Model Overview

The 3tic/Orion-Qwen3-1.7B-CPT-v2603 is a 2 billion parameter base model derived from the Qwen3-1.7B-Base architecture. Developed by 3tic, this model has undergone extensive continuous pre-training (CPT) on a specialized dataset exceeding 20 billion tokens of Chinese and Japanese light novel content. It features a substantial context window of 32768 tokens.

Key Capabilities

• Specialized Domain Knowledge: Optimized for processing and generating text related to Chinese and Japanese light novels.
• Large Context Window: Supports a 32768 token context length, beneficial for understanding long-form narratives.
• Foundation Model: Serves as a robust base model for further fine-tuning on specific downstream tasks within its specialized domain.

Ideal Use Cases

• Light Novel Generation: Creating new content or expanding existing narratives in the style of Chinese and Japanese light novels.
• Text Analysis: Analyzing and extracting information from extensive light novel texts.
• Domain-Specific Fine-tuning: As a starting point for developing applications requiring deep understanding of light novel language and themes.