pyrecall

Forgetting detection and skill rollback for fine-tuned LLMs.

Fine-tune on new data and your model quietly loses what it already knew — coding ability, reasoning, safety guardrails. pyrecall catches it before it ships.

Install

pip install pyrecall   # Python 3.11–3.14, CPU-only hardware supported

Quickstart

from pyrecall import Model

model = Model("meta-llama/Llama-3.2-1B")
model.snapshot("before")
model.learn("data.jsonl", epochs=3)

if not model.check().is_healthy:
    model.rollback(to="before")

pyrecall init --model meta-llama/Llama-3.2-1B
pyrecall snapshot before_v1
pyrecall learn train.jsonl --snapshot-after after_v1
pyrecall check --before before_v1 --after after_v1
# exit 0 → ship   exit 2 → pyrecall rollback before_v1

How it works

Benchmarks 180 prompts across 9 skill categories (reasoning, coding, safety, math, multilingual, and more) using log-likelihood scoring. After training, check diffs the scores and flags any category that drops past your threshold (default 10%). Only the LoRA adapter is stored per snapshot — a few hundred MB, not the full model.

Any causal LM on HuggingFace Hub is supported — Llama, Mistral, Phi, Gemma, Qwen, Falcon, GPT-2/Neo/J, and more. LoRA targets are auto-detected.

Docs

Full CLI reference, Python API, experiment tracker integrations (W&B, MLflow, Neptune), custom benchmarks, per-category thresholds, and more:

pyrecall.github.io/Pyrecall

Contributing

git clone https://github.com/Arths17/Pyrecall
pip install -e ".[dev]"
pytest

Open an issue before large changes. MIT — LICENSE.

Download files

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

Source Distribution

pyrecall-0.11.6.tar.gz (150.6 kB view details)

Uploaded Jun 17, 2026 Source

Built Distribution

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.

pyrecall-0.11.6-py3-none-any.whl (86.9 kB view details)

Uploaded Jun 17, 2026 Python 3

File details

Details for the file pyrecall-0.11.6.tar.gz.

File metadata

Download URL: pyrecall-0.11.6.tar.gz
Upload date: Jun 17, 2026
Size: 150.6 kB
Tags: Source
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for pyrecall-0.11.6.tar.gz
Algorithm	Hash digest
SHA256	`3766747d9a5c7562d27df409bb8788af50ea8d0937230ca1fa49c8f27090e4e4`
MD5	`f50e1bea3fe42575b13927db1f810134`
BLAKE2b-256	`f1dd0a34069dc89dd4d421db4e7e8e5d7c682836c01ccd33c3b893b489bf554b`

See more details on using hashes here.

Provenance

The following attestation bundles were made for pyrecall-0.11.6.tar.gz:

Publisher: publish.yml on Pyrecall/Pyrecall

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: pyrecall-0.11.6.tar.gz
- Subject digest: 3766747d9a5c7562d27df409bb8788af50ea8d0937230ca1fa49c8f27090e4e4
- Sigstore transparency entry: 1846007541
- Sigstore integration time: Jun 17, 2026
Source repository:
- Permalink: Pyrecall/Pyrecall@3731f4f892b75cd73f0fc1d5ff87d322c5b97fc3
- Branch / Tag: refs/tags/v0.11.6
- Owner: https://github.com/Pyrecall
- Access: public
Publication detail:
- Token Issuer: https://token.actions.githubusercontent.com
- Runner Environment: github-hosted
- Publication workflow: publish.yml@3731f4f892b75cd73f0fc1d5ff87d322c5b97fc3
- Trigger Event: push