-
Notifications
You must be signed in to change notification settings - Fork 152
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
chore: update lora notebook + fix order inputs in llama #967
base: main
Are you sure you want to change the base?
Conversation
@@ -4,6 +4,6 @@ peft==0.12.0 | |||
Jinja2==3.1.4 | |||
matplotlib==3.7.5 | |||
datasets==3.1.0 | |||
accelerate==1.2.0 | |||
accelerate==1.0.1 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Which package the 1.2.0 version had conflicts with?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
1.2.0 isn't compatible with python 3.8. I will probably revert this and add a comment that we don't support 3.8 in this use case
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
yes please add a comment
@@ -134,7 +133,7 @@ | |||
{ | |||
"data": { | |||
"application/vnd.jupyter.widget-view+json": { | |||
"model_id": "9775e413ec264b2eb14ee53dbc381474", |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Since the PR is red, I suggest some updates:
# Apply LoRA to the model.
# The 'target_modules' parameter can be set to "all-linear" to apply LoRA to all linear modules.
# By default, only the 'c_attn' projection layers are fine-tuned with LoRA.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
You can remove the extra spaces between the comment and the function:
# Since we've already handled padding and labels, we can use a custom data collator
def data_collator(features):
56d002c
to
c0094ae
Compare
No description provided.