Instructions to use aovolkov/Orpo-granite-4.0-micro with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use aovolkov/Orpo-granite-4.0-micro with Transformers:

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

pipe = pipeline("text-generation", model="aovolkov/Orpo-granite-4.0-micro")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("aovolkov/Orpo-granite-4.0-micro")
model = AutoModelForCausalLM.from_pretrained("aovolkov/Orpo-granite-4.0-micro")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use aovolkov/Orpo-granite-4.0-micro with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "aovolkov/Orpo-granite-4.0-micro"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "aovolkov/Orpo-granite-4.0-micro",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/aovolkov/Orpo-granite-4.0-micro

SGLang

How to use aovolkov/Orpo-granite-4.0-micro 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 "aovolkov/Orpo-granite-4.0-micro" \
    --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": "aovolkov/Orpo-granite-4.0-micro",
		"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 "aovolkov/Orpo-granite-4.0-micro" \
        --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": "aovolkov/Orpo-granite-4.0-micro",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use aovolkov/Orpo-granite-4.0-micro with Docker Model Runner:
```
docker model run hf.co/aovolkov/Orpo-granite-4.0-micro
```

Orpo-granite-4.0-micro

File size: 2,099 Bytes

d18d41f

{
  "architectures": [
    "GraniteMoeHybridForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "attention_multiplier": 0.015625,
  "bos_token_id": 100257,
  "dtype": "float16",
  "embedding_multiplier": 12,
  "eos_token_id": 100257,
  "hidden_act": "silu",
  "hidden_size": 2560,
  "initializer_range": 0.1,
  "intermediate_size": 8192,
  "layer_types": [
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention",
    "attention"
  ],
  "logits_scaling": 10,
  "mamba_chunk_size": 256,
  "mamba_conv_bias": true,
  "mamba_d_conv": 4,
  "mamba_d_head": 40,
  "mamba_d_state": 256,
  "mamba_expand": 2,
  "mamba_n_groups": 1,
  "mamba_n_heads": 128,
  "mamba_proj_bias": false,
  "max_position_embeddings": 131072,
  "model_type": "granitemoehybrid",
  "normalization_function": "rmsnorm",
  "num_attention_heads": 40,
  "num_experts_per_tok": 0,
  "num_hidden_layers": 40,
  "num_key_value_heads": 8,
  "num_local_experts": 0,
  "output_router_logits": false,
  "pad_token_id": 100256,
  "position_embedding_type": "rope",
  "residual_multiplier": 0.22,
  "rms_norm_eps": 1e-05,
  "rope_parameters": {
    "rope_theta": 10000000,
    "rope_type": "default"
  },
  "router_aux_loss_coef": 0.01,
  "shared_intermediate_size": 8192,
  "tie_word_embeddings": true,
  "time_step_limit": [
    0.0,
    {
      "__float__": "Infinity"
    }
  ],
  "time_step_max": 0.1,
  "time_step_min": 0.001,
  "transformers_version": "5.4.0",
  "use_cache": true,
  "vocab_size": 100352
}