llama3-platypus

Supervised Fine-Tuned Model

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the open_platypus dataset. It achieves the following results on the evaluation set:

• Loss: 0.6769
• Accuracy: 0.8116

Model description

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the open_platypus dataset.

Intended uses & limitations

How to use

You can use this model directly with a pipeline for text classification. Here is an example:

import logging
from transformers import pipeline

# Set up logging
logging.basicConfig(filename='model_output.log', level=logging.INFO)

# Load the model using the pipeline API for text generation
model_name = "2nji/llama3-platypus"
generator = pipeline('text-generation', model=model_name)

# Example prompt
prompt = "Hello! How can AI help humans in daily life?"

# Generate response
try:
    responses = generator(prompt, max_length=50)  # Adjust max_length as needed
    response_text = responses[0]['generated_text']
    print("Model response:", response_text)

    # Log the output
    logging.info("Sent prompt: %s", prompt)
    logging.info("Received response: %s", response_text)

except Exception as e:
    logging.error("Error in generating response: %s", str(e))

Training and evaluation data

The model was fine-tuned on the open_platypus dataset.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 3.0

Training results

The model was trained on a single NVIDIA H100 GPU with the following results:

• Loss: 0.6769
• Accuracy: 0.8116

Framework versions

• PEFT 0.11.1
• Transformers 4.42.3
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1