Delta compression for LLM fine-tunes - lossless or LoRA-equivalent SVD compression

These details have been verified by PyPI

Project links

GitHub Statistics

Maintainers

gagansuie

These details have not been verified by PyPI

Project description

∴ Sparse

Delta Compression for Fine-tuned Models and Datasets

Compress your 14GB fine-tune to 1.4GB (lossless) or 50MB (LoRA-equivalent). Reconstruct in 4 seconds.

Verified: GPT-2 compression → reconstruction → identical inference output ✅

Quick Start • How It Works • CLI • Python API

What Sparse Does

Sparse compresses fine-tuned models and datasets as deltas from their base versions.

Compression Mode	Size (7B)	Quality	Use Case
Lossless	~1.4 GB	100%	Production, quality-critical
Lossy (SVD)	~50 MB	~95-99%	Sharing, size-critical
Dataset Delta	60-80% savings	100%	Derivative datasets

Key benefit: Works on models you've already trained - no LoRA required during training.

Works with: Full fine-tunes, RLHF, model merges, translated/augmented datasets

Quick Start

pip install sparse-llm

Compress a Fine-tune

# Lossless compression (~1.4GB for 7B model)
sparse compress meta-llama/Llama-2-7b-hf ./my-finetune -o ./my-delta

# OR: Lossy compression (~50MB, LoRA-equivalent quality)
sparse compress-lossy meta-llama/Llama-2-7b-hf ./my-finetune -o ./my-delta --rank 16

Reconstruct from Delta

# From lossless delta
sparse reconstruct meta-llama/Llama-2-7b-hf ./my-delta -o ./reconstructed-model

# From lossy delta
sparse reconstruct-lossy meta-llama/Llama-2-7b-hf ./my-delta -o ./reconstructed-model

Dataset Delta

# Compress derivative dataset
sparse dataset-compress squad squad_v2 -o ./squad_v2_delta

# Reconstruct
sparse dataset-reconstruct ./squad_v2_delta

How It Works

Fine-tuned Model (14GB)  -  Base Model (14GB)  =  Delta (1.4GB or 50MB)
                                                        ↓
                                              Reconstruct: Base + Delta

Lossless: Sparse + INT8 encoding → ~10% of original size, 100% quality
Lossy (SVD): Low-rank approximation → ~0.4% of original, ~95-99% quality

CLI Reference

# Lossless compression (100% quality)
sparse compress <base> <finetune> -o <output>
sparse reconstruct <base> <delta> [-o <output>]

# Lossy compression (~50MB, LoRA-equivalent quality)
sparse compress-lossy <base> <finetune> -o <output> [--rank 16]
sparse reconstruct-lossy <base> <delta> [-o <output>]

# Dataset commands
sparse dataset-compress <base> <derivative> -o <output>
sparse dataset-reconstruct <delta_dir>
sparse dataset-estimate <base> <derivative>

# Info
sparse info <path>

Python API

from core import compress_delta, reconstruct_from_delta
from core import compress_delta_svd_full, reconstruct_from_svd_delta

# Lossless compression
manifest = compress_delta(
    base_model_id="meta-llama/Llama-2-7b-hf",
    finetune_model_id="./my-finetune",
    output_path="./my-delta"
)
print(f"Compression: {manifest.compression_ratio:.1f}x")  # ~10x

# Extract LoRA (lossy, LoRA-equivalent)
manifest = compress_delta_svd_full(
    base_model_id="meta-llama/Llama-2-7b-hf",
    finetune_model_id="./my-finetune",
    output_path="./my-svd-delta",
    rank=16  # Like LoRA rank
)
print(f"Compression: {manifest.compression_ratio:.1f}x")  # ~280x

# Reconstruct (lossless)
model = reconstruct_from_delta("meta-llama/Llama-2-7b-hf", "./my-delta")

# Reconstruct from extracted LoRA
model = reconstruct_from_svd_delta("meta-llama/Llama-2-7b-hf", "./my-lora-delta")

Dataset API

from core import compress_dataset_delta, reconstruct_from_dataset_delta

# Compress
manifest = compress_dataset_delta("squad", "squad_v2", "./squad_v2_delta")
print(f"Savings: {manifest['size_stats']['savings_pct']:.1f}%")

# Reconstruct
dataset = reconstruct_from_dataset_delta("./squad_v2_delta")

Performance

All optimizations are automatic - no configuration needed:

Rust SIMD acceleration: 5-10x faster compression
Base model caching: ~20s saved per compression
Smart heuristics: 10-20% better compression ratios
GPU reconstruction: 2-3x faster on CUDA
Lazy loading: 50-70% memory reduction for 30B+ models

Typical speedup: ~60s → ~8-12s (5-8x faster)

📚 Advanced optimizations: See API_REFERENCE.md for MmapDeltaStorage, DifferentialCompressor, and other utilities.

Sparse vs LoRA

	LoRA/PEFT	Sparse
When applied	During training	After training
Works on existing models	❌	✅
Lossless option	❌	✅

Key insight: sparse compress-lossy gives you LoRA-sized files (~50MB) from models that weren't trained with LoRA.

Requirements

Python 3.9+
PyTorch 2.0+
transformers
Rust (included in wheel, no setup needed)

License

Apache 2.0 - See LICENSE for details.

Free for personal and commercial use.

Project details

These details have been verified by PyPI

Project links

GitHub Statistics

Maintainers

gagansuie

These details have not been verified by PyPI

Release history Release notifications | RSS feed

This version

0.0.5

Jan 5, 2026

0.0.4

Jan 4, 2026

0.0.3

Jan 4, 2026

0.0.2

Jan 3, 2026

0.0.1

Jan 3, 2026

Download files

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

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

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.

sparse_llm-0.0.5-cp39-abi3-win_amd64.whl (636.0 kB view details)

Uploaded Jan 5, 2026 CPython 3.9+Windows x86-64

sparse_llm-0.0.5-cp39-abi3-manylinux_2_34_x86_64.whl (888.6 kB view details)

Uploaded Jan 5, 2026 CPython 3.9+manylinux: glibc 2.34+ x86-64

sparse_llm-0.0.5-cp39-abi3-macosx_11_0_arm64.whl (680.6 kB view details)

Uploaded Jan 5, 2026 CPython 3.9+macOS 11.0+ ARM64

File details

Details for the file sparse_llm-0.0.5-cp39-abi3-win_amd64.whl.

File metadata

Download URL: sparse_llm-0.0.5-cp39-abi3-win_amd64.whl
Upload date: Jan 5, 2026
Size: 636.0 kB
Tags: CPython 3.9+, Windows x86-64
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for sparse_llm-0.0.5-cp39-abi3-win_amd64.whl
Algorithm	Hash digest
SHA256	`c7e1022abb38e84ebc1a524ae73da3b6dd58358b42d111cf2b655ac67c620d80`
MD5	`036252a42fa6b2df9542a64ed944de09`
BLAKE2b-256	`6494d8759fb0521010439efe9f63b5526b1c7105ad662d336565ac9ab411ee3f`

See more details on using hashes here.

Provenance

The following attestation bundles were made for sparse_llm-0.0.5-cp39-abi3-win_amd64.whl:

Publisher: build-artifacts.yml on gagansuie/sparse

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Statement:
- Statement type: https://in-toto.io/Statement/v1
- Predicate type: https://docs.pypi.org/attestations/publish/v1
- Subject name: sparse_llm-0.0.5-cp39-abi3-win_amd64.whl
- Subject digest: c7e1022abb38e84ebc1a524ae73da3b6dd58358b42d111cf2b655ac67c620d80
- Sigstore transparency entry: 791047063
- Sigstore integration time: Jan 5, 2026
Source repository:
- Permalink: gagansuie/sparse@8cbb068574893743730212770d795ae343a148fa
- Branch / Tag: refs/heads/main
- Owner: https://github.com/gagansuie
- Access: public
Publication detail:
- Token Issuer: https://token.actions.githubusercontent.com
- Runner Environment: github-hosted
- Publication workflow: build-artifacts.yml@8cbb068574893743730212770d795ae343a148fa
- Trigger Event: push

File details

Details for the file sparse_llm-0.0.5-cp39-abi3-manylinux_2_34_x86_64.whl.

File metadata

Download URL: sparse_llm-0.0.5-cp39-abi3-manylinux_2_34_x86_64.whl
Upload date: Jan 5, 2026
Size: 888.6 kB
Tags: CPython 3.9+, manylinux: glibc 2.34+ x86-64
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for sparse_llm-0.0.5-cp39-abi3-manylinux_2_34_x86_64.whl
Algorithm	Hash digest
SHA256	`a6a8214c55ba64241f7e0951d01aacb423cd2f3fdf75ce72fa37e59dd1b76ac7`
MD5	`8c7031c68bb9cd1dfde449cb2101d1c7`
BLAKE2b-256	`a1db4fbff9c90cbcf36e26441bd67d125f1ced734200c40de7b1077f7d96848f`

See more details on using hashes here.

Provenance

The following attestation bundles were made for sparse_llm-0.0.5-cp39-abi3-manylinux_2_34_x86_64.whl:

Publisher: build-artifacts.yml on gagansuie/sparse

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Statement:
- Statement type: https://in-toto.io/Statement/v1
- Predicate type: https://docs.pypi.org/attestations/publish/v1
- Subject name: sparse_llm-0.0.5-cp39-abi3-manylinux_2_34_x86_64.whl
- Subject digest: a6a8214c55ba64241f7e0951d01aacb423cd2f3fdf75ce72fa37e59dd1b76ac7
- Sigstore transparency entry: 791047053
- Sigstore integration time: Jan 5, 2026
Source repository:
- Permalink: gagansuie/sparse@8cbb068574893743730212770d795ae343a148fa
- Branch / Tag: refs/heads/main
- Owner: https://github.com/gagansuie
- Access: public
Publication detail:
- Token Issuer: https://token.actions.githubusercontent.com
- Runner Environment: github-hosted
- Publication workflow: build-artifacts.yml@8cbb068574893743730212770d795ae343a148fa
- Trigger Event: push

File details

Details for the file sparse_llm-0.0.5-cp39-abi3-macosx_11_0_arm64.whl.

File metadata

Download URL: sparse_llm-0.0.5-cp39-abi3-macosx_11_0_arm64.whl
Upload date: Jan 5, 2026
Size: 680.6 kB
Tags: CPython 3.9+, macOS 11.0+ ARM64
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for sparse_llm-0.0.5-cp39-abi3-macosx_11_0_arm64.whl
Algorithm	Hash digest
SHA256	`29d43f28b2838192812499b9df5604b62b10cb8518c970a39185289661f30ddd`
MD5	`580512a7c76f156e9811642f57b8c3c2`
BLAKE2b-256	`8daa86e9411985cf297b912e3b0afbb659a759d18432353052beaa2520de0e60`

See more details on using hashes here.

Provenance

The following attestation bundles were made for sparse_llm-0.0.5-cp39-abi3-macosx_11_0_arm64.whl:

Publisher: build-artifacts.yml on gagansuie/sparse

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Statement:
- Statement type: https://in-toto.io/Statement/v1
- Predicate type: https://docs.pypi.org/attestations/publish/v1
- Subject name: sparse_llm-0.0.5-cp39-abi3-macosx_11_0_arm64.whl
- Subject digest: 29d43f28b2838192812499b9df5604b62b10cb8518c970a39185289661f30ddd
- Sigstore transparency entry: 791047057
- Sigstore integration time: Jan 5, 2026
Source repository:
- Permalink: gagansuie/sparse@8cbb068574893743730212770d795ae343a148fa
- Branch / Tag: refs/heads/main
- Owner: https://github.com/gagansuie
- Access: public
Publication detail:
- Token Issuer: https://token.actions.githubusercontent.com
- Runner Environment: github-hosted
- Publication workflow: build-artifacts.yml@8cbb068574893743730212770d795ae343a148fa
- Trigger Event: push

sparse-llm 0.0.5

Navigation

Verified details

Project links

GitHub Statistics

Maintainers

Unverified details

Meta

Classifiers

Project description

∴ Sparse

What Sparse Does

Quick Start

Compress a Fine-tune

Reconstruct from Delta

Dataset Delta

How It Works

CLI Reference

Python API

Dataset API

Performance

Sparse vs LoRA

Requirements

License

Project details

Verified details

Project links

GitHub Statistics

Maintainers

Unverified details

Meta

Classifiers

Release history Release notifications | RSS feed

Download files

Source Distributions

Built Distributions

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance