ewchampion/Ternary-Bonsai-8B-TQ2_0-GGUF overview
Ternary Bonsai 8B TQ2 0 GGUF Requantized from prism ml/Ternary Bonsai 8B gguf https://huggingface.co/prism ml/Ternary Bonsai 8B gguf 's F16 baseline to GGML's …
Runs locally from ~2.47 GB disk (4 GB VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
| File | Type | Quantization | Size | Link |
|---|---|---|---|---|
| Ternary-Bonsai-8B-TQ2_0.gguf | GGUF | GGUF | 2.47 GB | Download |
Model Details
| Model ID | ewchampion/Ternary-Bonsai-8B-TQ2_0-GGUF |
|---|---|
| Author | ewchampion |
| Pipeline | — |
| License | apache-2.0 |
| Base model | prism-ml/Ternary-Bonsai-8B-gguf |
| Last modified | 2026-07-09T16:53:00.000Z |
Model README
---
license: apache-2.0
base_model: prism-ml/Ternary-Bonsai-8B-gguf
---
Ternary-Bonsai-8B-TQ2_0-GGUF
Requantized from prism-ml/Ternary-Bonsai-8B-gguf's F16 baseline to GGML's native TQ2_0 ternary type (~2.06 bits/weight), using stock llama.cpp (tag b9498) — not PrismML's custom fork/packing.
Sanity-checked via llama-server's OpenAI-compatible endpoint on CPU (Metal has no TQ1_0/TQ2_0 matmul kernels at this llama.cpp version — this model must run with -ngl 0 / CPU-only until upstream adds Metal support):
- Chat coherence: pass ("What is the capital of France?" -> "The capital of France is Paris.")
- Native tool-calling (Qwen3
<tool_call>format): pass, clean structuredtool_callsoutput.
License: Apache 2.0 (carried over from base model).
Run ewchampion/Ternary-Bonsai-8B-TQ2_0-GGUF with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.
Source: Hugging Face · Compare models