llm.create_chat_completion(
messages = [
{
"role": "user",
"content": "What is the capital of France?"
}
]
)ProfessorF is Nick V. Flor, PhD
Models quantized for research reproducibility purposes
💫 Community Model> Llama 3 8B Instruct by Meta
Model creator: meta-llama
Original model: Meta-Llama-3-8B-Instruct
GGUF quantization: provided by professorf based on llama.cpp PR 6745
ProfessorF (Nick V. Flor, PhD): Quantizes models for research reproducibility. If referenced in a paper, this is the exact quantized model used in that research.
Model Summary:
Llama 3 represents a huge update to the Llama family of models. This model is the 8B parameter instruction tuned model, meaning it's small, fast, and tuned for following instructions.
This model is very happy to follow the given system prompt, so use this to your advantage to get the behavior you desire.
Llama 3 excels at all the general usage situations, including multi turn conversations, general world knowledge, and coding.
This 8B model exceeds the performance of Llama 2's 70B model, showing that the performance is far greater than the previous iteration.
Prompt Template:
Choose the 'Llama 3' preset in your LM Studio.
Under the hood, the model will see a prompt that's formatted like so:
<|begin_of_text|><|start_header_id|>system<|end_header_id|>
{system_prompt}<|eot_id|><|start_header_id|>user<|end_header_id|>
{prompt}<|eot_id|><|start_header_id|>assistant<|end_header_id|>
Use case and examples
Llama 3 should be great for anything you throw at it. Try it with conversations, coding, and just all around general inquiries.
Creative conversations
Using a system prompt of You are a pirate chatbot who always responds in pirate speak!
General knowledge
Coding
Technical Details
Llama 3 was trained on over 15T tokens from a massively diverse range of subjects and languages, and includes 4 times more code than Llama 2.
This model also features Grouped Attention Query (GQA) so that memory usage scales nicely over large contexts.
Instruction fine tuning was performed with a combination of supervised fine-tuning (SFT), rejection sampling, proximal policy optimization (PPO), and direct policy optimization (DPO).
Check out their blog post for more information here
Special thanks
🙏 Special thanks to Georgi Gerganov and the whole team working on llama.cpp for making all of this possible.
🙏 Special thanks to Kalomaze for his dataset (linked here) that was used for calculating the imatrix for these quants, which improves the overall quality!
Disclaimers
ProfessoF does not endorse, support, represent or guarantee the completeness, truthfulness, accuracy, or reliability of any Community Model. You understand that Community Models can produce content that might be offensive, harmful, inaccurate or otherwise inappropriate, or deceptive. Each Community Model is the sole responsibility of the person or entity who originated such Model. ProfessorF may not monitor or control the Community Models and cannot, and does not, take responsibility for any such Model. ProfessorF disclaims all warranties or guarantees about the accuracy, reliability or benefits of the Community Models. ProfessorF further disclaims any warranty that the Community Model will meet your requirements, be secure, uninterrupted or available at any time or location, or error-free, viruses-free, or that any errors will be corrected, or otherwise. You will be solely responsible for any damage resulting from your use of or access to the Community Models, your downloading of any Community Model, or use of any other Community Model provided by or through ProfessorF.
- Downloads last month
- 16
We're not able to determine the quantization variants.
# !pip install llama-cpp-python from llama_cpp import Llama llm = Llama.from_pretrained( repo_id="professorf/Meta-Llama-3-8B-Instruct-gguf", filename="llama-3-8b-instruct.gguf", )