Text Generation
Transformers
Safetensors
English
Chinese
Russian
yue2
music-generation
orbitquant
quantization
4-bit precision
custom-code
8-bit precision
Instructions to use WaveCut/YuE2-3B-OrbitQuant-W4A4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use WaveCut/YuE2-3B-OrbitQuant-W4A4 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="WaveCut/YuE2-3B-OrbitQuant-W4A4")# Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("WaveCut/YuE2-3B-OrbitQuant-W4A4", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use WaveCut/YuE2-3B-OrbitQuant-W4A4 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "WaveCut/YuE2-3B-OrbitQuant-W4A4" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "WaveCut/YuE2-3B-OrbitQuant-W4A4", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/WaveCut/YuE2-3B-OrbitQuant-W4A4
- SGLang
How to use WaveCut/YuE2-3B-OrbitQuant-W4A4 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 "WaveCut/YuE2-3B-OrbitQuant-W4A4" \ --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": "WaveCut/YuE2-3B-OrbitQuant-W4A4", "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 "WaveCut/YuE2-3B-OrbitQuant-W4A4" \ --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": "WaveCut/YuE2-3B-OrbitQuant-W4A4", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use WaveCut/YuE2-3B-OrbitQuant-W4A4 with Docker Model Runner:
docker model run hf.co/WaveCut/YuE2-3B-OrbitQuant-W4A4
File size: 6,034 Bytes
f0c91ec | 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 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 | from __future__ import annotations
import argparse
from pathlib import Path
def main() -> None:
parser = argparse.ArgumentParser(
description="Prepare kernel-builder output for a platform wheel build."
)
parser.add_argument("project", type=Path)
parser.add_argument("--version", required=True)
parser.add_argument(
"--torch-requirement",
default="torch>=2.11",
help=(
"torch dependency for the wheel metadata; non-stable-ABI variants "
'must pin the torch minor they were built against (e.g. "torch>=2.9,<2.10")'
),
)
args = parser.parse_args()
pyproject = args.project / "pyproject.toml"
text = pyproject.read_text(encoding="utf-8")
old_version = 'version = "0.1.0"'
if text.count(old_version) != 1:
raise RuntimeError("generated pyproject must contain one stub version")
text = text.replace(old_version, f'version = "{args.version}"', 1)
requires_python = 'requires-python = ">=3.9"'
if text.count(requires_python) != 1:
raise RuntimeError("generated pyproject must contain one Python requirement")
text = text.replace(
requires_python,
f'{requires_python}\ndependencies = ["{args.torch_requirement}"]',
1,
)
pyproject.write_text(text, encoding="utf-8")
cmake = args.project / "CMakeLists.txt"
cmake_text = cmake.read_text(encoding="utf-8")
for required in (False, True):
marker = " REQUIRED" if required else ""
development = (
f"find_package(Python3{marker} COMPONENTS Development "
"Development.SABIModule Interpreter)"
)
if cmake_text.count(development) != 1:
raise RuntimeError(
"generated CMake must contain one Python development lookup"
)
cmake_text = cmake_text.replace(
development,
f"find_package(Python3{marker} COMPONENTS Development.SABIModule Interpreter)",
1,
)
cmake.write_text(cmake_text, encoding="utf-8")
setup = args.project / "setup.py"
setup_text = setup.read_text(encoding="utf-8")
shutil_import = "from shutil import which, move\n"
if setup_text.count(shutil_import) != 1:
raise RuntimeError("generated setup must contain one shutil import")
setup_text = setup_text.replace(
shutil_import,
"from shutil import copy2, move, which\n",
1,
)
ninja_path = 'ninja_executable_path = Path(ninja.BIN_DIR) / "ninja"'
if setup_text.count(ninja_path) != 1:
raise RuntimeError("generated setup must contain one Ninja executable path")
setup_text = setup_text.replace(
ninja_path,
'ninja_executable_path = Path(ninja.BIN_DIR) / '
'("ninja.exe" if os.name == "nt" else "ninja")',
1,
)
for cache_tool in ("sccache", "ccache"):
availability = f'return which("{cache_tool}") is not None'
if setup_text.count(availability) != 1:
raise RuntimeError(
f"generated setup must contain one {cache_tool} availability check"
)
setup_text = setup_text.replace(
availability,
f'return os.name != "nt" and which("{cache_tool}") is not None',
1,
)
cmake_args_hook = (
' if "CMAKE_ARGS" in os.environ:\n'
' cmake_args += [item for item in os.environ["CMAKE_ARGS"].split(" ") '
"if item]\n"
)
if setup_text.count(cmake_args_hook) != 1:
raise RuntimeError("generated setup must contain one CMAKE_ARGS hook")
setup_text = setup_text.replace(
cmake_args_hook,
cmake_args_hook
+ ' cmake_make_program = os.environ.get("ORBITQUANT_CMAKE_MAKE_PROGRAM")\n'
+ " if cmake_make_program:\n"
+ ' cmake_args.append(f"-DCMAKE_MAKE_PROGRAM:FILEPATH={cmake_make_program}")\n',
1,
)
build_temp = " build_temp = Path(self.build_temp) / ext.name"
if setup_text.count(build_temp) != 1:
raise RuntimeError("generated setup must contain one extension build temp")
setup_text = setup_text.replace(
build_temp,
' build_temp_root = os.environ.get("ORBITQUANT_BUILD_TEMP", '
"self.build_temp)\n"
" build_temp = (Path(build_temp_root) / ext.name).resolve()",
1,
)
windows_multi_config = (
' if sys.platform == "win32":\n'
" # Move the dylib one folder up for discovery."
)
if setup_text.count(windows_multi_config) != 1:
raise RuntimeError("generated setup must contain one Windows output move")
setup_text = setup_text.replace(
windows_multi_config,
' if sys.platform == "win32" and (extdir / cfg).is_dir():\n'
" # Move the dylib one folder up for discovery.",
1,
)
build_call = (
" subprocess.run(\n"
' ["cmake", "--build", str(build_temp), *build_args], '
"cwd=build_temp, check=True\n"
" )\n"
)
if setup_text.count(build_call) != 1:
raise RuntimeError("generated setup must contain one wheel CMake build call")
setup_text = setup_text.replace(
build_call,
build_call
+ "\n"
+ ' package_name = ext.name.split(".", 1)[0]\n'
+ " generated_ops = (\n"
+ ' Path(ext.sourcedir) / "torch-ext" / package_name / "_ops.py"\n'
+ " )\n"
+ ' copy2(generated_ops, extdir / "_ops.py")\n',
1,
)
zip_safe = " zip_safe=False,\n"
if setup_text.count(zip_safe) != 1:
raise RuntimeError("generated setup must contain one zip-safe option")
setup_text = setup_text.replace(
zip_safe,
' options={"bdist_wheel": {"py_limited_api": "cp39"}},\n' + zip_safe,
1,
)
setup.write_text(setup_text, encoding="utf-8")
if __name__ == "__main__":
main()
|