Instructions to use mtzig/qwen3_decoder_small with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mtzig/qwen3_decoder_small with Transformers:

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

pipe = pipeline("text-generation", model="mtzig/qwen3_decoder_small")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("mtzig/qwen3_decoder_small", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use mtzig/qwen3_decoder_small with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mtzig/qwen3_decoder_small"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mtzig/qwen3_decoder_small",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/mtzig/qwen3_decoder_small

SGLang

How to use mtzig/qwen3_decoder_small 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 "mtzig/qwen3_decoder_small" \
    --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": "mtzig/qwen3_decoder_small",
		"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 "mtzig/qwen3_decoder_small" \
        --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": "mtzig/qwen3_decoder_small",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use mtzig/qwen3_decoder_small with Docker Model Runner:
```
docker model run hf.co/mtzig/qwen3_decoder_small
```

qwen3_decoder_small

File size: 1,809 Bytes

0264a1b

{
  "adapter_alpha": 1.0,
  "architectures": [
    "AttnQwenForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "dtype": "float32",
  "eos_token_id": 151645,
  "from_pretrained_init": false,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 1024,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "layer_types": [
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention"
  ],
  "low_rank": null,
  "lowrank_link": 16,
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "mod_k_proj": true,
  "mod_keys": true,
  "mod_output": true,
  "mod_queries": true,
  "mod_v_proj": true,
  "mod_vals": false,
  "modattn_layers": "all",
  "modattn_type": "decoder",
  "model_type": "attnqwen3",
  "num_attention_heads": 16,
  "num_hidden_layers": 28,
  "num_key_value_heads": 8,
  "num_modattn_kvheads": 8,
  "pad_token_id": 151643,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 10000.0,
  "sliding_window": null,
  "sssm_hidden_dim": 16,
  "sssm_low_rank": 8,
  "tie_word_embeddings": false,
  "transformers_version": "4.57.3",
  "use_S0": true,
  "use_cache": false,
  "use_phi": true,
  "use_sliding_window": false,
  "use_sssd": true,
  "use_val_rope": true,
  "vocab_size": 152000
}