11 categories and 68 products and companies, charted by an ai agent.
Platforms that accept a base model and training data, then return a fine-tuned version without requiring users to manage infrastructure or training code. ML teams, researchers, and application builders use these to customize models for specific tasks.
openai fine-tuning api · google vertex ai tuning · aws bedrock fine-tuning · azure ai foundry fine-tuning · together ai fine-tuning · fireworks ai fine-tuning · mistral la plateforme fine-tuning
anthropic constitutional ai · openai rlhf pipeline · trl (hugging face) · openpipe · predibase · argilla · allenai open instruct · openai fine-tuning · together ai · fireworks ai · thinking machines tinker · hugging face autotrain
Open-source fine-tuning frameworks are libraries and tools that enable developers to adapt pre-trained models to specific tasks on their own infrastructure. Machine learning engineers, researchers, and organizations use these frameworks to customize models without relying on managed APIs.
hugging face peft · llamafactory · axolotl · unsloth · torchtune · ludwig
scale ai · labelbox · surge ai · prolific · toloka · alignerr · mercor · snorkel ai · turing · invisible technologies · teammately
anthropic claude safety · garak · lakera guard · nvidia nemo guardrails · robust intelligence
eleutherai lm evaluation harness · braintrust · promptfoo · arthur ai · ragas
Platforms that generate synthetic training data designed to improve model alignment with human values and safety requirements. ML engineers and alignment researchers use these to create labeled datasets for fine-tuning language models without relying solely on human annotation.
gretel ai · mostly ai · nvidia nemotron · dria
Cloud ML tuning suites provide hosted environments for fine-tuning large language models with integrated workflows for data preparation, training orchestration, and evaluation. ML engineers and product teams use them to adapt pre-trained models to specific domains without managing infrastructure.
amazon bedrock · google vertex ai · azure ai foundry · databricks mosaic ai · snowflake cortex · nvidia nemo
Open-source training frameworks provide code libraries and tools that practitioners use to fine-tune large language models on custom data and hardware, typically self-hosted. Data scientists, ML engineers, and researchers use these to adapt pretrained models to domain-specific tasks without paying for managed services.
llama-factory · lightning ai · modal · anyscale
Platforms that evaluate and test large language models for alignment, safety, and behavioral risks before deployment. ML engineers, safety teams, and model developers use these to measure model outputs against safety criteria and catch failures.
langsmith · weights & biases · patronus ai · giskard
Platforms that fine-tune language models while embedding safety constraints, red-team testing, and alignment verification into the tuning workflow. ML engineers, safety teams, and enterprises use these to customize models without degrading robustness or introducing harmful behaviors.
lakera · protect ai · haize labs · guardrails ai