Examples Gallery

CodefyUI ships a library of ready-to-run example graphs under examples/. Whenever the active tab has an empty canvas, the gallery appears right on the canvas — pick a card and the graph loads into the tab, ready to Run. The other ways to open it are listed under When the canvas is not empty. You can also run any example headless with the CLI Graph Runner.

The gallery is organized into ordered sections:

Section	Contents
Quick Start	The three pinned starters: Train CNN on MNIST, Inference CNN on MNIST, and Api-Function (graph-as-a-function demo).
Advanced Examples	Every other runnable builtin example — LLM (Word Embedding Analogy with the offline demo-16d backend, Sentence Similarity (zh-TW) with a real sentence encoder, Train a Causal LM on TinyStories, and the two retrieval examples RAG, fully local and RAG with a chat API), Diffusion (Forward Process, Toy Sampling, Mini U-Net), Classical ML (Iris KNN, tabular pipeline), Transformer (MoE routing), RNN, RL (RLHF reward + KL), VLA (Train a VLA on PushWorld — needs a CUDA GPU and about an hour; the recipe is in its README), and the remaining trainers (GPT-Mini, ResNet-CIFAR10, and the measured ResNet-18 / CIFAR-10 baseline — see Reproducing Baselines).
Plugin Examples	Examples shipped by installed plugins (and any unrecognized categories). Only shown when present.
Model Architectures	15 classic architecture walkthroughs, always listed last: ResNet, ConvNeXt, EfficientNet, UNet, ViT, SwinTransformer, BERT, GPT, LLaMA, DiT, LSTM TimeSeries, BiGRU SpeechRecognition, Seq2Seq Attention, DQN Atari, PPO Robotics.

On disk the examples are grouped by topic folder: Classical/, Diffusion/, LLM/, Model_Architecture/, RL/, RNN/, Transformer/, Usage_Example/, and VLA/.

Every listed example runs offline out of the box, with four exceptions:

• Train a Causal LM on TinyStories downloads the TinyStories corpus from the Hugging Face Hub and the gpt2 BPE ranks on its first run, and needs a GPU with headroom for a 203,668,480-parameter model. Its card leads with both requirements; the full recipe, the token budgets and the memory levers are in the README.md beside the graph, at examples/LLM/TrainCausalLM-TinyStories/. Both downloads are cached, so later runs are offline too.
• Sentence Similarity (zh-TW) needs the sentence-embeddings pack, which is a one-off install from the Package Center (toolbar > Settings > Optional packs) or cdui packs install sentence-embeddings — a run never downloads it for you. Once the pack is in, the example runs offline on CPU in a few seconds. See Optional Packs.
• RAG, fully local needs two downloads rather than one: qwen2.5-0.5b-instruct from the rag pack, and the multilingual-e5-small item of sentence-embeddings — about 1.5 GB together. Installing rag brings that pack's Python packages but no encoder, so the second item has to be picked as well. With both in, nothing leaves the machine: the documents, the search and the generation all happen locally, at a few tokens per second on a CPU, so expect the answer to take anywhere from a few seconds to tens of seconds — an estimate from the model size rather than a measurement, and much faster on a GPU.
• RAG with a chat API is that same retrieval chain with LLMChat in the last box, so it needs only multilingual-e5-small — plus somewhere to send the prompt. Out of the box that is a local Ollama with ollama pull qwen2.5:0.5b, which still keeps everything on this machine; switching provider to a hosted model sends the retrieved chunks to a third party and needs a key in the environment.

The two RL architecture graphs (DQN Atari, PPO Robotics) feed their networks from a synthetic observation tensor (TensorCreate, randn) instead of a live gym environment, so no ale-py/mujoco install is needed — swap in an EnvWrapper node to drive them from a real environment.

Loading an example

• In the app — open a new (empty) tab; the gallery overlay appears on the canvas. Pick a card and the graph loads into the tab, ready to Run.

• From the CLI — point run_graph.py at the graph's JSON:

  cd backend
  python run_graph.py ../examples/Usage_Example/CNN-MNIST/TrainCNN-MNIST/graph.json

When the canvas is not empty

The Template Gallery opens from the toolbar's Templates button, from Browse all templates on the empty-canvas overlay, and from the sidebar's Templates tab. It groups the examples by category, and selecting one shows its description, its node and connection counts, and whether it is built in or comes from a plugin. Each example offers two actions:

• Open in new tab leaves the current graph alone.
• Insert into this canvas adds the example to the graph you are editing: the inserted nodes get fresh ids and are placed below your current graph, so nothing is overwritten, and one undo removes them.

In the sidebar's Templates tab, click an example to insert it, or drag it onto the canvas where you want it.

A good first run

Load Train CNN on MNIST, then:

1. Record node outputs and Persist weights between runs are both on by default — check them in the Settings popover (Recording & Inspection and Training Behavior).
2. Click Run and watch the live loss chart in the Training tab.
3. Click a Conv2d node to inspect its kernels and activations in the Teaching Inspector.
4. Run again — with weights persisted, the model keeps learning across runs.

Training also saves model_weights.pt (under backend/data/models/). After that, load Inference CNN on MNIST — it classifies test_digit.png, a real MNIST digit bundled under backend/data/images/, using the weights you just trained.