A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 19:45 UTC
[Bugfix] Fix tqdm progress bar when SamplingParams.n > 1 (#12428)
Signed-off-by: Yuchen Yan <740987012@qq.com>
Yuchen Yan committed
Mar 9, 2025 at 12:14 UTC
73ae0b44e96caa66e8b5d46badece2a3186ce304
1 file changed
+3
-1
vllm/entrypoints/llm.py
+3
-1
@@ -1394,7 +1394,9 @@ class LLM:
1394
pbar.postfix = (
1395
f"est. speed input: {in_spd:.2f} toks/s, "
1396
f"output: {out_spd:.2f} toks/s")
1397
- pbar.update(1)
1397
+ pbar.update(len(output.outputs))
1398
+ else:
1399
+ pbar.update(1)
1400
1401
if use_tqdm:
1402
pbar.close()