Low-code framework for building custom AI

Ludwig, a declarative deep learning framework hosted by the Linux Foundation AI & Data, announced support for Python 3.12, PyTorch 2.7+, Pydantic 2, Transformers 5, and Ray 2.54, enabling users to train, fine-tune, and deploy AI models using a YAML config file with zero boilerplate Python. The framework now includes advanced features such as PatchTST and N-BEATS encoders for time-series forecasting, VLM fine-tuning for models like LLaVA and Qwen2-VL, and GRPO alignment, among others.

Declarative deep learning framework for LLMs, multimodal models, and tabular AI. Docs https://ludwig.ai · Getting Started https://ludwig.ai/latest/getting started/ · Examples https://ludwig.ai/latest/examples · Discord Ludwig is a declarative deep learning framework that lets you train, fine-tune, and deploy AI models — from LLM fine-tuning to tabular classification — using a YAML config file and zero boilerplate Python. Fine-tune Llama-3.1 with LoRA in one config file model type: llm base model: meta-llama/Llama-3.1-8B adapter: type: lora trainer: type: finetune epochs: 3 input features: - name: instruction type: text output features: - name: response type: text ludwig train --config model.yaml --dataset my data.csv Tech stack: Python 3.12 · PyTorch 2.7+ · Pydantic 2 · Transformers 5 · Ray 2.54 Ludwig is hosted by the Linux Foundation AI & Data https://lfaidata.foundation/ . | Feature | Description | |---|---| | PatchTST & N-BEATS encoders | State-of-the-art timeseries forecasting encoders with MASE/sMAPE metrics | | Advanced PEFT adapters | PiSSA, EVA, CorDA/LoftQ initializers; TinyLoRA, OFT, HRA, WaveFT, LN-Tuning, VBLoRA, C3A adapter types | | VLM fine-tuning | Train LLaVA, Qwen2-VL, InternVL via is multimodal: true with gated cross-attention | | HyperNetwork combiner | Conditioning-based feature fusion — one feature generates weights for others | | Nash-MTL & Pareto-MTL | Game-theoretic and preference-based multi-task loss balancing | | LLM config generation | ludwig generate config "describe your task" — LLM writes the YAML for you | | ModelInspector | Architecture analysis, weight collection, feature importance proxy | | Ray Serve & KServe | Distributed and Kubernetes-native model deployment shims | | GRPO alignment | Reward-model-free RLHF via Group Relative Policy Optimization | | torchao quantization + QAT | PyTorch-native int4/int8/float8 with Quantization-Aware Training | | Multi-adapter PEFT | Multiple named LoRA adapters with weighted merging TIES, DARE, SVD | | Native Optuna executor | GPT/TPE/CMA-ES samplers, pruning, resumable SQLite/PostgreSQL storage | | Timeseries forecasting | model.forecast dataset, horizon=N API with TimeseriesOutputFeature | | Muon & ScheduleFreeAdamW | New optimizers for large-scale pretraining and fine-tuning | | Image segmentation decoders | UNet, SegFormer, FPN decoders for semantic segmentation | pip install ludwig core pip install ludwig full all optional dependencies pip install ludwig llm LLM fine-tuning only Requires Python 3.12+. See contributing https://github.com/ludwig-ai/ludwig/blob/main/CONTRIBUTING.md for a full dependency matrix. Ludwig supports the full LLM fine-tuning spectrum: | Technique | Config key | |---|---| | Supervised fine-tuning SFT | trainer.type: finetune | | DPO / KTO / ORPO / GRPO alignment | trainer.type: dpo or kto , orpo , grpo | | LoRA / DoRA / VeRA / PiSSA | adapter.type: lora or dora , vera , lora + init weights: pissa | | 4-bit QLoRA bitsandbytes | quantization.bits: 4 | | torchao + QAT | quantization.backend: torchao | | Multi-adapter with merging | adapters: dict + merge: block | | VLM vision-language | is multimodal: true | model type: llm base model: meta-llama/Llama-3.1-8B quantization: bits: 4 adapter: type: lora prompt: template: | Instruction: {instruction} Input: {input} Response: input features: - name: prompt type: text output features: - name: output type: text trainer: type: finetune learning rate: 0.0001 batch size: 1 gradient accumulation steps: 16 epochs: 3 learning rate scheduler: decay: cosine warmup fraction: 0.01 backend: type: local export HUGGING FACE HUB TOKEN="<your token " ludwig train --config model.yaml --dataset "ludwig://alpaca" input features: - name: review text type: text encoder: type: bert - name: star rating type: number - name: product image type: image encoder: type: dinov2 output features: - name: recommended type: binary ludwig train --config model.yaml --dataset reviews.csv ludwig generate config "I have a CSV with age, income, education level, and I want to predict loan default" ludwig predict --model path results/experiment run/model --dataset new data.csv ludwig serve --model path results/experiment run/model POST http://localhost:8000/predict LLM Fine-Tuning - Supervised fine-tuning SFT on instruction/response pairs - Alignment training : DPO, KTO, ORPO, GRPO reward-model-free RLHF - PEFT adapters : LoRA, DoRA, VeRA, LoRA+, TinyLoRA, OFT, HRA, WaveFT, LN-Tuning, VBLoRA, C3A - LoRA initializers : PiSSA, EVA, CorDA, LoftQ for improved convergence - Multi-adapter PEFT : multiple named adapters on one base model, switchable at runtime; merge with TIES, DARE, SVD, magnitude pruning - Quantization : 4-bit/8-bit QLoRA bitsandbytes , torchao int4/int8/float8 with QAT - VLM fine-tuning : LLaVA, Qwen2-VL, InternVL via is multimodal: true - Sequence packing for efficient training on variable-length inputs - Paged and 8-bit optimizers for memory-efficient training Multimodal & Tabular Models - Input modalities : text, numbers, categories, binary, sets, bags, sequences, images, audio, timeseries, vectors, dates - Text encoders : any HuggingFace Transformer BERT, RoBERTa, ModernBERT, Qwen3, Llama-3.1, etc. , plus Mamba-2, Jamba - Image encoders : DINOv2, ConvNeXt, EfficientNet, ViT, CAFormer, ConvFormer, PoolFormer, TIMM 1000+ models - Timeseries encoders : PatchTST, N-BEATS, CNN, RNN, Transformer; MASE and sMAPE metrics; model.forecast API - Combiners : concat, transformer, tab transformer, FT-Transformer, TabNet, TabPFN v2, HyperNetwork, ProjectAggregate, GatedFusion, Perceiver - Multi-task learning : multiple output features in a single model; Nash-MTL, Pareto-MTL, FAMO, GradNorm, uncertainty loss balancing - Image segmentation : UNet, SegFormer, FPN decoders Training Infrastructure - Distributed training : HuggingFace Accelerate with DDP, FSDP, DeepSpeed zero-code changes - Ray backend : training across a Ray cluster, larger-than-memory datasets via Ray Data - Automatic batch size selection and learning rate range test - Mixed precision fp16/bf16 , gradient checkpointing, gradient accumulation - Optimizers : AdamW, Adafactor, SGD, Muon, ScheduleFreeAdamW, Lion, paged/8-bit variants - Learning rate schedulers : cosine, linear, polynomial, reduce-on-plateau, OneCycleLR - Model Soup : uniform and greedy checkpoint averaging for better generalization at zero inference cost - Modality dropout for robust multimodal models Hyperparameter Optimization - Executors : Ray Tune ASHA, PBT, Bayesian and native Optuna auto/GP/TPE/CMA-ES - Optuna persistence : SQLite or PostgreSQL for resumable HPO runs - Pruning with Optuna's MedianPruner and HyperbandPruner - Search spaces : uniform, log-uniform, choice, randint, quantized - Full Ludwig config is searchable — any nested parameter can be a hyperparameter Production & Deployment - REST API : FastAPI server with Prometheus metrics and structured logging ludwig serve - vLLM serving : OpenAI-compatible API with PagedAttention and continuous batching - Ray Serve : distributed deployment with auto-scaling and traffic splitting - KServe : Kubernetes-native deployment with Open Inference Protocol v2 - Model export : SafeTensors default , torch.export .pt2 bundles, ONNX - HuggingFace Hub : ludwig upload hf hub — push model + auto-generated model card - Docker : prebuilt containers at ludwigai/ludwig https://hub.docker.com/u/ludwigai Tooling & Integrations - Experiment tracking : TensorBoard, Weights & Biases, Comet ML, MLflow, Aim Stack - Model inspection : ModelInspector — weight enumeration, architecture summary, feature importance proxy - Visualizations : learning curves, confusion matrices, calibration plots, ROC curves, hyperopt analysis - AutoML : ludwig.automl.auto train — give it a dataset and a time budget; the YAML-driven search space samples encoder/combiner/decoder combinations and validates them before training - Dataset quality checks : from ludwig.utils.dataset quality import check dataset quality — validates a DataFrame before training missing values, class imbalance, near-duplicate columns, ID leakage, … - OpenML integration : load any OpenML task directly — OpenMLLoader fetches by task ID and caches locally as Parquet - LLM config generation : ludwig generate config "describe your task" — LLM writes the YAML - K-fold cross-validation : ludwig experiment --k fold N - Dataset Zoo : 70+ built-in benchmark datasets ludwig://mnist , ludwig://alpaca , … | Use Case | Link | |---|---| | LLM instruction tuning LoRA + QLoRA | examples/llm https://ludwig.ai/latest/examples/llm/llm finetuning | | DPO / GRPO alignment | examples/llm/alignment https://ludwig.ai/latest/examples/llm/alignment | | Advanced PEFT PiSSA, OFT, VBLoRA, … | examples/llms/peft advanced https://ludwig.ai/latest/examples/llms/peft advanced | | VLM fine-tuning LLaVA, Qwen2-VL | examples/vlm https://github.com/ludwig-ai/ludwig/tree/main/examples/vlm | | Use Case | Link | |---|---| | Binary classification Titanic | examples/titanic https://ludwig.ai/latest/examples/titanic | | Tabular classification census income | examples/adult census income https://ludwig.ai/latest/examples/adult census income | | Multimodal classification | examples/multimodal classification https://ludwig.ai/latest/examples/multimodal classification | | Multi-task learning | examples/multi task https://ludwig.ai/latest/examples/multi task | | Use Case | Link | |---|---| | Timeseries forecasting PatchTST, N-BEATS | examples/forecasting https://ludwig.ai/latest/examples/forecasting | | Weather forecasting | examples/weather https://ludwig.ai/latest/examples/weather | | Image classification MNIST | examples/mnist https://ludwig.ai/latest/examples/mnist | | Semantic segmentation | examples/semantic segmentation https://ludwig.ai/latest/examples/semantic segmentation | | Use Case | Link | |---|---| | Text classification | examples/text classification https://ludwig.ai/latest/examples/text classification | | Named entity recognition | examples/ner tagging https://ludwig.ai/latest/examples/ner tagging | | Machine translation | examples/machine translation https://ludwig.ai/latest/examples/machine translation | | Speech recognition | examples/speech recognition https://ludwig.ai/latest/examples/speech recognition | | Speaker verification | examples/speaker verification https://ludwig.ai/latest/examples/speaker verification | - Zero boilerplate — no training loop, no data pipeline, no evaluation code. The YAML config is the entire program. - Best-in-class LLM support — full spectrum from LoRA to GRPO alignment, torchao QAT, and VLM fine-tuning, all in config. - Multimodal out of the box — mix text, images, numbers, audio, and timeseries with one config change. - Scale without code changes — go from laptop → multi-GPU → Ray cluster by changing backend.type . - Expert control when you need it — every activation function, scheduler, and optimizer is configurable. - Reproducible research — every run is logged and the full config is saved. Compare experiments with ludwig visualize . - Ludwig: A Type-Based Declarative Deep Learning Toolbox https://arxiv.org/pdf/1909.07930.pdf 2019 - Declarative Machine Learning Systems https://arxiv.org/pdf/2107.08148.pdf 2021 - Ludwig's State-of-the-Art Benchmarks https://openreview.net/pdf?id=hwjnu6qW7E4 - Discord https://discord.gg/CBgdrGnZjy — ask questions, share what you've built - GitHub Issues https://github.com/ludwig-ai/ludwig/issues — bugs and feature requests - X / Twitter https://twitter.com/ludwig ai — announcements - Medium https://medium.com/ludwig-ai — tutorials and deep-dives