Samantha v6

Introduction

Samantha-v6 is a pre-trained language model created by merging different models using the mergekit tool. It is designed for text generation tasks and leverages advanced merging techniques, such as the TIES method, to combine the strengths of individual models.

Architecture

Samantha-v6 is built by merging two models: Guilherme34/Samantha-raw-mergenotlora and NeverSleep/Lumimaid-v0.2-8B. The model leverages the architecture of the transformers library, ensuring compatibility with various applications in text generation.

Training

The model was created using the TIES merge method, which involves a strategic combination of pre-trained models to enhance performance. The configuration includes parameters such as density and weight for the Guilherme34/Samantha-raw-mergenotlora model, and normalization was applied with a float16 data type to optimize the model's precision and efficiency.

Guide: Running Locally

To run Samantha-v6 locally, follow these steps:

1. Install Dependencies: Ensure you have Python and the transformers library installed.

   pip install transformers
   

2. Download Model: Use the Hugging Face transformers library to download the model.

   from transformers import AutoModelForCausalLM, AutoTokenizer
   
   tokenizer = AutoTokenizer.from_pretrained("Guilherme34/Samantha-v6")
   model = AutoModelForCausalLM.from_pretrained("Guilherme34/Samantha-v6")
   

3. Run Inference: Use the tokenizer and model to generate text.

   input_text = "Your input text here"
   inputs = tokenizer(input_text, return_tensors="pt")
   outputs = model.generate(**inputs)
   print(tokenizer.decode(outputs[0], skip_special_tokens=True))
   

4. Cloud GPUs: For optimal performance, especially with large models, consider using cloud-based GPU services such as AWS EC2, Google Cloud Platform, or Azure.

License

The model and its associated files are subject to the licenses specified by the original model creators. Users should refer to the respective licenses for each base model used in the merge to ensure compliance.