limbed

ColBERT embedding and MaxSim scoring without PyTorch. Uses a native C++ extension (ONNX Runtime + tokenizers-cpp) so you don't need to pull in 2 GB of deep learning dependencies just to encode some text.

Install

pip install limbed-py

Only runtime deps are numpy and huggingface-hub.

Usage

from limbed import LateEmbedder, compute_maxsim

model = LateEmbedder()  # downloads thlurte/mxbai-edge-colbert-v0-17m-onnx

q = model.encode_queries("What is late interaction?")
d = model.encode_docs("ColBERT computes token-level similarity.")

score = compute_maxsim(q[0], d[0])
print(score)

You can also point it at a local directory with model.onnx and tokenizer.json:

model = LateEmbedder("./my-model/")

Models

Alias	Repo	Size	Dim	Notes
`mxbai-edge-colbert-v0-17m`	`thlurte/mxbai-edge-colbert-v0-17m-onnx`	66 MB	48	Default
`mxbai-edge-colbert-v0-32m`	`thlurte/mxbai-edge-colbert-v0-32m-onnx`	124 MB	64
`colbertv2.0`	`thlurte/colbertv2.0-onnx`	438 MB	128	Standard ColBERTv2.0 BERT-based model
`answerai-colbert-small-v1`	`thlurte/answerai-colbert-small-v1-onnx`	135 MB	96	Lightweight, high-performance model
`jina-colbert-v2`	`thlurte/jina-colbert-v2-onnx`	2.23 GB	128	XLM-RoBERTa multilingual model
`lateon`	`thlurte/lateon-onnx`	580 MB	128	Case-sensitive: use `do_lower_case=False`

Any ColBERT ONNX model should work if you put model.onnx and tokenizer.json in a folder and pass the path.

Benchmarks

The following benchmark was run on CPU using 20 queries (max length 32) and 20 documents (max length 256), comparing limbed against fastembed execution:

Performance (Throughput & Speedup)

Model	Operation	`limbed` Throughput	`fastembed` Throughput	Speedup (Wall-clock)
ColBERTv2.0	Queries	71.3 QPS	31.6 QPS	2.25x
ColBERTv2.0	Documents	93.8 DPS	66.1 DPS	1.42x
Jina ColBERT v2	Queries	6.2 QPS	5.0 QPS	1.25x
Jina ColBERT v2	Documents	10.1 DPS	5.2 DPS	1.94x

How it works

Tokenization and inference run in C++ via a nanobind extension
GIL is released during encode and MaxSim calls, so you can run multiple threads
Punctuation tokens are masked out of document embeddings (standard ColBERT behavior)
Embeddings are L2-normalized by default
CPU only for now

Docker & Alpine Linux Compatibility

Because the underlying precompiled ONNX Runtime library is linked against glibc, this package will not run out-of-the-box on Alpine Linux images (e.g., python:3.10-alpine).

If deploying via Docker, it is highly recommended to use a Debian-based slim image:

FROM python:3.10-slim

If you must use Alpine, you will need to install the compatibility layer: apk add --no-cache gcompat.

Supported Platforms & Architectures

Precompiled wheels are published to PyPI for the following environments:

Operating System	Architecture	Python Versions	Notes
Linux	`x86_64`, `aarch64`	3.9, 3.10, 3.11, 3.12, 3.13, 3.14	Built on `manylinux_2_28` (glibc-based)
macOS	`arm64` (Apple Silicon)	3.9, 3.10, 3.11, 3.12, 3.13, 3.14	SDK/deployment target macOS 13.3+
Windows	`AMD64` (x86_64)	3.9, 3.10, 3.11, 3.12, 3.13, 3.14

[!NOTE] Other platforms (such as Intel-based macOS or ARM-based Windows) will fall back to compilation from the source distribution (sdist). This requires a local C++ compiler (supporting C++17) and CMake.

License

MIT. See LICENSE.

License
- OSI Approved :: MIT License
Operating System
- OS Independent
Programming Language
- Python :: 3

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

limbed_py-0.1.5.tar.gz (20.4 kB view details)

Uploaded Aug 16, 2026 Source

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

limbed_py-0.1.5-cp313-cp313-win_amd64.whl (6.7 MB view details)

Uploaded Aug 16, 2026 CPython 3.13Windows x86-64

limbed_py-0.1.5-cp313-cp313-manylinux_2_28_x86_64.whl (21.3 MB view details)

Uploaded Aug 16, 2026 CPython 3.13manylinux: glibc 2.28+ x86-64

limbed_py-0.1.5-cp313-cp313-manylinux_2_28_aarch64.whl (18.8 MB view details)

Uploaded Aug 16, 2026 CPython 3.13manylinux: glibc 2.28+ ARM64

limbed_py-0.1.5-cp313-cp313-macosx_13_0_arm64.whl (16.6 MB view details)

Uploaded Aug 16, 2026 CPython 3.13macOS 13.0+ ARM64

limbed_py-0.1.5-cp312-cp312-win_amd64.whl (6.7 MB view details)

Uploaded Aug 16, 2026 CPython 3.12Windows x86-64

limbed_py-0.1.5-cp312-cp312-manylinux_2_28_x86_64.whl (21.3 MB view details)

Uploaded Aug 16, 2026 CPython 3.12manylinux: glibc 2.28+ x86-64

limbed_py-0.1.5-cp312-cp312-manylinux_2_28_aarch64.whl (18.8 MB view details)

Uploaded Aug 16, 2026 CPython 3.12manylinux: glibc 2.28+ ARM64

limbed_py-0.1.5-cp312-cp312-macosx_13_0_arm64.whl (16.6 MB view details)

Uploaded Aug 16, 2026 CPython 3.12macOS 13.0+ ARM64

limbed_py-0.1.5-cp311-cp311-win_amd64.whl (6.7 MB view details)

Uploaded Aug 16, 2026 CPython 3.11Windows x86-64

limbed_py-0.1.5-cp311-cp311-manylinux_2_28_x86_64.whl (21.3 MB view details)

Uploaded Aug 16, 2026 CPython 3.11manylinux: glibc 2.28+ x86-64

limbed_py-0.1.5-cp311-cp311-manylinux_2_28_aarch64.whl (18.8 MB view details)

Uploaded Aug 16, 2026 CPython 3.11manylinux: glibc 2.28+ ARM64

limbed_py-0.1.5-cp311-cp311-macosx_13_0_arm64.whl (16.6 MB view details)

Uploaded Aug 16, 2026 CPython 3.11macOS 13.0+ ARM64

limbed_py-0.1.5-cp310-cp310-win_amd64.whl (6.7 MB view details)

Uploaded Aug 16, 2026 CPython 3.10Windows x86-64

limbed_py-0.1.5-cp310-cp310-manylinux_2_28_x86_64.whl (21.3 MB view details)

Uploaded Aug 16, 2026 CPython 3.10manylinux: glibc 2.28+ x86-64

limbed_py-0.1.5-cp310-cp310-manylinux_2_28_aarch64.whl (18.8 MB view details)

Uploaded Aug 16, 2026 CPython 3.10manylinux: glibc 2.28+ ARM64

limbed_py-0.1.5-cp310-cp310-macosx_13_0_arm64.whl (16.6 MB view details)

Uploaded Aug 16, 2026 CPython 3.10macOS 13.0+ ARM64

limbed_py-0.1.5-cp39-cp39-win_amd64.whl (6.7 MB view details)

Uploaded Aug 16, 2026 CPython 3.9Windows x86-64

limbed_py-0.1.5-cp39-cp39-manylinux_2_28_x86_64.whl (21.3 MB view details)

Uploaded Aug 16, 2026 CPython 3.9manylinux: glibc 2.28+ x86-64

limbed_py-0.1.5-cp39-cp39-manylinux_2_28_aarch64.whl (18.8 MB view details)

Uploaded Aug 16, 2026 CPython 3.9manylinux: glibc 2.28+ ARM64

limbed_py-0.1.5-cp39-cp39-macosx_13_0_arm64.whl (16.6 MB view details)

Uploaded Aug 16, 2026 CPython 3.9macOS 13.0+ ARM64

File details

Details for the file limbed_py-0.1.5.tar.gz.

File metadata

Download URL: limbed_py-0.1.5.tar.gz
Upload date: Aug 16, 2026
Size: 20.4 kB
Tags: Source
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for limbed_py-0.1.5.tar.gz
Algorithm	Hash digest
SHA256	`3470f66f08e9c25df7a6225531050948f47e3dc348c7fe18fde04751d5ea9874`
MD5	`ba7852e8ceb80309f00e5042b86d6f4d`
BLAKE2b-256	`42b216ad77932b2cdc4f3a4d1e3b25b6a2f0c2d741f7c0f261f99c0865070c64`

Algorithm	Hash digest
SHA256	`4e21b9815dad85741b21acd1cffcbd932609aac35bd5312e08bc58ead1839531`
MD5	`69471628b31f0a6a9ff8c29048e94251`
BLAKE2b-256	`e874b01d697db74ba36be6e6a6f572619a6368653cf806acfdfe9a2b695ca3ea`

Algorithm	Hash digest
SHA256	`508a8658fd87118a1ba4aeeb40a9d22c08172bf993083f778080fb1c3367601f`
MD5	`d58127ae246fc0dbe87faed73bf609d4`
BLAKE2b-256	`1c78a3db7ffbfe490ce497724d45960914c22cd74db5069514e5c3aacc0c0ab1`

Algorithm	Hash digest
SHA256	`e8d93fe44b4576a7edc5ca39e7116c91ee798182e7796bfbf3172181066dcc70`
MD5	`26487368b6d788589ab82bb349487635`
BLAKE2b-256	`8a7dc63dae4fded72a6faf30380eb75f566853ef16a20d1e1b52feb3061c1a90`

Algorithm	Hash digest
SHA256	`41547bb895c941c4bb3571bb2c444877335f57265ad85e4fabbce7c4880f8369`
MD5	`2208495faeaffe84afa1d15a653356ea`
BLAKE2b-256	`27e2e20dfa2d109b186d9b1a752503a6f14cf9f517e4564480a729bc060db52b`

Algorithm	Hash digest
SHA256	`0a9c4822bd92a508e168b59cc145a4786c872a56d451be5d756c9858c0dc9550`
MD5	`a838fb003dad1ff2ae9a16b64c562535`
BLAKE2b-256	`85bc9f62b34b21abc38000f0d5c411e6dfc8505c64793aadbb73f6aedb5b1e2b`

Algorithm	Hash digest
SHA256	`b6a6d5067a9c673dc9c4bc3e53d38e6c5fdd7111b98b1b21383f74b886791650`
MD5	`a7a14e4cfb344f930f43e94af0fec076`
BLAKE2b-256	`1c2bf4e7875d89c08418b0cd7cec41e3bbf5288b9ea62e56713f20b240396d2f`

Algorithm	Hash digest
SHA256	`7528bcc4b4229f0fd4c727100649e9ff6d125cd6876715ac74592b38f826f823`
MD5	`f4e1f7722b9bb70ee3a2e694c0ebed5f`
BLAKE2b-256	`803a91bcc8b9fe8b0b4138f0a97a3454d0582d9aaf0067271d59387f4607a9fc`

Algorithm	Hash digest
SHA256	`7eba5843eb175ad27f86024661c3ace2a136f87c91e51142447caec7402cb4b3`
MD5	`1566e47bba9e72839000901f007e953f`
BLAKE2b-256	`a7e78afa29e5aa65f92b135d68bad8a9ac45759a24779167b34da7dff568fc3d`

Algorithm	Hash digest
SHA256	`29d64021446cc2f4a9caffc0d5b01e5b821e2fa0118b826c01d1959d0dac9f46`
MD5	`0c267463efa65b870bad9b456713f95b`
BLAKE2b-256	`031bbf03c02936ed81be48bedb4b96b583450a6457bc62db043439f0badb849e`

Algorithm	Hash digest
SHA256	`ee9f5412b3ce96b8ee0e652b91f400d042086e2a7f63f5e6acaca9ea70e572b1`
MD5	`34acbb4d3990c42def59be73aa3ffd95`
BLAKE2b-256	`322887fc4b8817be0ab046c8ccf3631466f919a1ae0d0bc140353db800f1c2a3`

Algorithm	Hash digest
SHA256	`409c251bb9f40e43ed74725f09b4b1e7aede410befcd641de95fe57a6eb2bdf2`
MD5	`a996dd2d9d05cd408c2ee04c1b89120c`
BLAKE2b-256	`742038ca891c95cc8c335250d8932c6231306b456f03759c7e69483b7d751fd3`

Algorithm	Hash digest
SHA256	`5d73e8fa5ec0ddffaaf644560463f7fe53ecc1dc10dd57e6a6deaf6b42b159c8`
MD5	`dcdeeeee28bf65dda6852ebf3377cd72`
BLAKE2b-256	`d941a5f70da18c95b3542347128abea3626838fa3098399e0baf3cc5d63bfac4`

Algorithm	Hash digest
SHA256	`5faa2869923ecc8a41143b504ddd5b53ad6d540d5cbf2089c1a8717d8eb736eb`
MD5	`1fdc3972a1aac7a9c3a5b0cc16695105`
BLAKE2b-256	`2318893fd19f3055a02e36b8eb5e36d9e6ba303a43d4b2c45623f25f917b511b`

Algorithm	Hash digest
SHA256	`db5672277da169a0506564b0be658cc52b771553f8484b162c88d3bcda3abfb3`
MD5	`aaad298ee1e2a4d8e3ea7c5bdc11341b`
BLAKE2b-256	`0bc38c3029c214cca9a2001632507fd1d4128cba4adff5fae97490203fbb7125`

Algorithm	Hash digest
SHA256	`b1cadaca1084f77301817b43047e53aa4cb6ea20ed9c603a2cd35d6625d65358`
MD5	`4f196527c8b7f3aa4182424bd15ac050`
BLAKE2b-256	`a165f0d5dc89510c7e304f272497453fc01be8101fd9ef4e3663b664475db651`

Algorithm	Hash digest
SHA256	`00fe95dca521d946bf9dbd7d74ee4518a6b53852a80ca54e96bd2f9b6a93cf98`
MD5	`cff584a6898fe103c36f554cdc68420e`
BLAKE2b-256	`7c959685922479ffe806ed67f38f96d58747148363f739e7f8414d4fe6e1603a`

Algorithm	Hash digest
SHA256	`271dd007928c391cf1104d0589d040749366bc30645f7f1a2469ebf357dcc752`
MD5	`52c11197c60c5ae6e6afc5df1523103a`
BLAKE2b-256	`8ca6eea4ee74d45387ab5146abce7e8f88f900451b5a0b30f3695c2595e5727e`

Algorithm	Hash digest
SHA256	`a99ddf2d50a9de97ba7ffa2afa4e71e5a9052a42596b1745caf60fc5f8476a9f`
MD5	`72b7a04c55557b3b28dd64e7418ef021`
BLAKE2b-256	`ca14a17d90fdaa02ae32a741037a9b5223b60617f6f5d31e9c5e2dc0cf95bb08`

Algorithm	Hash digest
SHA256	`5327a0308bdbb6e521fe875952bb35d2e7e39c6b86b0b7ef94f836af7f0bb3fc`
MD5	`f8e8de815327f9e6d7002d394ab1b05c`
BLAKE2b-256	`dfc71789f7c886ad26523e406c981faeebbd9139c6dce2bb101c049f5aa86bb8`

Algorithm	Hash digest
SHA256	`c4de5b267c8ec2e005ae5d4d28761122dae22345eb4f0272300015f8729f8bf0`
MD5	`9d707536be62745f51956cfc16578303`
BLAKE2b-256	`ea43d3164842d3096263eee501d21fadfdd30d6e4c09b19dc68fbe2bca4b8699`

limbed-py 0.1.5

limbed

Install

Usage

Models

Benchmarks

Performance (Throughput & Speedup)

How it works

Docker & Alpine Linux Compatibility

Supported Platforms & Architectures

License

Key dates

1 maintainer

License

Requires

Tags

Classifiers

Download files

Source Distribution

Built Distributions

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance