The Blossom V5 series models is fully trained using high-quality data distilled from gpt-4-0125-preview, resulting in significant improvements.
Introduction
Blossom is a conversational large language model, fine-tuned on the Blossom Orca/Wizard/Chat/Math mixed dataset based on the Meta-Llama-3-8B pre-trained model. Blossom possesses robust general capabilities and context comprehension. Additionally, the high-quality Chinese and English datasets used for training have been made open source.
Training was conducted in two stages. The first stage used 40K Wizard, 40K Orca, 10K Math single-turn instruction datasets, training for 1 epoch; the second stage used 10K Blossom chat multi-turn dialogue dataset, and 10% randomly sampled data from the first stage, training for 3 epochs.
Inference
Inference is performed in the form of dialogue continuation.
Single-turn dialogue
A chat between a human and an artificial intelligence bot. The bot gives helpful, detailed, and polite answers to the human's questions.
|Human|: hello
|Bot|:
Multi-turn dialogue
A chat between a human and an artificial intelligence bot. The bot gives helpful, detailed, and polite answers to the human's questions.
|Human|: hello
|Bot|: Hello! How can I assist you today?<|end_of_text|>
|Human|: Generate a random number using python
|Bot|:
Note: At the end of the Bot's output in the historical conversation, append a <|end_of_text|>.
