Text Generation
Transformers
Safetensors
taonet
trust-remote-code
sentencepiece
custom-architecture
custom_code
Instructions to use TaoTern/TaoNet-mini-A2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use TaoTern/TaoNet-mini-A2 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="TaoTern/TaoNet-mini-A2", trust_remote_code=True)# Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("TaoTern/TaoNet-mini-A2", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use TaoTern/TaoNet-mini-A2 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "TaoTern/TaoNet-mini-A2" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "TaoTern/TaoNet-mini-A2", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/TaoTern/TaoNet-mini-A2
- SGLang
How to use TaoTern/TaoNet-mini-A2 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 "TaoTern/TaoNet-mini-A2" \ --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": "TaoTern/TaoNet-mini-A2", "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 "TaoTern/TaoNet-mini-A2" \ --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": "TaoTern/TaoNet-mini-A2", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use TaoTern/TaoNet-mini-A2 with Docker Model Runner:
docker model run hf.co/TaoTern/TaoNet-mini-A2
File size: 2,411 Bytes
fd448dd | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 | """Model architecture registry and factory."""
from typing import Dict, Type, Optional, Any
import torch
from taoTrain.core import BaseModel
from taoTrain.config import ModelConfig
# Global registry for model architectures
_ARCHITECTURE_REGISTRY: Dict[str, Type[BaseModel]] = {}
def register_architecture(name: str):
"""Decorator to register a custom model architecture."""
def decorator(cls: Type[BaseModel]):
if name in _ARCHITECTURE_REGISTRY:
raise ValueError(f"Architecture '{name}' is already registered")
_ARCHITECTURE_REGISTRY[name] = cls
return cls
return decorator
def get_registered_architectures() -> Dict[str, Type[BaseModel]]:
"""Get all registered architectures."""
return _ARCHITECTURE_REGISTRY.copy()
def get_model(
config: Any,
device: Optional[torch.device] = None,
) -> BaseModel:
"""
Create a model instance from config.
Args:
config: ModelConfig instance
device: Device to create model on (defaults to CPU)
Returns:
Model instance
"""
if device is None:
device = torch.device('cpu')
# Handle both raw model configs and full training configs used by wrappers.
config_with_arch = config
if not hasattr(config_with_arch, "architecture_type") and hasattr(config, "model"):
config_with_arch = config.model
# Handle both enum and string values
arch_type = config_with_arch.architecture_type
if isinstance(arch_type, str):
arch_name = arch_type
else:
arch_name = arch_type.value
if arch_name not in _ARCHITECTURE_REGISTRY:
raise ValueError(
f"Unknown architecture: {arch_name}. "
f"Available: {list(_ARCHITECTURE_REGISTRY.keys())}"
)
model_class = _ARCHITECTURE_REGISTRY[arch_name]
model = model_class(config).to(device)
return model
def register_builtin_architectures():
"""Register all built-in architectures."""
# Import here to register (avoid circular imports)
from . import transformer # noqa: F401
from . import taonet # noqa: F401
from . import gamma_net # noqa: F401
from . import multimodal_wrapper # noqa: F401
# Auto-register built-in architectures when module is imported
register_builtin_architectures()
|