Command Line Interface
Contents
Command Line Interface¶
llm-atc train¶
Launch a train job on a cloud provider
llm-atc train [OPTIONS]
Options
- --model_type <model_type>¶
Required LLM type to train. Run llm-atc show-models to see a list of supported models
- --finetune_data <finetune_data>¶
Required local/cloud URI to finetuning data. (e.g ~/mychat.json, s3://my_bucket/my_chat.json)
- --checkpoint_bucket <checkpoint_bucket>¶
Required object store bucket name
- --checkpoint_store <checkpoint_store>¶
Required object store type [‘S3’, ‘GCS’, ‘AZURE’, ‘R2’, ‘IBM’]
- -n, --name <name>¶
Required Name of this model run.
- --description <description>¶
description of this model run
- -c, --cluster <cluster>¶
Name of skypilot cluster. If name matches existing cluster, will use this cluster
- --cloud <cloud>¶
Which cloud provider to use.
- --envs <envs>¶
Environment variable to set on the remote node. It can be specified multiple times. Examples:
1.
--envs MY_ENV=1
: set$MY_ENV
on the cluster to be 1.2.
--envs MY_ENV2=$HOME
: set$MY_ENV2
on the cluster to be the same value of$HOME
in the local environment where the CLI command is run.3.
--envs MY_ENV3
: set$MY_ENV3
on the cluster to be the same value of$MY_ENV3
in the local environment.
- --region <region>¶
which region to train in. Defaults to any region
- --zone <zone>¶
which zone to train in. Defaults to any zone
- --accelerator <accelerator>¶
Required Which GPU type to use
- --detach_setup <detach_setup>¶
launch task non-interactively. Don’t stream setup logs
- --detach_run <detach_run>¶
Perform execution non-interactively. Calling terminal doesn’t hang on run
- --no_setup¶
Skip setup. Faster if cluster is already provisioned and UP
- Default
False
llm-atc serve¶
Create a cluster to serve an openAI.api_server using FastChat and vLLM
llm-atc serve [OPTIONS]
Options
- -n, --name <name>¶
Required name of model to serve
- --source <source>¶
object store path for llm-atc finetuned model checkpoints.e.g. s3://<bucket-name>/<path>/<to>/<checkpoints>
- -e, --envs <envs>¶
Environment variable to set on the remote node. It can be specified multiple times. Examples:
1.
--envs MY_ENV=1
: set$MY_ENV
on the cluster to be 1.2.
--envs MY_ENV2=$HOME
: set$MY_ENV2
on the cluster to be the same value of$HOME
in the local environment where the CLI command is run.3.
--envs MY_ENV3
: set$MY_ENV3
on the cluster to be the same value of$MY_ENV3
in the local environment.
- --accelerator <accelerator>¶
Which gpu instance to use for serving
- -c, --cluster <cluster>¶
Name of skypilot cluster. If name matches existing cluster, will use this cluster
- --cloud <cloud>¶
which cloud provider to use for deployment
- --region <region>¶
which region to deploy. Defaults to any region
- --zone <zone>¶
which zone to deploy. Defaults to any zone
- --no_setup¶
skip setup step
- Default
False
- --detach_setup¶
Don’t connect to this session
- Default
False
- -d, --detach_run¶
Don’t connect to this session
- Default
False
llm-atc list¶
List models created by llm-atc. For models that are done, their status is permanently marked as available. Jobs that are pending require a cluster to be UP in order to update their status TODO: add checks for status of runs
llm-atc list [OPTIONS]
Options
- --limit <limit>¶
Limit of number of models to print
- --model_type <model_type>¶
Filter models by model type
- --name <name>¶
Filter models by name. Matches against model names with pattern name included