Instructions to use tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF with Transformers:

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

pipe = pipeline("text-generation", model="tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF", dtype="auto")

llama-cpp-python

How to use tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF",
	filename="LongWriter-Zero-32B-Q2_K.gguf",
)

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

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF:Q2_K
# Run inference directly in the terminal:
llama-cli -hf tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF:Q2_K

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF:Q2_K
# Run inference directly in the terminal:
llama-cli -hf tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF:Q2_K

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 tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF:Q2_K
# Run inference directly in the terminal:
./llama-cli -hf tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF:Q2_K

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 tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF:Q2_K
# Run inference directly in the terminal:
./build/bin/llama-cli -hf tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF:Q2_K

Use Docker

docker model run hf.co/tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF:Q2_K

LM Studio
Jan

vLLM

How to use tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tensorblock/THU-KEG_LongWriter-Zero-32B-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": "tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF:Q2_K

SGLang

How to use tensorblock/THU-KEG_LongWriter-Zero-32B-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 "tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF with Ollama:
```
ollama run hf.co/tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF:Q2_K
```

Unsloth Studio new

How to use tensorblock/THU-KEG_LongWriter-Zero-32B-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 tensorblock/THU-KEG_LongWriter-Zero-32B-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 tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF to start chatting

Docker Model Runner
How to use tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF with Docker Model Runner:
```
docker model run hf.co/tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF:Q2_K
```

Lemonade

How to use tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF:Q2_K

Run and chat with the model

lemonade run user.THU-KEG_LongWriter-Zero-32B-GGUF-Q2_K

List all available models

lemonade list

THU-KEG/LongWriter-Zero-32B - GGUF

Join our Discord to learn more about what we're building ↗

This repo contains GGUF format model files for THU-KEG/LongWriter-Zero-32B.

The files were quantized using machines provided by TensorBlock, and they are compatible with llama.cpp as of commit b5753.

Our projects

Forge

An OpenAI-compatible multi-provider routing layer.
🚀 Try it now! 🚀
Awesome MCP Servers	TensorBlock Studio

A comprehensive collection of Model Context Protocol (MCP) servers.	A lightweight, open, and extensible multi-LLM interaction studio.
👀 See what we built 👀	👀 See what we built 👀

Prompt template

A conversation between the user and the assistant. The user provides a writing/general task, and the assistant completes it. The assistant first deeply thinks through the writing/answering process in their mind before providing the final written work to the user. The assistant should engage in comprehensive and in-depth planning to ensure that every aspect of the writing/general task is detailed and well-structured. If there is any uncertainty or ambiguity in the writing request, the assistant should reflect, ask themselves clarifying questions, and explore multiple writing approaches to ensure the final output meets the highest quality standards. Since writing is both a creative and structured task, the assistant should analyze it from multiple perspectives, considering coherence, clarity, style, tone, audience, purpose, etc.. Additionally, the assistant should review and refine the work to enhance its expressiveness. The writing thought process and the final written work should be enclosed within <think> </think> and <answer> </answer> tags, respectively, as shown below: <think>A comprehensive strategy for writing that encompasses detailed planning and structural design—including brainstorming, outlining, style selection, audience adaptation, self-reflection, quality assurance, etc..</think> <answer>The final written work after thorough optimization and refinement.</answer> <|user|>: {system_prompt} <|assistant|>:

Model file specification

Filename	Quant type	File Size	Description
LongWriter-Zero-32B-Q2_K.gguf	Q2_K	12.313 GB	smallest, significant quality loss - not recommended for most purposes
LongWriter-Zero-32B-Q3_K_S.gguf	Q3_K_S	14.392 GB	very small, high quality loss
LongWriter-Zero-32B-Q3_K_M.gguf	Q3_K_M	15.935 GB	very small, high quality loss
LongWriter-Zero-32B-Q3_K_L.gguf	Q3_K_L	17.247 GB	small, substantial quality loss
LongWriter-Zero-32B-Q4_0.gguf	Q4_0	18.640 GB	legacy; small, very high quality loss - prefer using Q3_K_M
LongWriter-Zero-32B-Q4_K_S.gguf	Q4_K_S	18.784 GB	small, greater quality loss
LongWriter-Zero-32B-Q4_K_M.gguf	Q4_K_M	19.851 GB	medium, balanced quality - recommended
LongWriter-Zero-32B-Q5_0.gguf	Q5_0	22.638 GB	legacy; medium, balanced quality - prefer using Q4_K_M
LongWriter-Zero-32B-Q5_K_S.gguf	Q5_K_S	22.638 GB	large, low quality loss - recommended
LongWriter-Zero-32B-Q5_K_M.gguf	Q5_K_M	23.262 GB	large, very low quality loss - recommended
LongWriter-Zero-32B-Q6_K.gguf	Q6_K	26.886 GB	very large, extremely low quality loss
LongWriter-Zero-32B-Q8_0.gguf	Q8_0	34.821 GB	very large, extremely low quality loss - not recommended

Downloading instruction

Command line

Firstly, install Huggingface Client

pip install -U "huggingface_hub[cli]"

Then, downoad the individual model file the a local directory

huggingface-cli download tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF --include "LongWriter-Zero-32B-Q2_K.gguf" --local-dir MY_LOCAL_DIR

If you wanna download multiple model files with a pattern (e.g., *Q4_K*gguf), you can try:

huggingface-cli download tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF --local-dir MY_LOCAL_DIR --local-dir-use-symlinks False --include='*Q4_K*gguf'

Downloads last month: 24

GGUF

Model size

33B params

Architecture

qwen2

Hardware compatibility

2-bit

3-bit

Model tree for tensorblock/THU-KEG_LongWriter-Zero-32B-GGUF

Base model

Qwen/Qwen2.5-32B

Finetuned

THU-KEG/LongWriter-Zero-32B

Quantized

(9)

this model

tensorblock
/

THU-KEG_LongWriter-Zero-32B-GGUF