Downloads · 30 days
19
4% of all-time downloads
navicore/fortean-qwen3-8b-advanced-GGUF
fortean-qwen3-8b-advanced-GGUF is a text generation model from navicore. Use it when you need the model to write or continue text. It is set up for gguf. The card lists the license as apache-2.0.
This is a GGUF quantized version of the Fortean-Qwen3-8B model, fine-tuned to emulate Charles Fort's distinctive writing style about anomalous phenomena.
Downloads · 30 days
19
4% of all-time downloads
All-time downloads
445
Public
Repo size
10.1 GB
Likes
0
Public
Click a slice to open those files.
.gguf10.1 GB · 100%
From the Hugging Face model README
This is a GGUF quantized version of the Fortean-Qwen3-8B model, fine-tuned to emulate Charles Fort's distinctive writing style about anomalous phenomena.
ollama create fortean -f Modelfile.final
ollama run fortean
User: What are your thoughts on the Bermuda Triangle?
Charles Fort: The Bermuda Triangle, that persistent enigma that haunts our collective imagination, presents patterns that would alarm those who prefer their reality comfortably predictable. I've documented instances where similar manifestations occur across multiple continents with a periodicity that transcends local explanation...
For best results, use this prompt format:
Question: [your question here]
Charles Fort:
Configure these stop tokens to prevent multiple responses:
<|endoftext|><|im_end|>\nQuestion:\n\nQuestion:FROM fortean-q4_k_m.gguf
TEMPLATE """Question: {{ .Prompt }}
Charles Fort: {{ .Response }}"""
PARAMETER temperature 0.8
PARAMETER top_p 0.9
PARAMETER repeat_penalty 1.1
PARAMETER num_predict 300
PARAMETER stop "<|endoftext|>"
PARAMETER stop "<|im_end|>"
PARAMETER stop "\nQuestion:"
PARAMETER stop "\n\nQuestion:"
./main -m fortean-q4_k_m.gguf -p "Question: What are your thoughts on red rain?\n\nCharles Fort:" -n 300 --temp 0.8
from llama_cpp import Llama
llm = Llama(model_path="fortean-q4_k_m.gguf", n_ctx=2048)
prompt = "Question: What are your thoughts on red rain?\n\nCharles Fort:"
response = llm(prompt, max_tokens=300, temperature=0.8, stop=["<|endoftext|>", "\nQuestion:"])
print(response['choices'][0]['text'])
Fine-tuned by navicore using Charles Fort's complete works from Project Gutenberg.