Instructions to use prism-ml/Ternary-Bonsai-27B-gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use prism-ml/Ternary-Bonsai-27B-gguf with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="prism-ml/Ternary-Bonsai-27B-gguf",
	filename="Ternary-Bonsai-27B-F16.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use prism-ml/Ternary-Bonsai-27B-gguf with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf prism-ml/Ternary-Bonsai-27B-gguf:F16
# Run inference directly in the terminal:
llama cli -hf prism-ml/Ternary-Bonsai-27B-gguf:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf prism-ml/Ternary-Bonsai-27B-gguf:F16
# Run inference directly in the terminal:
llama cli -hf prism-ml/Ternary-Bonsai-27B-gguf:F16

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf prism-ml/Ternary-Bonsai-27B-gguf:F16
# Run inference directly in the terminal:
./llama-cli -hf prism-ml/Ternary-Bonsai-27B-gguf:F16

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf prism-ml/Ternary-Bonsai-27B-gguf:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf prism-ml/Ternary-Bonsai-27B-gguf:F16

Use Docker

docker model run hf.co/prism-ml/Ternary-Bonsai-27B-gguf:F16

LM Studio
Jan

vLLM

How to use prism-ml/Ternary-Bonsai-27B-gguf with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "prism-ml/Ternary-Bonsai-27B-gguf"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "prism-ml/Ternary-Bonsai-27B-gguf",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/prism-ml/Ternary-Bonsai-27B-gguf:F16

Ollama
How to use prism-ml/Ternary-Bonsai-27B-gguf with Ollama:
```
ollama run hf.co/prism-ml/Ternary-Bonsai-27B-gguf:F16
```

Unsloth Studio

How to use prism-ml/Ternary-Bonsai-27B-gguf with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for prism-ml/Ternary-Bonsai-27B-gguf to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for prism-ml/Ternary-Bonsai-27B-gguf to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for prism-ml/Ternary-Bonsai-27B-gguf to start chatting

Atomic Chat new
Docker Model Runner
How to use prism-ml/Ternary-Bonsai-27B-gguf with Docker Model Runner:
```
docker model run hf.co/prism-ml/Ternary-Bonsai-27B-gguf:F16
```

Lemonade

How to use prism-ml/Ternary-Bonsai-27B-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull prism-ml/Ternary-Bonsai-27B-gguf:F16

Run and chat with the model

lemonade run user.Ternary-Bonsai-27B-gguf-F16

List all available models

lemonade list

Generates gibberish and repeats itself in the middle of thinking and token generation

#19

by rbtasd - opened 3 days ago

Discussion

rbtasd

3 days ago

•

edited 3 days ago

Looks like model is broken or something.
it generates random text and repeats that until the token-windows is filled. An image of the situation in OpenCode is attached below.

Command:

llama-server -m "path-to\Ternary-Bonsai-27B-Q2_0.gguf" --temp 0.7 --top-p 0.95 --top-k 20 -ngl 99
I used other commands too, but no difference.

Sample: opencode

Llama.cpp chat interface

Windows 11
Prism llama.cpp fork
Rtx 3080 ti

SeraphimTimo

3 days ago

•

edited 3 days ago

I've tried it, so far there's no looping, but there are syntax hallucinations while coding.

RTX 5060 Ti, windows 11, prism llama.cpp fork,

@echo off
.\build-prism\llama-server.exe ^
-m "models\ternary-bonsai-27b\Ternary-Bonsai-27B-Q2_0.gguf" ^
--mmproj "models\ternary-bonsai-27b\Ternary-Bonsai-27B-mmproj-BF16.gguf" ^
--alias "ternary-bonsai-27b" ^
--models-max 1 ^
--models-autoload ^
--no-mmap ^
--mlock ^
-np 1 ^
--no-mmproj-offload ^
-ngl 99 ^
--flash-attn on ^
--fit-ctx 8192 ^
--image-min-tokens 1024 ^
-ctk q8_0 ^
-ctv q8_0 ^
--cache-ram 2048 ^
--ctx-checkpoints 8 ^
-t 6 ^
--ubatch-size 1024 ^
--reasoning-budget 2048 ^
--temp 0.7 ^
--top-p 0.95 ^
--top-k 20 ^
--repeat-penalty 1.15

pause

rbtasd

3 days ago

.\build-prism\llama-server.exe ^
-m "models\ternary-bonsai-27b\Ternary-Bonsai-27B-Q2_0.gguf" ^
--mmproj "models\ternary-bonsai-27b\Ternary-Bonsai-27B-mmproj-BF16.gguf" ^
--alias "ternary-bonsai-27b" ^
--models-max 1 ^
--models-autoload ^
--no-mmap ^
--mlock ^
-np 1 ^
--no-mmproj-offload ^
-ngl 99 ^
--flash-attn on ^
--fit-ctx 8192 ^
--image-min-tokens 1024 ^
-ctk q8_0 ^
-ctv q8_0 ^
--cache-ram 2048 ^
--ctx-checkpoints 8 ^
-t 6 ^
--ubatch-size 1024 ^
--reasoning-budget 2048 ^
--temp 0.7 ^
--top-p 0.95 ^
--top-k 20 ^
--repeat-penalty 1.15

I tried with you're command, still the same problem. It's weird, because I don't have this problem with any other models like vanilla Qwen-35b, Gemma12b end a lot more

SeraphimTimo

2 days ago

Yeah, I'm a bit confused as well. Since we're supposed to be using the same model and parameters, I would have expected similar results.
I'm not an expert on CUDA or llama.cpp internals, so this is really just me being curious and trying to compare our setups as closely as possible.

I'm using the official PrismML release prism-b9591-62061f9, specifically:
llama-prism-b1-62061f9-bin-win-cuda-12.4-x64.zip
cudart-llama-bin-win-cuda-12.4-x64.zip
https://github.com/PrismML-Eng/llama.cpp/releases#release-prism-b9591-62061f9

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment