mirror of
https://github.com/huggingface/text-generation-inference.git
synced 2025-04-24 16:32:12 +00:00
571ac9b507
* Use Hub kernels for Marlin and cutlass quantization kernels * Use hub kernels for MoE/GPTQ-Marlin MoE * Use attention kernels from the Hub * Cache the kernels in the Docker image * Update moe kernels * Support loading local kernels for development * Support latest moe kernels * Update to moe 0.1.1 * CI: download locked kernels for server tests * Fixup some imports * CI: activate venv * Fix unused imports * Nix: add attention/moe/quantization kernels * Update hf-kernels to 0.1.5 * Update kernels * Update tgi-nix flake for hf-kernels * Fix EOF * Take `load_kernel` out of a frequently-called function * Hoist another case of kernel loading out of a somewhat hot function * marlin-kernels -> quantization * attention -> paged-attention * EOF fix * Update hf-kernels, fixup Docker * ipex fix * Remove outdated TODO
89 lines
2.0 KiB
TOML
89 lines
2.0 KiB
TOML
[project]
|
|
name = "text-generation-server"
|
|
version = "2.0.5-dev0"
|
|
description = "Text Generation Inference Python gRPC Server"
|
|
readme = "README.md"
|
|
requires-python = ">=3.9"
|
|
authors = [
|
|
{name = "Olivier Dehaene", email = "olivier@huggingface.co"},
|
|
{name = "Nicolas Patry", email = "nicolas@huggingface.co"},
|
|
]
|
|
dependencies = [
|
|
"einops>=0.8.0",
|
|
"grpc-interceptor>=0.15.4",
|
|
"grpcio>=1.67.0",
|
|
"grpcio-reflection>=1.67.0",
|
|
"grpcio-status>=1.67.0",
|
|
"hf-kernels>=0.1.5",
|
|
"hf-transfer>=0.1.8",
|
|
"loguru>=0.7.3",
|
|
"numpy>=1.26,<3",
|
|
"opentelemetry-api>=1.27.0",
|
|
"opentelemetry-exporter-otlp>=1.27.0",
|
|
"opentelemetry-instrumentation-grpc>=0.50b0",
|
|
"pillow>=11.1.0",
|
|
"prometheus-client>=0.21.0",
|
|
"protobuf>=5.28.3",
|
|
"py-cpuinfo>=9.0.0",
|
|
"rich>=13.8.1",
|
|
"safetensors>=0.4.5",
|
|
"scipy>=1.13.1",
|
|
"sentencepiece>=0.2.0",
|
|
"tokenizers>=0.20.3",
|
|
"typer>=0.15.1",
|
|
"transformers>=4.48.0"
|
|
]
|
|
|
|
[build-system]
|
|
requires = ["hf-kernels>=0.1.2", "setuptools"]
|
|
build-backend = "setuptools.build_meta"
|
|
|
|
[tool.kernels.dependencies]
|
|
"kernels-community/paged-attention" = ">=0.0.2"
|
|
"kernels-community/moe" = ">=0.1.1"
|
|
"kernels-community/quantization" = ">=0.0.3"
|
|
|
|
[project.scripts]
|
|
text-generation-server = "text_generation_server.cli:app"
|
|
|
|
[project.optional-dependencies]
|
|
accelerate = [
|
|
"accelerate>=1.2.1,<2",
|
|
]
|
|
bnb = [
|
|
"bitsandbytes>=0.45.0",
|
|
]
|
|
compressed-tensors = [
|
|
"compressed-tensors>=0.9.0",
|
|
]
|
|
peft = [
|
|
"peft>=0.14.0",
|
|
]
|
|
outlines = [
|
|
"outlines>=0.1.13",
|
|
]
|
|
dev = [
|
|
"grpcio-tools>=1.51.1,<2.0",
|
|
"pytest>=7.3.0,<8"
|
|
]
|
|
quantize = [
|
|
"texttable>=1.6.7,<2",
|
|
"datasets>=2.21,<3",
|
|
]
|
|
gen = [
|
|
"grpcio-tools>=1.69.0",
|
|
"mypy-protobuf>=3.6.0",
|
|
]
|
|
|
|
[tool.pytest.ini_options]
|
|
markers = ["private: marks tests as requiring an admin hf token (deselect with '-m \"not private\"')"]
|
|
|
|
[tool.isort]
|
|
profile = "black"
|
|
|
|
[tool.uv]
|
|
package = true
|
|
|
|
[tool.setuptools.packages.find]
|
|
include = ["text_generation_server*"]
|