Closed shivchander closed 7 months ago
sample yaml structure:
created_by: ...
domain: ...
seed_examples:
- answer: |
chat, download, generate, init, diff, serve, test and train
question: |
What sub-commands does `lab` has?
- answer: |
`lab train` works locally for Linux (with or without GPUs) and MacBook with M chips. There is a notebook can be used to train and test the model on the cloud.
question: |
What options are there for training using InstructLab CLI?
- answer: |
run `lab download`
question: |
How to download the model using the InstructLab CLI
- answer: |
run `lab init`
question: |
how to initialize the workspace using the InstructLab CLI
task_description: 'How to the `lab` command for the InstructLab CLI'
document:
repo: https://github.com/instruct-lab/cli
commit: 951999a
pattern: README.md
document: repo: https://github.com/instruct-lab/cli commit: 951999a pattern: README.md
Does pattern
allow for globbing? docs/**/*.md
? Or multiple values ['file1.md', 'file2.md']
?
yes, pattern will allow globbing. the intension is to have consistent way to allow point to either single file or multiple files under a folder
As per the discussions here: https://github.com/instruct-lab/cli/issues/702,
We have a new model on how to knowledge contributions should be formatted. Could we update the current knowledge branch of the taxonomy to reflect this.