Instructions to use UnicomAI/DeepSeek-DQ3_K_M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use UnicomAI/DeepSeek-DQ3_K_M 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 UnicomAI/DeepSeek-DQ3_K_M:Q3_K_M
# Run inference directly in the terminal:
llama cli -hf UnicomAI/DeepSeek-DQ3_K_M:Q3_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf UnicomAI/DeepSeek-DQ3_K_M:Q3_K_M
# Run inference directly in the terminal:
llama cli -hf UnicomAI/DeepSeek-DQ3_K_M:Q3_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 UnicomAI/DeepSeek-DQ3_K_M:Q3_K_M
# Run inference directly in the terminal:
./llama-cli -hf UnicomAI/DeepSeek-DQ3_K_M:Q3_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 UnicomAI/DeepSeek-DQ3_K_M:Q3_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf UnicomAI/DeepSeek-DQ3_K_M:Q3_K_M

Use Docker

docker model run hf.co/UnicomAI/DeepSeek-DQ3_K_M:Q3_K_M

LM Studio
Jan
Ollama
How to use UnicomAI/DeepSeek-DQ3_K_M with Ollama:
```
ollama run hf.co/UnicomAI/DeepSeek-DQ3_K_M:Q3_K_M
```

Unsloth Studio

How to use UnicomAI/DeepSeek-DQ3_K_M 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 UnicomAI/DeepSeek-DQ3_K_M 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 UnicomAI/DeepSeek-DQ3_K_M to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for UnicomAI/DeepSeek-DQ3_K_M to start chatting

Docker Model Runner
How to use UnicomAI/DeepSeek-DQ3_K_M with Docker Model Runner:
```
docker model run hf.co/UnicomAI/DeepSeek-DQ3_K_M:Q3_K_M
```

Lemonade

How to use UnicomAI/DeepSeek-DQ3_K_M with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull UnicomAI/DeepSeek-DQ3_K_M:Q3_K_M

Run and chat with the model

lemonade run user.DeepSeek-DQ3_K_M-Q3_K_M

List all available models

lemonade list

Atomic Chat

Add pipeline tag and transformers library

by nielsr HF Staff - opened Jun 16, 2025

base: refs/heads/main

←

from: refs/pr/2

Discussion Files changed

-3

Files changed (1) hide show

README.md +4 -3

README.md CHANGED Viewed

@@ -1,13 +1,14 @@
 ---
-license: apache-2.0
 base_model:
 - deepseek-ai/DeepSeek-R1
 - deepseek-ai/DeepSeek-V3
 ---
 # Quantitative Analysis of Performance Drop in DeepSeek Model Quantization
 Enbo Zhao<sup>1,2</sup>, Yi Shen<sup>1,2</sup>, Shuming Shi<sup>1,2</sup>, Jieyun Huang<sup>1,2</sup>, Zhihao Chen<sup>1,2</sup>, Ning Wang<sup>1,2</sup>, Siqi Xiao<sup>1,2</sup>, Jian Zhang<sup>1,2</sup>, Kai Wang<sup>1,2</sup>, Shiguo Lian<sup>1,2</sup>
@@ -89,4 +90,4 @@ Recently, there is a high demand for deploying DeepSeek-R1 and V3 locally, possi
 | C‑Eval          | -               | 87.0                    | 86.7                | 86.8                 | 86.2                 |
 | **Average**     | -               | 77.78                   | 77.65               | 77.91                | 77.35                |
 | **Weighted avg.**   | -               | 79.94                   | 79.71               | 79.97                | 79.40                |
-| **Accuracy drop**   | -               | -                       | 0.29%               | 0                    | 0.68%                |

 ---
 base_model:
 - deepseek-ai/DeepSeek-R1
 - deepseek-ai/DeepSeek-V3
+license: apache-2.0
+pipeline_tag: text-generation
+library_name: transformers
 ---
 # Quantitative Analysis of Performance Drop in DeepSeek Model Quantization
 Enbo Zhao<sup>1,2</sup>, Yi Shen<sup>1,2</sup>, Shuming Shi<sup>1,2</sup>, Jieyun Huang<sup>1,2</sup>, Zhihao Chen<sup>1,2</sup>, Ning Wang<sup>1,2</sup>, Siqi Xiao<sup>1,2</sup>, Jian Zhang<sup>1,2</sup>, Kai Wang<sup>1,2</sup>, Shiguo Lian<sup>1,2</sup>
 | C‑Eval          | -               | 87.0                    | 86.7                | 86.8                 | 86.2                 |
 | **Average**     | -               | 77.78                   | 77.65               | 77.91                | 77.35                |
 | **Weighted avg.**   | -               | 79.94                   | 79.71               | 79.97                | 79.40                |
+| **Accuracy drop**   | -               | -                       | 0.29%               | 0                  | 0.68%                |