lite-hermes

Subscribe to my YouTube channel

Subscribe

lite-hermes is a finetuned version of HelpingAI-Lite

License

This model is licensed under the Apache License 2.0, which allows for the usage, modification, and distribution of the model according to the terms specified in the license.

Language

The model supports English language.

Usage

To utilize the lite-hermes model, you can access the provided Colab notebook. The notebook allows you to run the model on both CPU and GPU. Feel free to make any necessary changes to adapt it to your specific requirements.

CPU and GPU code

!pip install transformers
!pip install sentencepiece
!pip install accelerate

import torch                        # allows Tensor computation with strong GPU acceleration
from transformers import pipeline   # fast way to use pre-trained models for inference
import os

# load model
HL_pipeline = pipeline(model="OEvortex/lite-hermes",
                            torch_dtype=torch.bfloat16,
                            trust_remote_code=True,
                            device_map="auto")

# define helper function
def get_completion_HL(input):
  system = f"""
  You are an expert Physicist.
  You are good at explaining Physics concepts in simple words.
  Help as much as you can.
  """
  prompt = f"#### System: {system}\n#### User: \n{input}\n\n#### Response from Lite:"
  print(prompt)
  HL_response = HL_pipeline(prompt,
                                  max_new_tokens=500
                                  )
  return HL_response[0]["generated_text"]

# let's prompt
prompt = "Explain the difference between nuclear fission and fusion."
# prompt = "Why is the Sky blue?"

print(get_completion_HL(prompt))