azure99/
blossom-v6.3:36b

17 yesterday

Blossom is a powerful, fully open-source large language model, including the training data.

8b 14b 30b 36b

yesterday

0127f1b07ebf · 22GB ·

model
seed_oss
·
36.2B
·
Q4_K_M
template
A chat between a user and an artificial intelligence assistant. The assistant gives helpful, detaile
params
{ "repeat_penalty": 1.05, "temperature": 0.5, "top_k": 50, "top_p": 0.85 }

Readme

BLOSSOM-V6.3

💻Github • 🚀Blossom Chat Demo

Introduction

Blossom is a powerful open-source conversational large language model that provides reproducible post-training data, dedicated to delivering an open, powerful, and cost-effective locally accessible general-purpose model for everyone.

The Blossom-V6.3 series improves the repeated-output issue in V6.2, adds an MoE version of the 30B-A3B model, and enhances the overall capability of the 8B model.

Chat Model Resource Base Model
Blossom-V6.3-36B Demo GGUF Ollama Seed-OSS-36B-Base
Blossom-V6.3-30B-A3B Demo GGUF Ollama Qwen3-30B-A3B-Base
Blossom-V6.3-14B Demo GGUF Ollama Qwen3-14B-Base
Blossom-V6.3-8B Demo GGUF Ollama Qwen3-8B-Base

You can find the training data here: Blossom-V6.3-SFT-Stage1 (1 epoch)、Blossom-V6.3-SFT-Stage2 (3 epoch).

Data Synthesis Workflow Overview

Primarily employs three cost-effective models: Deepseek-V3.1, Gemini 2.5 Flash, and Qwen3-235B-A22B-Instruct-2507 (denoted as A, B, C)—to regenerate responses under different scenarios using tailored synthesis strategies.

For example:

  • In objective scenarios like mathematics (where answers are unique), Model A first generates responses as a “teacher.” If reference answers exist in the source data, Model B verifies the correctness of A’s responses against them. If no reference answers exist, Model C generates a second response, and Model B checks consistency between A and C’s outputs. Inconsistent responses are filtered out.
  • For subjective scenarios, three models cross-evaluate each other. For instance, Models A and B generate responses to a question, and Model C evaluates which is better. The superior response may be retained as training data or used for preference data construction. To mitigate model bias, roles (respondent/evaluator) are randomly assigned to A, B, and C in each instance.

Additional rule-based filtering is applied, such as:

  • N-Gram filtering to remove data with many repetitions.
  • Discarding questions containing toxic content that triggers teacher model refusals.

Further technical details will be released in the future. The data is synthesized by the 🌸BlossomData framework.