# jina-ai/finetuner

:dart: Task-oriented embedding tuning for BERT, CLIP, etc.

Repository: https://github.com/jina-ai/finetuner
Canonical: https://ross.abutalabs.com/products/finetuner
Homepage: https://finetuner.jina.ai
Language: Python
License: Apache-2.0
License Family: permissive
Topics: fine-tuning, pretrained-models, few-shot-learning, negative-sampling, metric-learning, siamese-network, triplet-loss, transfer-learning, jina, neural-search, finetuning, similarity-learning, bert, openai-clip
Archived: true
Last push: 2024-03-11T08:05:13+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 1848, "days_push": 905, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1503, forks 66 (observed 2026-08-28T04:04:54.624647+00:00)

## What it is
Jina AI's Finetuner is a Python library for task-oriented fine-tuning of pretrained models like BERT and CLIP to produce better embeddings for neural search. It streamlines the fine-tuning workflow and can run training jobs on Jina AI Cloud GPU infrastructure.

## Use cases
- fine-tune BERT for semantic text search
- improve CLIP embeddings for text-to-image search
- train embeddings with few hundred labeled samples
- metric learning with triplet loss
- hard negative mining for similarity search
- fine-tune image similarity models

## When to choose
- you need domain-specific embeddings for search or recommendation
- you have limited labeled data and want few-shot fine-tuning
- you want cloud GPU training without managing infrastructure

## When to avoid
- you need fully local/offline training without Jina AI Cloud
- you need general-purpose LLM instruction fine-tuning rather than embedding tuning
- you require a very actively developed tool - releases have slowed

## Facets
- artifact type: library
- maturity: maintenance
- function: llm-training, machine-learning, deep-learning
- domain: machine-learning, deep-learning
- platform: python, cloud, windows
- tags: fine-tuning, embeddings, metric-learning, neural-search, few-shot-learning, bert, clip, search, linux, macos

## Member repositories
- jina-ai/finetuner (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:54.624647+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:32:49.493355+00:00, confidence not recorded.
  - readme: https://github.com/jina-ai/finetuner (fetched 2026-08-28T04:04:54.624647+00:00, sha 84b7d2f5d70b)
  - homepage: https://finetuner.jina.ai (fetched 2026-08-29T11:37:34.551251+00:00, sha 4c1e70afd3b4)
  - site_page: https://finetuner.jina.ai/get-started/installation (fetched 2026-08-29T11:37:34.553849+00:00, sha cfdf0fb95742)
  - registry_pypi: https://pypi.org/pypi/finetuner/json (fetched 2026-08-29T11:37:34.555371+00:00, sha df444df8d41d)
- Data as of 2026-08-30T08:39:29.467469+00:00.