Instructions to use naos-ku/GraphTokenLM with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use naos-ku/GraphTokenLM with Transformers:

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

pipe = pipeline("text-generation", model="naos-ku/GraphTokenLM", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("naos-ku/GraphTokenLM", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use naos-ku/GraphTokenLM with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/naos-ku/GraphTokenLM

SGLang

How to use naos-ku/GraphTokenLM 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 "naos-ku/GraphTokenLM" \
    --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": "naos-ku/GraphTokenLM",
		"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 "naos-ku/GraphTokenLM" \
        --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": "naos-ku/GraphTokenLM",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use naos-ku/GraphTokenLM with Docker Model Runner:
```
docker model run hf.co/naos-ku/GraphTokenLM
```

GraphTokenLM / config.json

naos-ku

Export GraphTokenLM to Hugging Face Hub

a342f8b 7 months ago

Raw History Blame Contribute Delete

871 Bytes

	{
	"architectures": [
	"GraphTokenLM"
	],
	"base_model": "Qwen/Qwen3-4B-Base",
	"bos_token_id": 151643,
	"dtype": "float32",
	"enable_lora": false,
	"eos_token_id": 151643,
	"freeze_llm": true,
	"gnn_hidden": 64,
	"gnn_hidden_dim": 64,
	"gnn_out": 64,
	"gnn_out_dim": 64,
	"gnn_type": "GIN",
	"graph_pooling": [
	"mean"
	],
	"hidden_size": 2560,
	"llm_name": "Qwen/Qwen3-4B-Base",
	"lpe_dim": 8,
	"model_type": "graph_token_lm",
	"node_feat_dim": 8,
	"node_pos_emb_dim": 8,
	"num_attention_heads": 32,
	"num_gnn_layers": 3,
	"num_graph_tokens": 4,
	"num_hidden_layers": 36,
	"num_max_nodes": 20,
	"num_proj_layers": 2,
	"pos_emb_dim": 8,
	"transformers_version": "4.57.1",
	"use_degree_emb": false,
	"vocab_size": 151936,
	"auto_map": {
	"AutoConfig": "glm.GraphTokenLMConfig",
	"AutoModelForCausalLM": "glm.GraphTokenLM"
	}
	}