Instructions to use inclusionAI/LLaDA-MoE-7B-A1B-Base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use inclusionAI/LLaDA-MoE-7B-A1B-Base with Transformers:

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

pipe = pipeline("text-generation", model="inclusionAI/LLaDA-MoE-7B-A1B-Base", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("inclusionAI/LLaDA-MoE-7B-A1B-Base", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use inclusionAI/LLaDA-MoE-7B-A1B-Base with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "inclusionAI/LLaDA-MoE-7B-A1B-Base"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "inclusionAI/LLaDA-MoE-7B-A1B-Base",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/inclusionAI/LLaDA-MoE-7B-A1B-Base

SGLang

How to use inclusionAI/LLaDA-MoE-7B-A1B-Base 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 "inclusionAI/LLaDA-MoE-7B-A1B-Base" \
    --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": "inclusionAI/LLaDA-MoE-7B-A1B-Base",
		"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 "inclusionAI/LLaDA-MoE-7B-A1B-Base" \
        --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": "inclusionAI/LLaDA-MoE-7B-A1B-Base",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use inclusionAI/LLaDA-MoE-7B-A1B-Base with Docker Model Runner:
```
docker model run hf.co/inclusionAI/LLaDA-MoE-7B-A1B-Base
```

Request to open source the training code.

by win10 - opened Sep 11, 2025

Discussion

win10

Sep 11, 2025

Request to open source the training source code. This is not transparent, nor is it a healthy open source project. If you want to promote community developer participation, please truly open source it.

m1ngcheng

inclusionAI org Sep 14, 2025

Request to open source the training source code. This is not transparent, nor is it a healthy open source project. If you want to promote community developer participation, please truly open source it.

Hi, we are finalizing the documentation and testing for this part of the scripts and other content, and we will open source it on GitHub. We will notify you at that time and update the link in the README.

win10

Sep 14, 2025

Request to open source the training source code. This is not transparent, nor is it a healthy open source project. If you want to promote community developer participation, please truly open source it.

Hi, we are finalizing the documentation and testing for this part of the scripts and other content, and we will open source it on GitHub. We will notify you at that time and update the link in the README.

I may have spoken a little too harshly.Without the official training code library reference, it would be difficult to carry out secondary development or migrate the architecture of existing models.

ZHZisZZ

Sep 21, 2025

@win10

Hi there, we’ve built dllm-trainer, a lightweight finetuning framework for diffusion language models on top of the 🤗 Transformers Trainer. Give it a try if you’d like to finetune LLaDA / LLaDA-MoE and Dream.

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