finetrainers Releases
55 releases of huggingface/finetrainers
- v5.48.0
Adds an optional OpenAPI spec route with gated access, fixes multiple bugs in upload, content-manager, and database components, and applies security patches for uuid and qs DoS advisories.
Jun 10, 2026
- v4.26.2
This is the final Strapi 4 release, marking it as End-Of-Life. It includes a critical security fix for relational filtering data exposure, Node.js 22 support, and dependency updates.
Jun 9, 2026
- v1.0.39Release v1.0.39
Adds credential folders, speech-to-text capabilities, and an LLM-backed TurnIntent classifier. Also includes security fixes for langchain integration and tiered LLM output prompts.
Jun 8, 2026
- v9.10.0
Focuses on training capabilities by adding LiteRT vector support and Knowledge Distillation training. It also includes a new URLRetrieve pipeline and reduces logging noise.
Jun 4, 2026
- v5.47.1
This release focuses on bug fixes, security patches, and performance optimizations. It addresses a ReDoS vulnerability, improves Danish localization, and refactors boot processes to load dependencies lazily.
Jun 3, 2026
- v5.47.0
Introduces a beta MCP server and a `publicationFilter` parameter for REST and document services. It includes fixes for nested relations, content manager access, and GraphQL inheritance issues.
May 28, 2026
- v5.46.1
A patch release focusing on bug fixes, security updates, and dependency bumps for the Strapi CMS platform, including improvements to the admin interface and database handling.
May 20, 2026
- v1.121.2Hyperswitch v1.121.2
A security-focused patch release for Hyperswitch that addresses a critical SQL injection vulnerability in analytics endpoints.
May 19, 2026
- v1.123.1Hyperswitch v1.123.1
This release introduces profile-based payment blocking, adds new wallet support via ACI connector, and enhances existing connectors like Stripe and Cybersource, alongside security fixes for SQL injection.
May 19, 2026
- v9.9.0
A major feature release for FineTainers introducing zero-dependency installation, support for Transformers v5, and the LiteRT-LM large language model.
May 12, 2026
- v9.8.0
A maintenance release primarily focused on security and compatibility, fixing issues with LiteLLM, Model2Vec, and protocol handling in request URLs.
Apr 29, 2026
- v1.123.0Hyperswitch v1.123.0
This release introduces AWS SES support for Docker images, implements NTID flow for wallet tokens, and fixes critical Stripe dispute status mapping bugs. It also adds support for configurable region-based endpoints in Noon and enables config-based UAS eligibility.
Apr 14, 2026
- v2.0Audio2Face-3D Samples v2.0
This release updates the Audio2Face-3D sample client configurations to version 2.0. It introduces new support for diffusion models and updates regression models to version 2.3.1. Additionally, the update enhances facial animation capabilities by adding 17 tongue blendshapes and refactoring the configuration structure.
Mar 11, 2026
- v0.1.321Tally v0.1.321
Enhanced CSV parsing error diagnostics to help users debug silently failing data imports, including verbose flags and health checks.
Jan 17, 2026
- v0.1.319Tally v0.1.319
Introduced CSV export functionality and dynamic transaction description transformation via rules, along with simplified supplemental data source requirements.
Jan 14, 2026
- v0.1.311Tally v0.1.311
Added support for folder-based CSV sources and European CSV formats, while fixing tag collection logic to prevent unexpected behavior.
Jan 10, 2026
- v0.1.292Tally v0.1.292
Replaced the `run` command with `up`, introduced report diffing for tracking changes, and added experimental supplemental data sources with Python-like syntax.
Jan 6, 2026
- v0.8.16
This release removes CLI-based account creation and improves the `dockerd` wrapper cleanup process. Additionally, it eliminates all Cloud dependencies, commands, and flags following the Earthly Cloud shutdown.
Jul 16, 2025
- v1.3Audio2Face-3D Samples v1.3
Jun 16, 2025
- v0.2.0
This release introduces channel-concatenated control conditioning for Wan2.1 and CogView4, adds support for FLUX.1-dev and Wan2.1 I2V models, and provides multiple attention providers for both training and inference.
Apr 25, 2025