Unverified Commit 55f6005f authored by fzyzcjy's avatar fzyzcjy Committed by GitHub
Browse files

Fix bench_one_batch_server (#6503)

parent 7222e1da
...@@ -149,7 +149,7 @@ def run_one_case( ...@@ -149,7 +149,7 @@ def run_one_case(
response = requests.post( response = requests.post(
url + "/generate", url + "/generate",
json={ json={
"input_ids": [input_ids for input_ids, _, _ in input_requests], "input_ids": [req.prompt for req in input_requests],
"sampling_params": { "sampling_params": {
"temperature": temperature, "temperature": temperature,
"max_new_tokens": output_len, "max_new_tokens": output_len,
...@@ -268,6 +268,7 @@ def run_benchmark(server_args: ServerArgs, bench_args: BenchArgs): ...@@ -268,6 +268,7 @@ def run_benchmark(server_args: ServerArgs, bench_args: BenchArgs):
input_len_step_percentage=bench_args.input_len_step_percentage, input_len_step_percentage=bench_args.input_len_step_percentage,
run_name=bench_args.run_name, run_name=bench_args.run_name,
result_filename=bench_args.result_filename, result_filename=bench_args.result_filename,
tokenizer=tokenizer,
) )
) )
finally: finally:
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment