ORIGINAL FORGE · PLAIN ENGLISH

Original Forge UI Glossary

Decode the labels before changing them. These definitions explain each term’s job, the term it is most often confused with, and the guide that shows it in a working task.

Terms defined63Source statedfdcbabEvidence checked31 Aug 2026
DIRECT ANSWER

Read a Forge control as one part of a chain.

VERIFIED

A model file supplies a component. A generation control changes the run. A memory control changes where the work fits. An editing tool adds another input. The same word—especially “model,” “weight,” or “strength”—can belong to different parts of that chain.

This glossary covers Original Forge interface and workflow language. It is not a general machine-learning encyclopedia, a compatibility promise for arbitrary files, or documentation for reForge, Forge Classic, or Forge Neo.

DO NOT MERGE THESE

Six pairs that look related but do different jobs.

01 · LOOKUP

Search the label you see in Forge.

Search accepts terms, aliases, and definition words. Category buttons narrow the same complete, server-rendered list; they do not load definitions after a click.

/

Showing all 63 terms

01

Forge & interface

6 terms
hash, revision, build versionVERIFIED

Commit

A specific recorded state of the repository, identified by a Git hash such as dfdcbab.

Not the same as

A branch can move after an update, while a commit hash names one fixed code state. A release date and commit date are not interchangeable.

plugin, add-on, extensions tabSTALE SNAPSHOT

Extension

Additional code that adds or changes WebUI behavior, commonly installed under the extensions directory.

Not the same as

An extension can be version-sensitive. Its presence in the interface does not prove compatibility with the current Forge commit.

built-in extension, extensions-builtinVERIFIED

Integrated extension

Feature code shipped inside the Original Forge repository under extensions-builtin, such as its ControlNet and LoRA integrations.

Not the same as

Integrated means bundled with this repository; it does not mean that every third-party model file or workflow is included or supported.

Stable Diffusion WebUI Forge, lllyasviel ForgeVERIFIED

Original Forge

The project in lllyasviel/stable-diffusion-webui-forge, built from the familiar Stable Diffusion WebUI interface with a resource-management backend and integrated features.

Not the same as

reForge, Forge Classic, and Forge Neo are separate downstream projects; their features and releases do not prove Original Forge behavior.

UI, preset, sd, xl, flux, allVERIFIED

UI preset

The top-level Forge selector that changes which model-oriented controls are shown; current choices include sd, xl, flux, and all.

Not the same as

A UI preset does not download or load a checkpoint and is not a model-family converter.

web interface, browser UIVERIFIED

WebUI

The browser interface used to control a locally running Forge process.

Not the same as

Closing the browser tab does not necessarily stop the Forge process; localhost is an address on your machine, not the application itself.

02

Models & files

14 terms
base model, Stable Diffusion checkpoint, model checkpointVERIFIED

Checkpoint

The primary model selection used to generate an image; in Forge it may be a traditional bundled checkpoint or a supported diffusion-model file.

Not the same as

A checkpoint is not a LoRA, VAE, text encoder, or ControlNet model, even when all use .safetensors files.

CLIP, clip_l, text encoderVERIFIED

CLIP-L

A CLIP text-encoder component used by some supported model layouts, selected through Forge’s VAE / Text Encoder control when supplied separately.

Not the same as

CLIP-L is not CLIP skip, and it is not interchangeable with T5-XXL simply because both process text.

low bits selector, Automatic, bnb-nf4, float8VERIFIED

Diffusion in Low Bits

The Forge model-loading control that chooses an automatic or explicit lower-bit representation for supported diffusion weights.

Not the same as

Automatic preserves the checkpoint’s own precision when possible; this control is not the same as selecting a different checkpoint.

UNet, DiT, transformer, denoiserVERIFIED

Diffusion model

The model component that performs the iterative denoising work; newer layouts can store it separately from the VAE and text encoders.

Not the same as

UNet and DiT describe different architectures. Treat the publisher’s model layout as the contract instead of calling every diffusion model a UNet.

textual inversion, TI embeddingVERIFIED

Embedding

A learned prompt token loaded by the textual-inversion system and used with a compatible base family.

Not the same as

An embedding is not a LoRA or checkpoint; its trigger belongs in the prompt rather than the Checkpoint selector.

float8, float16, precisionVERIFIED

FP8 / FP16

Floating-point formats used for model weights or computation; lower precision can reduce memory use but support and behavior depend on the component and hardware.

Not the same as

A precision label is not a model family, and changing precision cannot repair an incompatible VAE, encoder, or LoRA.

GGUF model, quantized model fileVERIFIED

GGUF

A supported container format commonly used for quantized diffusion-model or text-encoder weights in split model layouts.

Not the same as

GGUF names a file format, not one precision, architecture, or guarantee that an arbitrary file will load in Forge.

low-rank adaptation, network, style LoRAVERIFIED

LoRA

A smaller learned modification applied to a compatible base model, usually through an Extra Networks card or prompt token.

Not the same as

A LoRA cannot generate by itself and does not automatically cross SD 1.x, SDXL, or FLUX families.

architecture, base family, SD 1.5, SDXL, FLUXVERIFIED

Model family

A compatible architecture group that determines which checkpoints, LoRAs, text encoders, VAEs, and control models can work together.

Not the same as

Matching file extensions do not make assets from different families compatible.

bnb-nf4, BitsAndBytes NF4, 4-bitSTALE SNAPSHOT

NF4

A 4-bit BitsAndBytes quantization choice exposed by Forge for supported model-loading paths.

Not the same as

NF4 is not a universal mode for every checkpoint, and an NF4 filename or selector does not guarantee a speed improvement on every GPU.

low-bit, lower precision, compressed weightsVERIFIED

Quantization

Representing model weights with fewer bits to reduce their storage or memory cost, with quality, speed, and compatibility trade-offs that depend on the method and hardware.

Not the same as

Quantization does not shrink the image dimensions or guarantee faster generation.

T5, t5xxl, text encoderVERIFIED

T5-XXL

A large text-encoder component required by some FLUX model layouts and selected separately when it is not bundled.

Not the same as

A quantized T5 file changes storage and memory characteristics; it does not become the base diffusion model.

prompt encoder, CLIP, T5VERIFIED

Text encoder

A model component that converts prompt text into conditioning the diffusion model can use.

Not the same as

A text encoder is not the diffusion model or tokenizer. Split layouts may require more than one encoder.

variational autoencoder, decoder, vae fileVERIFIED

VAE

The component that converts between image pixels and the model’s latent representation, including the final decode into a viewable image.

Not the same as

A VAE does not replace the checkpoint. A missing or wrong VAE can fail at the decode stage even when sampling appears to run.

03

Generation controls

16 terms
number of batches, iterationsVERIFIED

Batch count

How many batches Forge runs one after another for the same Generate action.

Not the same as

Batch count multiplies total outputs and time; it is not the number processed simultaneously.

images at once, parallel batchVERIFIED

Batch size

How many images Forge asks the model to process together in one batch.

Not the same as

Batch size usually affects peak memory more directly than Batch count and is not the total number of repeated batches.

CFG, guidance scale, classifier-free guidanceVERIFIED

CFG Scale

The guidance strength applied between positive and unconditional conditioning in the standard CFG path.

Not the same as

CFG Scale is separate from Distilled CFG Scale; setting CFG to 1 disables the negative-conditioning branch in current processing code.

Clip skip, stop at last layersVERIFIED

CLIP skip

A compatibility control that chooses how early to stop through supported CLIP text-encoder layers.

Not the same as

CLIP skip is not CLIP-L, and it is not meaningful for every model family.

denoise, strengthVERIFIED

Denoising strength

In img2img, inpainting, or a high-resolution second pass, the fraction of the denoising process used to transform the starting latent.

Not the same as

It is not ControlNet weight or image opacity; higher values generally permit more departure from the source.

Distilled CFG, guidance distillationVERIFIED

Distilled CFG Scale

A separate guidance value passed to models that expose distilled guidance, including documented FLUX workflows.

Not the same as

It does not mean the same thing as standard CFG Scale. Do not copy one number into both fields by assumption.

highres fix, high-resolution pass, second passVERIFIED

Hires. fix

A two-stage txt2img workflow that first generates at one size, upscales, then denoises a higher-resolution pass.

Not the same as

It is not a simple post-process upscale; it adds another diffusion pass, memory load, and its own optional sampler, scheduler, checkpoint, VAE, encoder, prompt, and denoising controls.

image to image, image2imageVERIFIED

Img2img

The workflow that encodes a source image and denoises it toward a new result, controlled strongly by Denoising strength.

Not the same as

Img2img is not inpainting unless a mask limits where the edit is applied.

negative conditioningVERIFIED

Negative prompt

Text used for the unconditional or negative side of classifier-free guidance in workflows that use it.

Not the same as

In current Forge processing, CFG Scale 1 skips unconditional conditioning, so the negative prompt has no effect in that mode.

positive prompt, text promptVERIFIED

Prompt

Text that describes or conditions what the model should generate, interpreted through the active model’s encoders and training.

Not the same as

A prompt cannot compensate for an incompatible model component or a ControlNet map that encodes the wrong structure.

sampling method, Euler, DPM++VERIFIED

Sampler

The numerical method that advances the latent through the denoising process.

Not the same as

A sampler is not the noise schedule. Forge exposes Sampler and Schedule type as separate choices.

steps, iteration countVERIFIED

Sampling steps

The number of denoising steps requested from the sampler for a generation pass.

Not the same as

More steps are not automatically better; useful ranges are model- and sampler-dependent.

schedule type, noise schedule, Karras, normalVERIFIED

Scheduler

The rule that distributes noise levels or timesteps across the requested sampling steps.

Not the same as

A scheduler does not replace the sampler. Automatic lets Forge resolve the schedule associated with the selected sampler.

random seed, variation seedVERIFIED

Seed

The number used to initialize generation noise; keeping it fixed helps isolate the effect of one changed setting.

Not the same as

The same seed is not a cross-version promise when the model, sampler, scheduler, dimensions, software, or other inputs differ.

text to image, text2imageVERIFIED

Txt2img

The workflow that starts generation from text conditioning and an initial latent rather than from a source image.

Not the same as

ControlNet input images can guide a txt2img run, but that does not turn the tab into img2img.

resize model, ESRGAN, latent upscalerVERIFIED

Upscaler

The algorithm or model used to enlarge an image or latent, either between Hires. fix passes or as a post-processing operation.

Not the same as

An upscaler cannot recover ground-truth detail that was never present, and not every upscaler performs a diffusion pass.

04

Memory & loading

11 terms
Async swap method, asynchronous swapSTALE SNAPSHOT

Async

The alternative Forge swap method intended to overlap supported weight movement with other work when the environment can do so reliably.

Not the same as

Async is not guaranteed to be faster and can be less suitable on some systems; prove it with the same workload.

GPU Weights MB, weight budgetVERIFIED

GPU Weights

Forge’s megabyte target for how much model-weight data should remain on the GPU when supported swapping is active.

Not the same as

It is not total VRAM use. Setting it near the card’s full capacity can leave too little space for inference and cause fallback, severe slowdown, or an out-of-memory error.

compute headroom, runtime memory, activation memoryVERIFIED

Inference memory

The GPU memory needed while the model actually computes, beyond the space occupied by persistent weights.

Not the same as

A model fitting as weights does not prove that the requested image size, batch, attention work, VAE decode, or second pass will fit.

NeverOOM, historical checkboxSTALE SNAPSHOT

Never OOM

A February 2024 Forge memory feature described as splitting and offloading work to avoid some out-of-memory failures.

Not the same as

This is historical terminology, not a promise that every workload fits or that the same checkbox exists unchanged in the inspected build.

CPU offload, model offloading, swapVERIFIED

Offload

Keeping some model components or weights outside dedicated VRAM and moving them when needed to reduce peak GPU memory pressure.

Not the same as

Offload trades memory pressure for transfers and latency; it does not make RAM, bandwidth, or compute unlimited.

Queue swap methodSTALE SNAPSHOT

Queue

The conservative Forge swap-method choice documented for moving model weights in an ordered path.

Not the same as

Queue here is a memory-transfer strategy, not a list of pending image-generation jobs.

shared memory, shared VRAMSTALE SNAPSHOT

Shared GPU memory

System memory that the operating system can expose to GPU workloads when the platform supports it.

Not the same as

It is slower than dedicated VRAM and should not be added to dedicated VRAM as if both had equal performance.

CPU, Shared, offload locationVERIFIED

Swap Location

The Forge control that chooses where supported swapped weights wait outside their active GPU allocation; current labels are CPU and Shared.

Not the same as

The choice changes the memory and transfer path, not the model’s creative behavior.

Queue, Async, model swapVERIFIED

Swap Method

The Forge selector for the strategy used to move supported weights between GPU memory and the chosen swap location.

Not the same as

Queue and Async are strategies, not quality levels. Hardware and driver behavior determine which is useful.

CPU RAM, main memoryVERIFIED

System RAM

The computer’s main memory, used by Forge, Python, loaded files, offloaded weights, and the operating system.

Not the same as

System RAM is not VRAM. Moving weights out of VRAM can raise RAM use and transfer cost.

video memory, GPU memoryVERIFIED

VRAM

Memory available to the GPU for loaded weights, active computation, intermediate tensors, and other graphics or system use.

Not the same as

Advertised VRAM is not all available to Forge, and GPU Weights is not a live free-VRAM meter.

05

Editing & guidance

12 terms
alpha, transparency, RGBASTALE SNAPSHOT

Alpha channel

The per-pixel channel that records opacity in an image with transparency.

Not the same as

A checkerboard preview is only a display convention; verify that the saved file retains an alpha channel.

ControlNet weight, conditioning strengthVERIFIED

Control Weight

The strength assigned to a ControlNet unit’s influence during its active timestep range.

Not the same as

It is not LoRA weight, denoising strength, or GPU weight memory. More control weight is not automatically more accurate.

ControlNet Integrated, structure controlVERIFIED

ControlNet

An integrated workflow that feeds an additional structural signal—such as pose, edges, or depth—into compatible diffusion generation.

Not the same as

The interface does not bundle every control-model weight, and ControlNet does not by itself preserve identity, clothing, or exact pixels.

control model, ControlNet weightsVERIFIED

ControlNet model

A separate model that applies a specific kind of control signal during diffusion and must match the intended map type and base family.

Not the same as

It is not the preprocessor or base checkpoint. Appearing in the dropdown proves discovery, not compatibility.

canvas editor, mask canvasSTALE SNAPSHOT

Forge Canvas

Forge’s Gradio 4 image canvas for tasks such as drawing masks, panning, zooming, and editing image inputs.

Not the same as

Canvas interaction is an interface feature, not a generation model. Tablet and browser behavior in the announcement is dated evidence.

whole picture, only masked, crop regionVERIFIED

Inpaint area

The choice between processing the whole image or a crop around the masked region during inpainting.

Not the same as

Only masked changes the working crop and effective detail; it is not the same setting as Mask mode.

masked editing, local editVERIFIED

Inpainting

Generating a replacement for a masked region while using the surrounding image and selected masked-content settings as context.

Not the same as

Inpainting does not automatically preserve every unmasked pixel; crop mode, mask blur, denoising, model, and resolution still matter.

transparent image, transparency extensionSTALE SNAPSHOT

LayerDiffuse

An integrated Forge feature announced for generating or editing images with transparency-related workflows.

Not the same as

It is not ordinary background removal and requires a compatible workflow; the 2024 announcement is a dated feature snapshot.

inpaint mask, masked areaVERIFIED

Mask

A grayscale or painted selection that tells the inpainting workflow which region is targeted, according to the active mask mode.

Not the same as

White-versus-black meaning depends on Mask mode; inspect the selection instead of assuming from appearance alone.

mask feather, blur radiusVERIFIED

Mask blur

The control that softens the mask edge so the generated region can blend into nearby pixels.

Not the same as

It does not blur the whole output and cannot repair a mask placed over the wrong area.

ControlNet Pixel PerfectVERIFIED

Pixel Perfect

A ControlNet option that calculates preprocessor resolution from input and output geometry.

Not the same as

It does not validate model family, map type, control weight, or the quality of the detected map.

annotator, detector, pose detector, Canny, depthVERIFIED

Preprocessor

A ControlNet-stage tool that converts a source image into a control map such as edges, pose, or depth.

Not the same as

The preprocessor creates the signal; the ControlNet model consumes it. Use None when the input is already the required map.

06

Outputs & records

4 terms
reproduction record, settings record, test receiptVERIFIED

Generation receipt

This guide’s name for the minimum evidence needed to reproduce a result: build, files, prompt, seed, controls, console outcome, and saved output metadata.

Not the same as

It is an editorial testing term, not an official Forge button or file format.

generation parameters, PNG metadata, parameters stringVERIFIED

Infotext

Forge’s text record of generation inputs such as prompts, seed, steps, sampler, schedule, guidance, dimensions, and selected model details.

Not the same as

Infotext improves reproducibility but cannot embed every external file, extension version, environment detail, or hidden default.

latent image, latent spaceVERIFIED

Latent

The internal compressed representation that the diffusion process denoises before a VAE decodes it into pixels.

Not the same as

A latent preview is not necessarily the final decoded image; a VAE-stage problem can still produce a black or broken output.

PNG metadata tab, read parametersVERIFIED

PNG Info

The WebUI tab that reads embedded generation information from a supported image and can send recovered settings to another workflow.

Not the same as

A PNG can have metadata stripped by another application, and reading settings does not supply missing checkpoints or extensions.

02 · FROM TERM TO TASK

A definition tells you what a control is. The task guide tells you what to do next.

Use the glossary to remove ambiguity, then move to the canonical workflow. That page owns the sequence, safe baseline, failure checks, and evidence receipt.

03 · QUESTIONS USERS TYPE

The confusing parts, answered together.

These answers connect terms that are commonly searched as pairs. They stay inside the behavior verified for the inspected Original Forge code or carry a dated-evidence warning.

What is a checkpoint in Stable Diffusion WebUI Forge?

The checkpoint is the primary model selected for generation. It may bundle several components or, in a supported split layout, work with separately selected VAE and text encoders. It is not a LoRA or ControlNet model.

What is the difference between a checkpoint and a LoRA?

A checkpoint supplies the base generation model. A LoRA is a smaller learned change applied to a compatible base and cannot generate alone. Match its family and documented trigger, then prove its effect with the same seed.

What is the difference between a sampler and scheduler in Forge?

The sampler is the numerical denoising method. The scheduler distributes noise levels or timesteps across its steps. Forge exposes them separately, so a reproducible record includes both.

Why are CFG Scale and Distilled CFG Scale separate?

Standard CFG compares positive and unconditional conditioning. Distilled guidance is a separate value exposed for models trained to accept it. In current Forge, CFG 1 skips unconditional conditioning, so the negative prompt is ignored in that standard path.

What does GPU Weights mean in Forge?

GPU Weights is a megabyte target for supported model weights kept on the GPU during swapping. It is not total VRAM use. The run still needs inference headroom for active computation, image size, VAE decode, and other allocations.

Should GPU Weights equal my graphics card VRAM?

No. Using the full advertised VRAM as a weight target can leave no room for inference and cause fallback, severe slowdown, or an out-of-memory failure. Start from the memory guide’s conservative baseline and change one variable at a time.

What is the difference between Batch size and Batch count?

Batch size is how many images are processed together and usually raises peak memory more directly. Batch count is how many batches run sequentially. Total requested images are batch size multiplied by batch count.

Why does the negative prompt do nothing at CFG 1?

Current Forge processing skips the unconditional-conditioning branch when CFG Scale is 1. Because the negative prompt belongs to that branch, it has no effect in that mode. This does not mean every model should use a higher CFG.

What is the difference between a VAE and text encoder?

A text encoder turns prompt text into conditioning. A VAE converts between image pixels and the latent representation. They perform different jobs and neither replaces the base diffusion model.

What is the difference between a ControlNet preprocessor and model?

The preprocessor derives a map such as pose, edges, or depth from an image. The ControlNet model applies that map during diffusion. The map type, control model, and base-model family must agree.

Does Pixel Perfect make ControlNet compatible automatically?

No. Pixel Perfect calculates preprocessor resolution from geometry. It does not validate the control-model family, map type, active unit, control weight, or detected-map quality.

What does denoising strength change in img2img?

It controls how much of the denoising process is used from the starting image latent. Lower values usually preserve more source structure; higher values permit more change. Compare a short ladder with one fixed seed instead of treating a number as universal.

What is the difference between Hires. fix and an upscaler?

Hires. fix is a two-stage txt2img workflow with an upscale between diffusion passes. A post-process upscaler enlarges an existing image without necessarily running another diffusion pass. The correct route depends on whether you need regenerated detail or only a larger file.

Can PNG Info reproduce an image exactly?

It can recover embedded generation parameters when metadata is present, but exact reproduction can still require the same Forge commit, model files, hashes, extensions, hardware-sensitive paths, and defaults. Metadata may also be stripped after export.

Is Never OOM still a current Forge setting?

Never OOM is documented in a February 2024 announcement. Treat it as historical terminology rather than a guarantee about the inspected interface. Current memory diagnosis should use the actual controls and console output in your exact commit.

04 · EVIDENCE

Definitions trace to current code; user language traces to complete sources.

Current UI labels and behavior were checked against commit dfdcbab. Announcements and tutorials remain dated snapshots where they describe a particular interface generation.

Primary / direct sources

Local research reviewed: 9 grouped records, including 5 full video transcripts
O01 + O23

Original README and UI construction code

Product boundary and exact current labels.

O10–O11

Backend clarification and Never OOM announcement

Architecture and historical memory language.

O13–O17

Canvas, LayerDiffuse, FLUX, LoRA, and GGUF owner guides

Feature vocabulary and dated model-layout changes.

O20 + A03–A05

Extension record and secondary code maps

Intent discovery; consequential definitions rechecked in primary code.

VERIFIEDCurrent code

Exact labels, separate controls, paths, and processing relationships.

STALE SNAPSHOTDated owner guidance

Never OOM, Canvas, LayerDiffuse, FLUX loading, and swap recommendations.

UNKNOWNNot generalized

Arbitrary model compatibility, speed gains, extension support, and future maintenance.

Author Forge Field Guide editorial teamTechnical review Original Forge source at dfdcbabUpdated 31 Aug 2026