|
|
--- |
|
|
base_model: HuggingFaceTB/SmolLM2-135M-Instruct |
|
|
library_name: peft |
|
|
language: |
|
|
- en |
|
|
pipeline_tag: text-generation |
|
|
datasets: |
|
|
- flwrlabs/alpaca-gpt4 |
|
|
--- |
|
|
|
|
|
# Model Card for BlossomTune-SmolLM2-135M-Instruct-NLP |
|
|
|
|
|
This PEFT adapter has been trained by using [Flower](https://flower.ai/), a friendly federated AI framework. |
|
|
|
|
|
## Model Details |
|
|
|
|
|
Please check the following GitHub project for model details and evaluation (work in progress!): |
|
|
|
|
|
[https://github.com/ethicalabs-ai/BlossomTuneLLM](https://github.com/ethicalabs-ai/BlossomTuneLLM) |
|
|
|
|
|
## How to Get Started with the Model |
|
|
|
|
|
Use this model as: |
|
|
|
|
|
``` |
|
|
from peft import PeftModel |
|
|
from transformers import AutoModelForCausalLM |
|
|
|
|
|
base_model = AutoModelForCausalLM.from_pretrained("HuggingFaceTB/SmolLM2-135M-Instruct") |
|
|
model = PeftModel.from_pretrained(base_model, "ethicalabs/BlossomTune-SmolLM2-135M-Instruct-NLP") |
|
|
``` |
|
|
|
|
|
### Evaluation Results (Accuracy) |
|
|
|
|
|
- **STEM**: 25.30 % |
|
|
- **Social Sciences**: 27.17 % |
|
|
- **Humanities**: 23.95 % |
|
|
- **Average**: 25.47 % |
|
|
|
|
|
### Communication Budget |
|
|
|
|
|
7899.38 MB |
