Downloads · 30 days
5
7% of all-time downloads
dennisonb/qwen25-tax-3b-v2-GGUF
qwen25-tax-3b-v2-GGUF is a machine learning model from dennisonb. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Quantized GGUF of the v2 IRS Tax Code fine-tune of Qwen2.5-3B-Instruct.
Downloads · 30 days
5
7% of all-time downloads
All-time downloads
67
Public
Repo size
3.3 GB
Likes
0
Public
Click a slice to open those files.
.gguf3.3 GB · 100%
From the Hugging Face model README
Quantized GGUF of the v2 IRS Tax Code fine-tune of Qwen2.5-3B-Instruct.
The GGUF was produced by fusing all three training stages (SFT → DPO → GRPO) and exporting via mlx_lm.fuse + llama.cpp convert_hf_to_gguf.
| File | Quantization | Size |
|---|---|---|
qwen25-tax-3b-v2-q8_0.gguf | Q8_0 | ~3.3 GB |
# Modelfile
cat > Modelfile << 'EOF'
FROM qwen25-tax-3b-v2-q8_0.gguf
SYSTEM "You are a tax law assistant trained on the Internal Revenue Code (Title 26) and Treasury Regulations (26 CFR). You answer questions about US federal tax law accurately, cite relevant IRC sections, and note important exceptions and limitations. You do not provide personalised tax advice; always recommend consulting a qualified tax professional for individual situations."
PARAMETER temperature 0.3
PARAMETER top_p 0.9
PARAMETER top_k 40
PARAMETER repeat_penalty 1.1
PARAMETER num_ctx 4096
PARAMETER stop "<|endoftext|>"
PARAMETER stop "<|im_end|>"
EOF
ollama create qwen25-tax-3b-v2 -f Modelfile
ollama run qwen25-tax-3b-v2
See dennisonb/qwen25-tax-3b-v2-adapters for the raw LoRA adapter weights (SFT / DPO / GRPO).