Skip to content

Improve LISA layers_attribute mapping logics #746

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.

Sign up for GitHub

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

Merged
research4pan merged 1 commit into from
Apr 6, 2024
Merged

Improve LISA layers_attribute mapping logics #746

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
24 changes: 12 additions & 12 deletions src/lmflow/pipeline/finetuner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -304,19 +304,19 @@ def __init__(self, n_layers, interval_steps, model):
self.n_layers = n_layers
self.interval_steps = interval_steps
self.model = model

# Determine the way to access layers based on the model type
if self.model.__class__.__name__ == 'LlamaForCausalLM':
self.layers_attribute = 'model.model.layers' # Layer access path for LlamaForCausalLM
elif self.model.__class__.__name__ == 'Qwen2ForCausalLM':
self.layers_attribute = 'model.model.layers' # Layer access path for Qwen model
elif self.model.__class__.__name__ == 'MistralForCausalLM':
self.layers_attribute = 'model.model.layers'
elif self.model.__class__.__name__ == 'MixtralForCausalLM':
self.layers_attribute = 'model.model.layers'
elif self.model.__class__.__name__ == 'GemmaForCausalLM':
self.layers_attribute = 'model.model.layers'
elif self.model.__class__.__name__ == 'GPT2LMHeadModel':
self.layers_attribute = 'model.transformer.h'
class_to_layers_map = {
'LlamaForCausalLM': 'model.model.layers',
'Qwen2ForCausalLM': 'model.model.layers',
'MistralForCausalLM': 'model.model.layers',
'MixtralForCausalLM': 'model.model.layers',
'GemmaForCausalLM': 'model.model.layers',
'GPT2LMHeadModel': 'model.transformer.h',
}
model_class_name = self.model.__class__.__name__
if model_class_name in class_to_layers_map:
self.layers_attribute = class_to_layers_map[model_class_name]
else:
self.layers_attribute = training_args.lisa_layers_attribute
self.total_layers = len(eval('self.' + self.layers_attribute)) # Dynamically execute to get the number of layers
Expand Down