-                        req.image_size,
-                    )
+            jump_forward_reqs = batch.check_for_jump_forward(self.model_runner)
 
             self.forward_queue.extend(jump_forward_reqs)
             if batch.is_empty():
@@ -557,8 +591,8 @@ class ModelRpcServer:
 
     def handle_finished_requests(self, batch: Batch):
         output_rids = []
+        prev_output_strs = []
         output_tokens = []
-        output_and_jump_forward_strs = []
         output_hit_stop_str = []
         output_skip_special_tokens = []
         output_spaces_between_special_tokens = []
@@ -582,8 +616,8 @@ class ModelRpcServer:
                 )
             ):
                 output_rids.append(req.rid)
+                prev_output_strs.append(req.prev_output_str)
                 output_tokens.append(req.output_ids)
-                output_and_jump_forward_strs.append(req.output_and_jump_forward_str)
                 output_hit_stop_str.append(req.hit_stop_str)
                 output_skip_special_tokens.append(
                     req.sampling_params.skip_special_tokens
@@ -593,10 +627,8 @@ class ModelRpcServer:
                 )
 
                 meta_info = {
-                    "prompt_tokens": req.prompt_tokens,
-                    "completion_tokens": len(req.input_ids)
-                    + len(req.output_ids)
-                    - req.prompt_tokens,
+                    "prompt_tokens": len(req.origin_input_ids),
+                    "completion_tokens": len(req.prev_output_ids) + len(req.output_ids),
                     "completion_tokens_wo_jump_forward": req.completion_tokens_wo_jump_forward,
                     "finish_reason": FinishReason.to_str(req.finish_reason),
                     "hit_stop_str": req.hit_stop_str,
@@ -623,8 +655,8 @@ class ModelRpcServer:
             self.out_pyobjs.append(
                 BatchTokenIDOut(
                     output_rids,
+                    prev_output_strs,
                     output_tokens,
-                    output_and_jump_forward_strs,
                     output_hit_stop_str,
                     output_skip_special_tokens,
                     output_spaces_between_special_tokens,