Instructions to use TheBloke/deepseek-coder-6.7B-instruct-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use TheBloke/deepseek-coder-6.7B-instruct-GGUF with Transformers:

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("TheBloke/deepseek-coder-6.7B-instruct-GGUF", dtype="auto")

llama-cpp-python

How to use TheBloke/deepseek-coder-6.7B-instruct-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="TheBloke/deepseek-coder-6.7B-instruct-GGUF",
	filename="deepseek-coder-6.7b-instruct.Q2_K.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use TheBloke/deepseek-coder-6.7B-instruct-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf TheBloke/deepseek-coder-6.7B-instruct-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf TheBloke/deepseek-coder-6.7B-instruct-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf TheBloke/deepseek-coder-6.7B-instruct-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf TheBloke/deepseek-coder-6.7B-instruct-GGUF:Q4_K_M

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 TheBloke/deepseek-coder-6.7B-instruct-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf TheBloke/deepseek-coder-6.7B-instruct-GGUF:Q4_K_M

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 TheBloke/deepseek-coder-6.7B-instruct-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf TheBloke/deepseek-coder-6.7B-instruct-GGUF:Q4_K_M

Use Docker

docker model run hf.co/TheBloke/deepseek-coder-6.7B-instruct-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use TheBloke/deepseek-coder-6.7B-instruct-GGUF with Ollama:
```
ollama run hf.co/TheBloke/deepseek-coder-6.7B-instruct-GGUF:Q4_K_M
```

Unsloth Studio new

How to use TheBloke/deepseek-coder-6.7B-instruct-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 TheBloke/deepseek-coder-6.7B-instruct-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 TheBloke/deepseek-coder-6.7B-instruct-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for TheBloke/deepseek-coder-6.7B-instruct-GGUF to start chatting

Docker Model Runner
How to use TheBloke/deepseek-coder-6.7B-instruct-GGUF with Docker Model Runner:
```
docker model run hf.co/TheBloke/deepseek-coder-6.7B-instruct-GGUF:Q4_K_M
```

Lemonade

How to use TheBloke/deepseek-coder-6.7B-instruct-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull TheBloke/deepseek-coder-6.7B-instruct-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.deepseek-coder-6.7B-instruct-GGUF-Q4_K_M

List all available models

lemonade list

Missing tokenizer.model file

by whatever1983 - opened Nov 12, 2023

Discussion

whatever1983

Nov 12, 2023

@TheBloke : how did you use llama.cpp ./quantize to get those GGUFs when the tokenizer.model file isn't even uploaded by the model creator? If you managed to do a tokenizer.model file yourself, care to share the file on this repo? Also for the 33B one.

Thanks

TheBloke

Owner Nov 12, 2023

Yeah i couldn't make it using the standard Llama.cpp convert.py. Fortunately there was a PR to add support for the HF tokenizer format. That had a few problems and bugs at first but after a few fixes I was able to make working GGUFs.

That PR hasn't been merged as its still being reviewed. But if you want to make your own GGUFs of this model or any other without tokenizer.model you can use the covert.py from https://github.com/ggerganov/llama.cpp/pull/3633

eramax

Dec 5, 2023

•

edited Dec 5, 2023

Hey @TheBloke , Please provide a quantize for this model: https://huggingface.co/ise-uiuc/Magicoder-S-DS-6.7B I followed the same strategy (PR), but I wasn't successful.

LiuCi

Mar 16, 2024

@TheBloke hello, accorfing to you suggestions, it looks like the convet.py from https://github.com/ggerganov/llama.cpp/pull/3633 have been merged , so I have tried to quantize deepseek model by self but it failed like this. I have find many issuse on the github but can not fix it, I'd appreciate it if you had some advice.

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