modelexpress-client 0.5.1

Client library for Model Express gRPC server
Documentation 
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

# SPDX-FileCopyrightText: Copyright (c) 2025-2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: Apache-2.0

[build-system]
requires = ["setuptools>=61.0", "wheel"]
build-backend = "setuptools.build_meta"

[project]
name = "modelexpress"
version = "0.5.1"
description = "Python client for ModelExpress P2P GPU transfer service"
readme = "README.md"
license = "Apache-2.0"
requires-python = ">=3.10"
authors = [{ name = "NVIDIA Corporation", email = "sw-dl-dynamo@nvidia.com" }]
keywords = ["llm", "gpu", "transfer", "rdma", "nixl", "vllm"]
classifiers = [
    "Development Status :: 4 - Beta",
    "Intended Audience :: Developers",
    "Programming Language :: Python :: 3",
    "Programming Language :: Python :: 3.10",
    "Programming Language :: Python :: 3.11",
    "Programming Language :: Python :: 3.12",
    "Topic :: Scientific/Engineering :: Artificial Intelligence",
]
dependencies = [
    "grpcio>=1.66.2",
    "google-crc32c>=1.5.0",
    "huggingface_hub>=0.20.0",
    "numpy>=1.24.0",
    "protobuf>=5.27.2,<7.0.0",
    "pydantic>=2.0.0",
    "torch>=2.6.0",
    "runai-model-streamer[s3,gcs,azure]; sys_platform == 'linux'",
]

[project.optional-dependencies]
dev = [
    "opentelemetry-api>=1.41.1",
    "opentelemetry-sdk>=1.41.1",
    "pytest>=7.0.0",
    "pytest-asyncio>=0.21.0",
]
codegen = [
    # Keep generated Python stubs compatible with both protobuf 5 and 6.
    "grpcio-tools==1.66.2",
]
otel = [
    "opentelemetry-api>=1.20.0",
]
# Opt-in Prometheus metrics collector (MX_METRICS_ENABLED=1).
# Lazily imported; not required for normal operation.
metrics = [
    "prometheus-client>=0.20.0",
]
# Required for the VmmArena allocator (CudaVmmBackend uses cuMem* APIs).
# The pure-Python VmmArena tests don't need it; CudaVmmBackend does.
vmm = [
    "cuda-python>=12.0",
]

[project.urls]
Homepage = "https://github.com/ai-dynamo/modelexpress"
Repository = "https://github.com/ai-dynamo/modelexpress.git"
Documentation = "https://github.com/ai-dynamo/modelexpress/tree/main/docs"

[project.entry-points."vllm.general_plugins"]
modelexpress = "modelexpress:register_modelexpress_loaders"

[tool.setuptools.packages.find]
where = ["."]
include = ["modelexpress*"]

[tool.setuptools.package-data]
modelexpress = ["py.typed"]