Instructions to use ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use ShayonSarker/DeepSeek-Coder-1.3B-Base-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 ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf ShayonSarker/DeepSeek-Coder-1.3B-Base-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 ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf ShayonSarker/DeepSeek-Coder-1.3B-Base-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 ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF:Q4_K_M

Use Docker

docker model run hf.co/ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF:Q4_K_M

SGLang

How to use ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Ollama
How to use ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF with Ollama:
```
ollama run hf.co/ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF:Q4_K_M
```
Unsloth Desktop
Docker Model Runner
How to use ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF with Docker Model Runner:
```
docker model run hf.co/ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF:Q4_K_M
```

Lemonade

How to use ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.DeepSeek-Coder-1.3B-Base-GGUF-Q4_K_M

List all available models

lemonade list

Atomic Chat

DeepSeek-Coder 1.3B Base GGUF

Source model · HF release · Build hub

Pinned llama.cpp conversion of the official 1.3B base model. The previously listed deepseek-ai/deepseek-coder-1b repository does not exist; this release uses the official 1.3B base checkpoint.

Formats

File	Status	Purpose
`deepseek-coder-1.3b-base-F16.gguf`	Published	Reference quality
`deepseek-coder-1.3b-base-Q8_0.gguf`	Published	Higher-quality compact format
`deepseek-coder-1.3b-base-Q4_K_M.gguf`	Published	Smallest release format

Verified code completions

Verbatim deepseek-coder-1.3b-base-Q4_K_M.gguf completions, --temp 0, 24 new tokens. This is the base checkpoint, so it continues code rather than answering instructions. The prompt is shown as the def/for/SELECT line plus the already-given return or print(, and the model's text is reproduced as-is.

Prompt	Model completion
`def is_even(number):` `return`	`number % 2 == 0`
`def reverse_text(text):` `return`	`text[::-1]`
`def factorial(number):` `if number <= 1:` `return`	`1` `else:` `return number * factorial(number - 1)`
`def sum_list(numbers):` `return`	`sum(numbers)`
`def find_max(numbers):` `return`	`max(numbers)`
`for index, item in enumerate(items):` `print(`	`index, item)`

Python body completions are correct in all five cases. SQL is not: SELECT name FROM users WHERE active = continues into unrelated JavaScript-style ORM code, and SELECT COUNT(*) FROM orders WHERE total > returns a bare number. The def-style prompts are also filled with boilerplate follow-ups such as a generated is_odd helper, trimmed above.

Validation

WikiText-2 raw test evaluation, 8 chunks of 512 tokens. Lower perplexity is better.

Format	PPL	Ratio to F16
F16	18.0150	Baseline
Q8_0	18.0253	1.0006
Q4_K_M	18.3160	1.0167

A deterministic Q4_K_M code-completion smoke test completed successfully:

def add(a, b):
    return a + b

This is a base completion model, not an instruction-tuned assistant.

Build

python -m pip install -r requirements-build.txt
python build_gguf.py --model-id deepseek-ai/deepseek-coder-1.3b-base

The builder pins llama.cpp commit 6b790a9c291b5d7af3312bbf9f0c558aa023b13e and the upstream model revision. It does not upload or overwrite this repository.

License

DeepSeek Coder Model License v1.0. redistribution is allowed only under the license conditions, including the use-based restrictions. LICENSE contains the complete terms and must accompany redistributed derivatives.

Downloads last month: -

GGUF

Model size

1B params

Architecture

llama

Hardware compatibility

4-bit

8-bit

16-bit

Model tree for ShayonSarker/DeepSeek-Coder-1.3B-Base-GGUF

Base model

deepseek-ai/deepseek-coder-1.3b-base

Quantized

(19)

this model