Let bench_one_batch_server use sharegpt data to make expert distribution more natural (sgl-project#5573)

Author: fzyzcjy

python/sglang/bench_one_batch_server.py

@@ -22,6 +22,7 @@
 import numpy as np
 import requests
 
+from sglang.bench_serving import get_tokenizer, sample_random_requests
 from sglang.srt.entrypoints.http_server import launch_server
 from sglang.srt.server_args import ServerArgs
 from sglang.srt.utils import kill_process_tree
@@ -117,16 +118,19 @@ def run_one_case(
     input_len_step_percentage: float,
     run_name: str,
     result_filename: str,
+    tokenizer,
 ):
     requests.post(url + "/flush_cache")
-    input_lens = [
-        int(input_len * (1 + (i - (batch_size - 1) / 2) * input_len_step_percentage))
-        for i in range(batch_size)
-    ]
-    input_ids = [
-        [int(x) for x in np.random.randint(0, high=16384, size=(input_lens[i],))]
-        for i in range(batch_size)
-    ]
+    input_requests = sample_random_requests(
+        input_len=input_len,
+        output_len=output_len,
+        num_prompts=batch_size,
+        range_ratio=1.0,
+        tokenizer=tokenizer,
+        dataset_path="",
+        random_sample=True,
+        return_text=False,
+    )
 
     use_structured_outputs = False
     if use_structured_outputs:
@@ -145,8 +149,7 @@ def run_one_case(
     response = requests.post(
         url + "/generate",
         json={
-            # "text": texts,
-            "input_ids": input_ids,
+            "input_ids": [input_ids for input_ids, _, _ in input_requests],
             "sampling_params": {
                 "temperature": temperature,
                 "max_new_tokens": output_len,
@@ -228,6 +231,9 @@ def run_benchmark(server_args: ServerArgs, bench_args: BenchArgs):
     else:
         proc, base_url = launch_server_process(server_args)
 
+    tokenizer_id = server_args.tokenizer_path or server_args.model_path
+    tokenizer = get_tokenizer(tokenizer_id)
+
     # warmup
     if not bench_args.skip_warmup:
         print("=" * 8 + " Warmup Begin " + "=" * 8)
@@ -241,6 +247,7 @@ def run_benchmark(server_args: ServerArgs, bench_args: BenchArgs):
             input_len_step_percentage=bench_args.input_len_step_percentage,
             run_name="",
             result_filename="",
+            tokenizer=tokenizer,
         )
         print("=" * 8 + " Warmup End   " + "=" * 8 + "\n")
 

python/sglang/bench_serving.py

@@ -471,6 +471,10 @@ def get_model(pretrained_model_name_or_path: str) -> str:
 def get_tokenizer(
     pretrained_model_name_or_path: str,
 ) -> Union[PreTrainedTokenizer, PreTrainedTokenizerFast]:
+    assert (
+        pretrained_model_name_or_path is not None
+        and pretrained_model_name_or_path != ""
+    )
     if pretrained_model_name_or_path.endswith(
         ".json"
     ) or pretrained_model_name_or_path.endswith(".model"):
@@ -832,6 +836,7 @@ def sample_random_requests(
     tokenizer: PreTrainedTokenizerBase,
     dataset_path: str,
     random_sample: bool = True,
+    return_text: bool = True,
 ) -> List[DatasetRow]:
     input_lens = np.random.randint(
         max(int(input_len * range_ratio), 1),
@@ -892,10 +897,12 @@ def sample_random_requests(
             else:
                 ratio = (input_lens[i] + prompt_len - 1) // prompt_len
                 input_ids = (prompt_token_ids * ratio)[: input_lens[i]]
-            prompt = tokenizer.decode(input_ids)
+            input_content = input_ids
+            if return_text:
+                input_content = tokenizer.decode(input_content)
             input_requests.append(
                 DatasetRow(
-                    prompt=prompt,
+                    prompt=input_content,
                     prompt_len=int(input_lens[i]),
                     output_len=int(output_lens[i]),
                 )
@@ -905,15 +912,15 @@ def sample_random_requests(
         offsets = np.random.randint(0, tokenizer.vocab_size, size=num_prompts)
         input_requests = []
         for i in range(num_prompts):
-            prompt = tokenizer.decode(
-                [
-                    (offsets[i] + i + j) % tokenizer.vocab_size
-                    for j in range(input_lens[i])
-                ]
-            )
+            input_content = [
+                (offsets[i] + i + j) % tokenizer.vocab_size
+                for j in range(input_lens[i])
+            ]
+            if return_text:
+                input_content = tokenizer.decode(input_content)
             input_requests.append(
                 DatasetRow(
-                    prompt=prompt,
+                    prompt=input_content,
                     prompt_len=int(input_lens[i]),
                     output_len=int(output_lens[i]),
                 )