DarkIdol-Llama-3.1-8B-Instruct-1.2-Uncensored

Introduction

DarkIdol-Llama-3.1-8B-Instruct-1.2-Uncensored is a multilingual large language model (LLM) designed for various text generation tasks. It supports roleplay and conversational scenarios in 11 languages, including English, German, and French. The model is uncensored and focuses on providing quick and detailed responses.

Architecture

Llama 3.1 is an auto-regressive language model using an optimized transformer architecture. It utilizes supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to ensure alignment with human preferences for helpfulness and safety. The model supports multilingual text input and output, with a context length of 128k tokens.

Training

Llama 3.1 was pretrained on approximately 15 trillion tokens from publicly available sources. Fine-tuning involved over 25 million synthetically generated examples. Training was conducted on Meta's custom-built GPU cluster, using 39.3 million GPU hours on H100-80GB hardware. The model's training adhered to net-zero greenhouse gas emissions policies.

Guide: Running Locally

1. Install Prerequisites: Ensure transformers version >= 4.43.0 is installed.

   pip install --upgrade transformers
   

2. Load Model: Use the Transformers library to load and run the model.

   import transformers
   
   model_id = "meta-llama/Meta-Llama-3.1-8B-Instruct"
   pipeline = transformers.pipeline(
       "text-generation",
       model=model_id,
       model_kwargs={"torch_dtype": torch.bfloat16},
       device_map="auto",
   )
   

3. Generate Text: Provide input messages and generate text.

   messages = [
       {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
       {"role": "user", "content": "Who are you?"},
   ]
   
   outputs = pipeline(
       messages,
       max_new_tokens=256,
   )
   print(outputs[0]["generated_text"][-1])
   

Cloud GPUs: Consider using cloud GPUs such as AWS EC2 instances with GPU support for efficient processing.

License

This model is distributed under the Llama 3.1 Community License, available at Meta Llama License. The license allows for commercial and research use, provided compliance with its terms.