Skip to content

AliRezagholizadeh/FineTuning

Repository files navigation

license gemma
name FunctionGemma-270m-it-Extended-Mobile-Actions
base_model google/functiongemma-270m-it-270M
tags
function-calling
mobile-actions
gemma
Languages
English

Function Gemma

Function Gemma is a Google trained light-wight model with 270M parameters to enables users to make connection between their prompt and executive actions. You can find another variant of the model fine-tuned by Google over (google/mobile-actions)[https://huggingface.co/datasets/google/mobile-actions] here: litert-community/FunctionGemma_270M_Mobile_Actions. You can also use this FunctionGemma-270m-it-Extended-Mobile-Actions model variation fine-tuned over extended Mobile Actions tools set dataset in AliRGHZ/Mobile-Actions.

For deploying the model into your application, kindly follow the instructions available here. For further fine-tuning you can follow the instructions suggested by Google...

###Training Configuration:

  • Epochs: 2
  • Batch size: 4 per device
  • Gradient accumulation steps: 8
  • Learning rate: 1e-5
  • Scheduler: Cosine
  • Optimizer: AdamW (fused)
  • Precision: bfloat16
  • Gradient checkpointing: Enabled
  • Completion only loss: True (trains only on model outputs, not prompts)

###Training Infrastructure: Hardware: Google Colab A100 GPU Training time: ~24 minutes for 2 epochs Library versions: transformers==5.2.0 torch==2.10.0 huggingface_hub==1.5.0, trl==0.29.0, accelerate==1.13.0

About

A project to manage fine-tuning of a specific model (e.g. Function Gemma) on a dataset (e.g. Mobile Actions) as well as evaluate over a baseline.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages