44dummies/ted-llama3-8b-gguf

Ted AI - Llama 3 8B Fine-tuned

No-bullshit AI assistant with quantitative trading expertise and tool-calling capabilities.

Files

File	Description	Size
llama3-base.gguf	Llama 3 8B Q4_K_M base model	~4.6 GB
ted-lora.gguf	Ted LoRA adapter	~161 MB
Modelfile.local	Ollama Modelfile	-

Usage with Ollama

# Download files
wget https://huggingface.co/44dummies/ted-llama3-8b-gguf/resolve/main/llama3-base.gguf
wget https://huggingface.co/44dummies/ted-llama3-8b-gguf/resolve/main/ted-lora.gguf
wget https://huggingface.co/44dummies/ted-llama3-8b-gguf/resolve/main/Modelfile.local

# Create model
ollama create ted -f Modelfile.local

# Run
ollama run ted

Training Details

• Base Model: unsloth/llama-3-8b-bnb-4bit
• Method: LoRA (r=16, alpha=16)
• Training: 100 steps on custom dataset
• Focus: Direct personality, trading knowledge, tool-calling

Personality

Ted is direct, uses dark humor, skips disclaimers, and actually solves problems. Specializes in quantitative trading, risk management, and systematic approaches.