tags:
- llama-3
- gguf
- uncensored
- roleplay
license: llama3.1
language: - en
base_model: - meta-llama/Llama-3.1-8B-Instruct
pipeline_tag: text-generation
🌑 DarkIdol 1.3 - GGUF (CPU/Mac)
This is the GGUF (CPU/Apple Silicon) version of the DarkIdol 1.3 Uncensored model.
For the high-speed GPU version (EXL2), go here.
📥 Which version should I download?
| Filename | RAM Required | Use Case |
|---|---|---|
| DarkIdol-1.3-Q4_K_M.gguf | ~6-8 GB | ✅ Best for most users. Fast, low memory, negligible quality loss. |
| DarkIdol-1.3-Q8_0.gguf | ~10-12 GB | 💎 High Fidelity. Near-lossless. Recommended if you have 16GB+ RAM. |
| DarkIdol-1.3-Uncensored-FP16.gguf | ~18 GB | 💾 Uncompressed Master. Only for archiving or re-quantization. |
🚀 How to Run (llama.cpp)
./llama-cli -m DarkIdol-1.3-Q4_K_M.gguf \
-p "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nYou are DarkIdol.<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nHello!<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n" \
-n 512