Welcome to Super Transformer

Suptertransformer that auto loads Huggingface models

Introduction

This is a single line transformer for easy to load models from Huggingface. It is not to replace Huggingface Transformer process. It simplifies it and speed up the loading the process of the HuggingFace models

Usage

SuperTransformers download the model locally. The super class uses AutoTokenizer and AutoModelForCausalLM.from_pretrained.

Installation


              pip install bitsandbytes>=0.39.0
              pip install --upgrade accelerate transformers

How to run


            python SuperTransformer.py

Example of usage:


            # Load SuperTransformer Class,  (1) Loads Huggingface model, (2) System Prompt (3) Text/prompt (4)Max tokens
            SuperTransformers = SuperTransformers("EpistemeAI/ReasoningCore-3B-RE1-V2","You are a highly knowledgeable assistant with expertise in chemistry and physics. ","What is the area of a circle, radius=16, reason step by step", 2026)
            # 8-bit quantization
            SuperTransformers.HuggingFaceTransformer8bit()
            # or 4-bit quantization
            SuperTransformers.HuggingFaceTransformer4bit()

Returns model and tokenizer


            SuperTransformers = SuperTransformers("EpistemeAI/ReasoningCore-3B-RE1-V2")
            model, tokenizer = HuggingfaceTransfomer()  #returns the model and tokenizer

returns pipline as higher helper


            SuperTransformers = SuperTransformers("EpistemeAI/ReasoningCore-3B-RE1-V2")
            pipe = HuggingfacePipeline()  #returns the pipeline only
            output = pipe(self.text, max_new_tokens=self.max_new_tokens)  # Limit output length to save memory
            # Print the generated output
            print(output)

Try it today SuperTransformer-SHF