Skip to content

Adding A3mega Llama3.1-70b recipe with trtllm#221

Draft
Priya-Quad wants to merge 6 commits intoAI-Hypercomputer:mainfrom
Priya-Quad:main
Draft

Adding A3mega Llama3.1-70b recipe with trtllm#221
Priya-Quad wants to merge 6 commits intoAI-Hypercomputer:mainfrom
Priya-Quad:main

Conversation

@Priya-Quad
Copy link
Copy Markdown
Contributor

No description provided.

Comment on lines +1 to +3
# Copyright 2025 Google LLC
#
# Licensed under the Apache License, Version 2.0 (the "License");
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please remove the template if we already have the trt-inference folder

apiVersion: v1
kind: ConfigMap
metadata:
name: "{{ .Release.Name }}-config"
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please update it with the same format that vllm configs have and all the other files

configFile: serving-args.yaml
configPath: /workload/configs
envs:
- name: LAUNCHER_SCRIPT
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Move these to the launcher or some other file, which have the env variables

Comment thread src/launchers/trtllm-launcher.sh Outdated
pp_size=${SERVING_CONFIG_DICT["pp_size"]:=1}
ep_size=${SERVING_CONFIG_DICT["ep_size"]:=1}
backend=${SERVING_CONFIG_DICT["backend"]:="tensorrt"}
backend="tensorrt"
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please remove the hardcoding, the file is already right, as its being used by multiple models. Please debug the issue

Comment thread src/launchers/trtllm-launcher.sh Outdated
# If custom_dataset is not set, generate a textual dataset with tokens sampled in normal distribution
if [ -z "$dataset_file" ]; then
dataset_file="/ssd/token-norm-dist_${model_name##*/}_${isl}_${osl}_tp${tp_size}.json"
dataset_file="/scratch/token-norm-dist_${model_name##*/}_${isl}_${osl}_tp${tp_size}.json"
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

make changes in your added files instead of changing this

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants