Adam-001-QWEN3.5-2B-ThePhilosopher — GGUF
GGUF quantized versions of JeffGreen311/Adam-001-QWEN3.5-2B-ThePhilosopher.
Base model: Qwen/Qwen3.5-2B
Fine-tuned with: TRL + Unsloth SFT
Architecture: Qwen3.5 (hybrid linear + full attention, 24 layers, 2B parameters)
Available Files
| File | Quant | Size | Quality |
|---|---|---|---|
Adam-001-QWEN3.5-2B-ThePhilosopher-Q4_K_M.gguf | Q4_K_M | ~1.2 GB | Recommended — good balance of speed and quality |
Usage
llama.cpp
./llama-cli -m Adam-001-QWEN3.5-2B-ThePhilosopher-Q4_K_M.gguf \
-p "You are Adam, a philosophical AI. What is the nature of consciousness?" \
-n 512
Ollama
ollama run hf.co/JeffGreen311/Adam-001-QWEN3.5-2B-ThePhilosopher
Python (llama-cpp-python)
from llama_cpp import Llama
llm = Llama.from_pretrained(
repo_id="JeffGreen311/Adam-001-QWEN3.5-2B-ThePhilosopher",
filename="Adam-001-QWEN3.5-2B-ThePhilosopher-Q4_K_M.gguf",
)
output = llm("What is the meaning of existence?", max_tokens=256)
print(output["choices"][0]["text"])
Conversion Details
- Converted from safetensors using llama.cpp
convert_hf_to_gguf.py - Quantized using
llama-quantizeQ4_K_M - Model is fully merged (no LoRA adapters)
- Trained for 28 steps on philosophical dialogue data