kanshan/TensorRT-LLMs
1
0
Fork 0
mirror of https://github.com/NVIDIA/TensorRT-LLM.git synced 2026-01-14 06:27:45 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

[nvbug/5262268][fix] Fix trtllm-bench for llama 4 (#4104)

Browse Source 
[fix] Fix trtllm-bench for llama 4

Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com>
Co-authored-by: Zhihan Jiang <68881590+nvzhihanj@users.noreply.github.com>
This commit is contained in:
Mike Iovine 2025-05-09 00:27:57 -04:00 committed by GitHub
parent 57b2fe2019
commit d80dc40135
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 3 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
tensorrt_llm/llmapi/llm.py
View File

@ -295,8 +295,9 @@ class LLM:
inputs = prompt_inputs(inputs)
if not inputs.get("prompt") and inputs.get(
"prompt_token_ids") and not isinstance(self.input_processor,
DefaultInputProcessor):
"prompt_token_ids") and inputs.get(
"multi_modal_data") and not isinstance(
self.input_processor, DefaultInputProcessor):
# VLMs need to process/tokenize the prompt in their own way
prompt = self.tokenizer.decode(inputs['prompt_token_ids'])
inputs = TextPrompt(