# TRL - **Canonical URI:** https://www.vioscale.ai/software/trl - **Category:** Fine Tuning Platforms - **Homepage:** https://huggingface.co/docs/trl - **Also known as:** trl - **Profile claimed by vendor:** no - **Last updated:** 2026-08-26T19:36:22.478Z ## Vioscale score **56.9 / 100**, confidence 23% (low). Computed 2026-09-01. Composite of weighted, independently-sourced signals (no user reviews, no vendor payment). | Signal | Score | Weight | Contribution | Evidence present | |---|--:|--:|--:|:--:| | price_level | 0 | 0.052 | 0 | - | | reliability | 0 | 0.073 | 0 | - | | capabilities | 74.7 | 0.084 | 6.3 | ✓ | | repo_stars | 80.8 | 0.026 | 2.1 | ✓ | | integrations | 26 | 0.09355555555555557 | 2.4 | ✓ | | dependent_projects | 28.7 | 0.063 | 1.8 | ✓ | | dev_activity | 63.1 | 0.094 | 5.9 | ✓ | | release_cadence | 94.4 | 0.052 | 4.9 | ✓ | | security_posture | 0 | 0.073 | 0 | - | | package_downloads | 0 | 0.136 | 0 | - | | security_score | 0 | 0.042 | 0 | - | | pricing_transparency | 0 | 0.084 | 0 | - | | community_qa_activity | 0 | 0.063 | 0 | - | ## Pricing _As of 2026-08-21, [verify at source](https://huggingface.co/pricing). Independently observed._ free_open_source ## About A toolkit for enhancing transformer-based language models after initial training using various preference-alignment and optimization methods. It includes trainers integrated with the Hugging Face ecosystem and supports distributed training across multiple GPUs and clusters. _Independently observed._ ## Platform & deployment - **Platforms:** CLI - **Deployment:** Self-hosted ## Integrations (7) _Independently observed._ - transformers - PEFT - DeepSpeed - Accelerate - vLLM - torch_xla - torch.distributed ## Capabilities _The capabilities that matter for Fine Tuning Platforms. "-" = undocumented, not absent._ | Capability | Supported | |---|:--:| | **Capabilities** | | | Architecture model | Open source library local GPU | | Peft lora and qlora parameter efficient tuning | ✓ | | Rlhf and dpo preference alignment optimization | ✓ | | Serverless hosting of fine tuned adapters lorax | - | | Four bit and eight bit memory quantization | ✓ | | Flash attention and xformers compilation | - | | Distributed multi GPU orchestration fsdp deepspeed | ✓ | | Synthetic data generation and evaluation pipeline | - | | Native huggingface hub push pull integration | - | | Weights and biases wandb experiment tracking | - | | SOC2 type ii | - | | Mit or apache permissive oss license | ✓ | | Pricing model | Free open source | ## Release history | Version | Date | Type | |---|---|---| | [v1.10.0](https://github.com/huggingface/trl/releases/tag/v1.10.0) | 2026-08-13 | stable | | [v1.9.2](https://github.com/huggingface/trl/releases/tag/v1.9.2) | 2026-07-28 | stable | | [v1.9.1](https://github.com/huggingface/trl/releases/tag/v1.9.1) | 2026-07-26 | stable | | [v1.9.0](https://github.com/huggingface/trl/releases/tag/v1.9.0) | 2026-07-21 | stable | | [v1.8.0](https://github.com/huggingface/trl/releases/tag/v1.8.0) | 2026-07-09 | stable | | [v1.7.1](https://github.com/huggingface/trl/releases/tag/v1.7.1) | 2026-07-04 | stable | | [v1.7.0](https://github.com/huggingface/trl/releases/tag/v1.7.0) | 2026-06-25 | stable | | [v1.6.0](https://github.com/huggingface/trl/releases/tag/v1.6.0) | 2026-06-11 | stable | | [v1.5.1](https://github.com/huggingface/trl/releases/tag/v1.5.1) | 2026-05-27 | stable | | [v1.5.0](https://github.com/huggingface/trl/releases/tag/v1.5.0) | 2026-05-25 | stable | | [v1.4.0](https://github.com/huggingface/trl/releases/tag/v1.4.0) | 2026-05-09 | stable | | [v1.3.0](https://github.com/huggingface/trl/releases/tag/v1.3.0) | 2026-04-26 | stable | | [v1.2.0](https://github.com/huggingface/trl/releases/tag/v1.2.0) | 2026-04-17 | stable | | [v1.1.0](https://github.com/huggingface/trl/releases/tag/v1.1.0) | 2026-04-12 | stable | | [v1.0.0](https://github.com/huggingface/trl/releases/tag/v1.0.0) | 2026-03-31 | stable | _… and 5 earlier release(s)._ ## Security & compliance Known vulnerabilities: 0 (0 in the last 12 months) ([source](https://advisories.ecosyste.ms/api/v1/advisories?ecosystem=pypi&package_name=trl&per_page=100)). A count reflects scale and disclosure, not quality. ## Facts Every value below carries its source and our confidence. Facts are re-crawled on a freshness schedule. ### pricing | Attribute | Value | Source | Retrieved | Confidence | |---|---|---|---|---| | pricing.model | commercial | [link](https://huggingface.co/docs/trl) | 2026-08-26 | 40% (low) | | pricing.price_level | unknown | [link](https://huggingface.co/pricing) | 2026-08-21 | 60% (medium) | ### integrations | Attribute | Value | Source | Retrieved | Confidence | |---|---|---|---|---| | integrations.count | 7 | [link](https://huggingface.co/pricing) | 2026-08-21 | 60% (medium) | ### activity | Attribute | Value | Source | Retrieved | Confidence | |---|---|---|---|---| | activity.commits_last_30d | 100 | [link](https://github.com/huggingface/trl/pulse) | 2026-08-26 | 65% (medium) | ### adoption | Attribute | Value | Source | Retrieved | Confidence | |---|---|---|---|---| | adoption.github_stars | 19,154 | [link](https://github.com/huggingface/trl) | 2026-08-26 | 90% (high) | | adoption.dependent_repos | 52 | [link](https://packages.ecosyste.ms/api/v1/packages/lookup?repository_url=https%3A%2F%2Fgithub.com%2Fhuggingface%2Ftrl) | 2026-08-26 | 85% (high) | ### language | Attribute | Value | Source | Retrieved | Confidence | |---|---|---|---|---| | language.primary | Python | [link](https://github.com/huggingface/trl) | 2026-08-26 | 90% (high) | ### license | Attribute | Value | Source | Retrieved | Confidence | |---|---|---|---|---| | license.spdx | Apache-2.0 | [link](https://github.com/huggingface/trl) | 2026-08-26 | 95% (high) | ### release | Attribute | Value | Source | Retrieved | Confidence | |---|---|---|---|---| | release.cadence_days | 10 | [link](https://github.com/huggingface/trl/releases) | 2026-08-26 | 70% (medium) | ### security | Attribute | Value | Source | Retrieved | Confidence | |---|---|---|---|---| | security.vulnerabilities | `{"count":0,"source":"https://advisories.ecosyste.ms/api/v1/advisories?ecosystem=pypi&package_name=trl&per_page=100","last_12m":0,"max_severity":null}` | [link](https://advisories.ecosyste.ms/api/v1/advisories?ecosystem=pypi&package_name=trl&per_page=100) | 2026-08-26 | 90% (high) | --- *Source: Vioscale (https://www.vioscale.ai/software/trl). Independent, evidence-based software intelligence. Cite the canonical URI.*