You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am trying to extend the executor sampling config with a custom config and try to plumb it. But running into error while compiling. I am using the cpp backend (model_runner_cpp.py).
/code/tensorrt_llm/examples/llama# python ../run.py --max_output_len=20 --tokenizer_dir /code/tensorrt_llm/Llama-2-7b-hf/ --engine_dir /tmp/ws/models/compiled/engine/fp16/1-gpu --input_text "In Bash, how do I list all text files?" --output_logits_npy logits.npy --temperature 1.2 --repetition_penalty 1.4
System Info
Branch: v0.15
HW
I am trying to extend the executor sampling config with a custom config and try to plumb it. But running into error while compiling. I am using the cpp backend (
model_runner_cpp.py
).I used Llama2 7b model.
compilation commands:
runtime command:
Issue:
Here are my changes:
executor.txt
samplingConfig.txt
Who can help?
No response
Information
Tasks
examples
folder (such as GLUE/SQuAD, ...)Reproduction
I added already the diffs in the above description.
Expected behavior
Should able to compile.
actual behavior
Compilation error.
additional notes
N/A.
The text was updated successfully, but these errors were encountered: