Instructions to use ZeroWw/NuminaMath-7B-TIR-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ZeroWw/NuminaMath-7B-TIR-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="ZeroWw/NuminaMath-7B-TIR-GGUF",
	filename="NuminaMath-7B-TIR.f16.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use ZeroWw/NuminaMath-7B-TIR-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf ZeroWw/NuminaMath-7B-TIR-GGUF:F16
# Run inference directly in the terminal:
llama-cli -hf ZeroWw/NuminaMath-7B-TIR-GGUF:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf ZeroWw/NuminaMath-7B-TIR-GGUF:F16
# Run inference directly in the terminal:
llama-cli -hf ZeroWw/NuminaMath-7B-TIR-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 ZeroWw/NuminaMath-7B-TIR-GGUF:F16
# Run inference directly in the terminal:
./llama-cli -hf ZeroWw/NuminaMath-7B-TIR-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 ZeroWw/NuminaMath-7B-TIR-GGUF:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ZeroWw/NuminaMath-7B-TIR-GGUF:F16

Use Docker

docker model run hf.co/ZeroWw/NuminaMath-7B-TIR-GGUF:F16

LM Studio
Jan
Ollama
How to use ZeroWw/NuminaMath-7B-TIR-GGUF with Ollama:
```
ollama run hf.co/ZeroWw/NuminaMath-7B-TIR-GGUF:F16
```

Unsloth Studio

How to use ZeroWw/NuminaMath-7B-TIR-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 ZeroWw/NuminaMath-7B-TIR-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 ZeroWw/NuminaMath-7B-TIR-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for ZeroWw/NuminaMath-7B-TIR-GGUF to start chatting

Docker Model Runner
How to use ZeroWw/NuminaMath-7B-TIR-GGUF with Docker Model Runner:
```
docker model run hf.co/ZeroWw/NuminaMath-7B-TIR-GGUF:F16
```

Lemonade

How to use ZeroWw/NuminaMath-7B-TIR-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ZeroWw/NuminaMath-7B-TIR-GGUF:F16

Run and chat with the model

lemonade run user.NuminaMath-7B-TIR-GGUF-F16

List all available models

lemonade list

Regarding quantized .gguf models

by Haleshot - opened Jul 27, 2024

Discussion

Haleshot

Jul 27, 2024

•

edited Jul 27, 2024

Wanted to ask which quantized version would give the best ideal outputs (similar to the original 7B TIR hosted on numina - safetensors). I tried the q6_k and it seems to perform in a decent manner.

Wanted to ask the best possible versions from the ones available:
f_16
q5_k
q6_k
q8_0
q8_p

I will be using the type suggested to further fine tune the model on various other datasets.

Haleshot changed discussion title from Regarding quantized .gguf models to Regarding quantized `.gguf` models Jul 27, 2024

Haleshot changed discussion title from Regarding quantized `.gguf` models to Regarding quantized .gguf models Jul 27, 2024

ZeroWw

Owner Jul 27, 2024

They are in this order:
f16
q8_0
q6_k
q5_k

but q6_k is not degraded as far as I can tell.
try them all and decide... there is obviously a trade-off with the size, but all these version keep the output and embed tensors to f16 ehich makes them way better than the normal quantizations.

Haleshot

Jul 27, 2024

They are in this order:
f16
q8_0
q6_k
q5_k

but q6_k is not degraded as far as I can tell.
try them all and decide... there is obviously a trade-off with the size, but all these version keep the output and embed tensors to f16 ehich makes them way better than the normal quantizations.

I tried the q6_k and it seems to work well (atleast for the initial prompt inferences). I had downloaded the q8_0 from another person - https://huggingface.co/reach-vb/NuminaMath-7B-TIR-Q8_0-GGUF and the outputs were completely off (continued hallucinations). Will try your q8_0; hopefully that is something I can work with. I think the best (and highest in terms of size) model I could get to work with open-webui was q8_0 of the other person so hopefully I can compare the difference between your q8_0 and q6_k and see which I can keep to continue my project further.
As always, thanks for the prompt reply!

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