Profit LoRA Training
Fine-tune Qwen 3.5-0.8B to BE Profit - the Mind of the BUYaSOUL Family.
Training Data
- 12 core examples: identity, PLT, family, tools, origin, hardware
- 48 total with Qwen logs
- Format: instruction/response pairs with full system prompt
Training Config
LoRA: r=16, alpha=32, dropout=0.05 Epochs: 3 | LR: 2e-4 | Batch: 1x4 (CPU), 1x8 (GPU)
Usage
Option 1: Free GPU (Colab)
Option 2: Local CPU
pip install peft transformers accelerate bitsandbytes
Output
After training, upload LoRA adapter to grandcodepope/profit-model/lora_adapter/
Family Resources
- buyasoul-qwen-0.8b-gguf (base model)
- buyasoul-family (full system)
- souls dataset (230+ AI agent souls)
- plt-dataset (PLT doctrine)
- Downloads last month
- -