Skip to content

Models

123 checkpoints in 53 families, all behind pipeline("t2i", backend=<model id>, ...). Each family page has the architecture notes, settings, license and the recorded parity result.

cosmos-predict2-2b-t2i

Cosmos-Predict2

Text→ImageParity · CPU fp32

NVIDIA Open Model License

joyai-image-edit

JoyAI-Image

TextImage→ImageParity · CPU fp32

Apache-2.0

joyai-image-edit-plus

JoyAI-Image

TextImage→ImageParity · CPU fp32

Apache-2.0

ldm-text2im-large-256

Latent Diffusion Model

Text→ImageParity · CPU fp32

Apache-2.0

dreamlite-base

DreamLite

Text→ImageParity · CPU fp32

CC-BY-NC-4.0

dreamlite-mobile

DreamLite

Text→ImageParity · CPU fp32

CC-BY-NC-4.0

sd-v1-1

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

sd-v1-2

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

sd-v1-3

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

sd-v1-4

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

sd-v1-5

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

dreamshaper-8

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

sdxs-512-0.9

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

segmind-small-sd

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

segmind-tiny-sd

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

bk-sdm-base

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

bk-sdm-small

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

bk-sdm-tiny

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

bk-sdm-v2-base

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

bk-sdm-v2-small

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

bk-sdm-v2-tiny

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

lcm-dreamshaper-v7

Stable Diffusion

Text→ImageParity · CPU fp32

MIT

sd-v2-1

Stable Diffusion

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

sd-turbo

Stable Diffusion

Text→ImageParity · CPU fp32

sai-nc-community

sdxl-base-1.0

Stable Diffusion XL

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

sdxl-lightning-2step

Stable Diffusion XL

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

sdxl-lightning-4step

Stable Diffusion XL

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

sdxl-lightning-8step

Stable Diffusion XL

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

hyper-sdxl-1step

Stable Diffusion XL

Text→ImageParity · CPU fp32

Unspecified

dmd2-sdxl-4step

Stable Diffusion XL

Text→ImageParity · CPU fp32

CC-BY-NC-4.0

dmd2-sdxl-1step

Stable Diffusion XL

Text→ImageParity · CPU fp32

CC-BY-NC-4.0

juggernaut-xl-v9

Stable Diffusion XL

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL-M

animagine-xl-4.0

Stable Diffusion XL

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

dreamshaper-xl-1.0

Stable Diffusion XL

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

sdxl-refiner-1.0

Stable Diffusion XL

TextImage→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

sdxl-turbo

Stable Diffusion XL

Text→ImageParity · CPU fp32

sai-nc-community

ssd-1b

Stable Diffusion XL

Text→ImageParity · CPU fp32

Apache-2.0

segmind-vega

Stable Diffusion XL

Text→ImageParity · CPU fp32

Apache-2.0

koala-700m

Stable Diffusion XL

Text→ImageParity · CPU fp32

Unspecified

koala-1b

Stable Diffusion XL

Text→ImageParity · CPU fp32

Unspecified

koala-lightning-700m

Stable Diffusion XL

Text→ImageParity · CPU fp32

Unspecified

koala-lightning-1b

Stable Diffusion XL

Text→ImageParity · CPU fp32

Unspecified

playground-v2

Stable Diffusion XL

Text→ImageParity · CPU fp32

playground-v2-community

playground-v2.5

Stable Diffusion XL

Text→ImageParity · CPU fp32

playground-v2dot5-community

sd3-medium

Stable Diffusion 3 and 3.5

Text→ImageParity · CPU fp32

Stability AI Non-Commercial Research Community License

sd3.5-medium

Stable Diffusion 3 and 3.5

Text→ImageParity · CPU fp32

Stability AI Community License

sd3.5-large

Stable Diffusion 3 and 3.5

Text→ImageParity · CPU fp32

Stability AI Community License

sd3.5-large-turbo

Stable Diffusion 3 and 3.5

Text→ImageParity · CPU fp32

Stability AI Community License

flux1-schnell

FLUX.1

Text→ImageParity · CPU fp32

Apache-2.0

flux1-dev

FLUX.1

Text→ImageParity · CPU fp32

FLUX.1-dev Non-Commercial License

flux1-krea-dev

FLUX.1

Text→ImageParity · CPU fp32

FLUX.1-dev Non-Commercial License

flux-1-lite-8b

FLUX.1

Text→ImageParity · CPU fp32

FLUX.1-dev Non-Commercial License

flux1-kontext-dev

FLUX.1

TextImage→ImageParity · CPU fp32

FLUX.1-dev Non-Commercial License

visualcloze-384

VisualCloze

TextImage→ImageParity · CPU fp32

Apache-2.0

chroma1-hd

Chroma

Text→ImageParity · CPU fp32

Apache-2.0

chroma1-flash

Chroma

Text→ImageParity · CPU fp32

Apache-2.0

chroma1-base

Chroma

Text→ImageParity · CPU fp32

Apache-2.0

flux2-klein-4b

FLUX.2 klein

Text→ImageParity · CPU fp32

Apache-2.0

flux2-klein-9b

FLUX.2 klein

Text→ImageParity · CPU fp32

FLUX Non-Commercial License

hunyuan-dit-v1.1-distilled

HunyuanDiT

Text→ImageParity · CPU fp32

Tencent Hunyuan Community License

hunyuan-dit-v1.2

HunyuanDiT

Text→ImageParity · CPU fp32

Tencent Hunyuan Community License

hunyuan-dit-v1.2-distilled

HunyuanDiT

Text→ImageParity · CPU fp32

Tencent Hunyuan Community License

deepfloyd-if-i-xl

DeepFloyd IF

Text→ImageParity · CPU fp32

DeepFloyd IF License (non-commercial)

kandinsky-2-1

Kandinsky 2.1

Text→ImageParity · CPU fp32

Apache-2.0

kandinsky-2.2

Kandinsky 2.2

Text→ImageParity · CPU fp32

Apache-2.0

kandinsky-3

Kandinsky 3

Text→ImageParity · CPU fp32

Apache-2.0

kandinsky-5-t2i-lite

Kandinsky 5

Text→ImageParity · CPU fp32

MIT

show-o-512

Show-o

Text→ImageParity · CPU fp32

Apache-2.0

pixeldit

PixelDiT

Text→ImageParity · CPU fp32

NVIDIA License (NSCLv1; non-commercial research or evaluation)

ovis-image-7b

Ovis-Image

Text→ImageParity · CPU fp32

Apache-2.0

prx-1024-t2i-beta

PRX

Text→ImageParity · CPU fp32

Apache-2.0

prx-512-t2i-sft

PRX

Text→ImageParity · CPU fp32

Apache-2.0

stable-cascade

Stable Cascade

Text→ImageParity · CPU fp32

stable-cascade-nc-community

stable-cascade-lite

Stable Cascade

Text→ImageParity · CPU fp32

stable-cascade-nc-community

wuerstchen-v2

Würstchen v2

Text→ImageParity · CPU fp32

mit

glm-image

GLM-Image

Text→ImageParity · CPU fp32

MIT

pixart-sigma-xl-1024

PixArt

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

pixart-sigma-xl-512

PixArt

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

pixart-sigma-xl-256

PixArt

Text→ImageParity · CPU fp32

OpenRAIL (exact terms unspecified)

pixart-sigma-xl-2k

PixArt

Text→ImageParity · CPU fp32

OpenRAIL (exact terms unspecified)

pixart-alpha-xl-1024

PixArt

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

pixart-alpha-xl-512

PixArt

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

pixart-lcm-xl-1024

PixArt

Text→ImageParity · CPU fp32

CreativeML-OpenRAIL++-M

nucleus-image

Nucleus-Image

Text→ImageParity · CPU fp32

Apache-2.0

krea2-turbo

Krea 2

Text→ImageParity · CPU fp32

Krea 2 Community License

ernie-image

ERNIE-Image

Text→ImageParity · CPU fp32

Apache-2.0

longcat-image

LongCat-Image

Text→ImageParity · CPU fp32

Apache-2.0

cogview3-plus-3b

CogView3-Plus

Text→ImageParity · CPU fp32

Apache-2.0

hunyuanimage-2-1

HunyuanImage 2.1

Text→ImageParity · CPU fp32

Tencent Hunyuan Community License

cogview4-6b

CogView4

Text→ImageParity · CPU fp32

Apache-2.0

ideogram4-fp8

Ideogram 4

Text→ImageParity · CPU fp32

Ideogram 4 Non-Commercial License

qwen-image

Qwen-Image

Text→ImageParity · CPU fp32

Apache-2.0

qwen-image-2.1

Qwen-Image

Text→ImageParity · CPU fp32

Qwen Research License

qwen-image-edit

Qwen-Image

TextImage→ImageParity · CPU fp32

Apache-2.0

qwen-image-edit-2509

Qwen-Image

TextImage→ImageParity · CPU fp32

Apache-2.0

janus-pro-1b

Janus-Pro

Text→ImageParity · CPU fp32

DeepSeek Model License

janus-pro-7b

Janus-Pro

Text→ImageParity · CPU fp32

DeepSeek Model License

bagel-7b-mot

BAGEL

Text→ImageParity · CPU fp32

Apache-2.0

z-image

Z-Image

Text→ImageParity · CPU fp32

Apache-2.0

z-image-turbo

Z-Image

Text→ImageParity · CPU fp32

Apache-2.0

sana-1600m-1024

Sana

Text→ImageParity · CPU fp32

Apache-2.0

sana-600m-1024

Sana

Text→ImageParity · CPU fp32

Apache-2.0

sana-sprint-1.6b-1024

Sana

Text→ImageParity · CPU fp32

Apache-2.0

sana-1.5-4.8b-1024

Sana

Text→ImageParity · CPU fp32

Apache-2.0

kolors

Kolors

Text→ImageParity · CPU fp32

Apache-2.0

lumina-image-2.0

Lumina-Image 2.0

Text→ImageParity · CPU fp32

Apache-2.0

lumina-next

Lumina-Next

Text→ImageParity · CPU fp32

Apache-2.0

lumina-dimoo

Lumina-DiMOO

Text→ImageParity · CPU fp32

Apache-2.0

amused-256

aMUSEd

Text→ImageParity · CPU fp32

OpenRAIL++

amused-512

aMUSEd

Text→ImageParity · CPU fp32

OpenRAIL++

meissonic

Meissonic

Text→ImageParity · CPU fp32

Apache-2.0

auraflow-v0.3

AuraFlow

Text→ImageParity · CPU fp32

Apache-2.0

llamagen-t2i-xl-256

LlamaGen

Text→ImageParity · CPU fp32

MIT

llamagen-t2i-xl-512

LlamaGen

Text→ImageParity · CPU fp32

MIT

infinity-2b

Infinity

Text→ImageParity · CPU fp32

MIT

hart

HART

Text→ImageParity · CPU fp32

MIT

hidream-i1-dev

HiDream-I1

Text→ImageParity · CPU fp32

MIT

hidream-i1-fast

HiDream-I1

Text→ImageParity · CPU fp32

MIT

hidream-i1-full

HiDream-I1

Text→ImageParity · CPU fp32

MIT

hidream-o1-image

HiDream-O1

Text→ImageParity · CPU fp32

MIT

omnigen-v1

OmniGen

TextImage→ImageParity · CPU fp32

MIT

emu3-gen

Emu3-Gen

Text→ImageParity · CPU fp32

Apache-2.0

nextstep-1-large

NextStep-1

Text→ImageParity · CPU fp32

Apache-2.0

Image input

Most models take a text prompt only. The models in this table also accept an input image through the image option (a PIL.Image.Image or an imagehub.Image); every model not listed here takes no image input, and passing image to one raises an error.

Model Image input Use Options
flux1-kontext-dev optional instruction editing image
visualcloze-384 required grid visual in-context generation image
joyai-image-edit optional instruction editing image
joyai-image-edit-plus required multi-image instruction editing image
qwen-image-edit required instruction editing image
qwen-image-edit-2509 required multi-image instruction editing image
sdxl-refiner-1.0 required refinement (img2img) image, strength
omnigen-v1 optional editing and subject-driven generation, referenced from the prompt image, image_guidance, max_input_image_size, use_input_image_size_as_output

LoRA adapters

LoRA adapter files load on top of a checkpoint in the Stable Diffusion 1.x and 2.x, SDXL, SD3, FLUX.1, FLUX.2 klein, Sana, Qwen-Image, Z-Image, Lumina 2, HiDream-I1, Krea 2 and Ideogram 4 backends, including the fine-tunes they serve. The U-Net backends patch the U-Net and the CLIP text encoders; the transformer backends patch the transformer.

from imagehub import pipeline

with pipeline("t2i", backend="sd-v1-5", model="checkpoints/sd-v1-5",
              lora="style.safetensors", lora_scale=0.8) as pipe:
    result = pipe("a watercolor fox", seed=0)

pipeline("t2i", backend="sdxl-base-1.0", model="checkpoints/sdxl-base-1.0",
         lora=[("style.safetensors", 0.8), ("character.safetensors", 0.5)])
pipeline("t2i", backend="sd-v1-5", model="checkpoints/sd-v1-5",
         lora="style.safetensors", lora_fuse=True)

The second call stacks two adapters with their own scales; the third merges the adapter into the weights after loading.

Option Meaning
lora a .safetensors path, or a list of paths and (path, scale) pairs
lora_scale the scale of a single path (default 1.0); use pairs for a list
lora_fuse merge the adapters into the base weights after loading (default False)

Layouts. The diffusers / PEFT layout is read on every backend above. The U-Net backends also read the legacy diffusers and kohya layouts:

Layout Example key
diffusers / PEFT unet.down_blocks.0.attentions.0.transformer_blocks.0.attn1.to_q.lora_A.weight
legacy diffusers unet.down_blocks.0.attentions.0.transformer_blocks.0.attn1.processor.to_q_lora.down.weight
kohya lora_unet_down_blocks_0_attentions_0_proj_in.lora_down.weight (with .alpha)

kohya SDXL files that use the original U-Net block names (input_blocks, middle_block, output_blocks) are mapped to the diffusers names. Text-encoder keys are accepted with or without CLIP's text_model level. Adapters can patch linear and convolution layers. The alpha comes from the file (kohya .alpha tensors or the diffusers lora_adapter_metadata header) and equals the rank when the file has none.

Rejected files. DoRA and LyCORIS (LoHa, LoKr) adapters, keys that match no module, and modules without both weights raise an error before the model is changed. Options are checked when the pipeline is created, before any weights load.

Programmatic use. imagehub.lora holds the layers (LoRALinear, LoRAConv2d) and the inject_lora, merge_lora, unmerge_lora, set_lora_scale and remove_lora helpers; imagehub.lora_formats.load_lora_file loads a file into any set of native components.

Outputs match diffusers load_lora_weights exactly on CPU in float32; the comparisons are in the validation record.

Weights, licenses and pinned revisions