Instructions to use Refract-Labs/Orion-Flagship-Mini-Base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Refract-Labs/Orion-Flagship-Mini-Base with Transformers:

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

pipe = pipeline("text-generation", model="Refract-Labs/Orion-Flagship-Mini-Base", trust_remote_code=True)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("Refract-Labs/Orion-Flagship-Mini-Base", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Refract-Labs/Orion-Flagship-Mini-Base with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Refract-Labs/Orion-Flagship-Mini-Base"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Refract-Labs/Orion-Flagship-Mini-Base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Refract-Labs/Orion-Flagship-Mini-Base

SGLang

How to use Refract-Labs/Orion-Flagship-Mini-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 "Refract-Labs/Orion-Flagship-Mini-Base" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Refract-Labs/Orion-Flagship-Mini-Base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "Refract-Labs/Orion-Flagship-Mini-Base" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Refract-Labs/Orion-Flagship-Mini-Base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Refract-Labs/Orion-Flagship-Mini-Base with Docker Model Runner:
```
docker model run hf.co/Refract-Labs/Orion-Flagship-Mini-Base
```

Orion-Flagship-Mini-Base

File size: 1,146 Bytes

0975e0c

{
  "architecture": "orion_flagship_mini_t2",
  "architectures": [
    "OrionT2ForCausalLM"
  ],
  "auto_map": {
    "AutoConfig": "configuration_orion.OrionT2Config",
    "AutoModelForCausalLM": "modeling_orion.OrionT2ForCausalLM"
  },
  "bos_token_id": 1,
  "d_model": 768,
  "deliberation_cycles": 1,
  "deliberation_start": 9,
  "eos_token_id": 2,
  "expert_hidden": 1024,
  "implementation_version": 2,
  "model_name": "Orion Flagship",
  "model_type": "orion_t2",
  "n_experts": 16,
  "n_heads": 12,
  "n_kv_heads": 4,
  "n_layers": 12,
  "n_procedure_banks": 3,
  "procedure_bank_span": 4,
  "rms_norm_eps": 1e-06,
  "rope_theta": 10000.0,
  "router_aux_coef": 0.005,
  "router_z_coef": 0.0005,
  "sequence_length": 1024,
  "shared_expert": true,
  "tie_word_embeddings": false,
  "tokenizer_name": "mistralai/Mistral-7B-v0.3",
  "top_k": 2,
  "transformers_compatibility": "custom_code",
  "use_cache": false,
  "vault_key_parts": 256,
  "vault_read_layers": [
    3,
    7,
    11
  ],
  "vault_slots": 65536,
  "vault_top_component": 12,
  "vault_top_k": 4,
  "vocab_size": 32768,
  "working_state_layers": [
    3,
    7,
    11
  ]
}