Lumimaid-Magnum-v4-12B Model

Introduction

Lumimaid-Magnum-v4-12B is a text generation model developed by merging components from several existing models. It's designed to leverage advanced language processing capabilities for various applications in natural language generation. The model utilizes the Transformers library for its architecture.

Architecture

The architecture of Lumimaid-Magnum-v4-12B is a merge of contributions from the following base models:

• mistralai/Mistral-Nemo-Instruct-2407
• NeverSleep/Lumimaid-v0.2-12B
• Undi95/LocalC-12B-e2.0
• anthracite-org/magnum-v4-12b

The merge is performed using the DELLA merge method in mergekit, combining elements from these models to enhance the performance of the text generation tasks.

Training

The training process included a fine-tuning phase specifically on Claude input, focusing on a 16k context length. This was designed to improve the model’s adaptability and effectiveness in generating coherent text based on given inputs.

Guide: Running Locally

1. Set Up Environment:

   • Ensure you have Python installed.
   • Install the Transformers library using pip:

     pip install transformers
     

2. Clone the Repository:

   • Clone the model repository to your local machine:

     git clone https://huggingface.co/Undi95/Lumimaid-Magnum-v4-12B
     

3. Load the Model:

   • Use the Transformers library to load the model:

     from transformers import AutoModelForCausalLM, AutoTokenizer
     
     tokenizer = AutoTokenizer.from_pretrained("Undi95/Lumimaid-Magnum-v4-12B")
     model = AutoModelForCausalLM.from_pretrained("Undi95/Lumimaid-Magnum-v4-12B")
     

4. Run Inference:

   • Prepare your input and generate text:

     input_text = "<s>[INST] Your prompt here [/INST]"
     inputs = tokenizer(input_text, return_tensors="pt")
     outputs = model.generate(**inputs)
     print(tokenizer.decode(outputs[0], skip_special_tokens=True))
     

For optimized performance, especially with large models like Lumimaid-Magnum-v4-12B, consider using cloud GPUs available through services like AWS, Google Cloud, or Azure.

License

The model's license can be found in the repository. Ensure to review the terms and conditions before use to comply with any restrictions or obligations.