tags:
- text-to-image
- sd3
- lora
- diffusers
- template:sd-lora
- StableDiffusion3
- StableDiffusion3.5
- ai-toolkit
widget:
- text: >-
/@step 4000 weights:/ HST style autochrome photo of a drugged God
Christ-insect in wooden pillbox being pinned needle-poked by a giant
Marina Tsvetaeva in a long red gown, arcane symbolism, colorful HD
realistic film in Yury Annenkov & David Lynch styles, detailed face
worrying, full-height red-robed woman roams European city circa 1930
output:
url: samples/1729711624663__000004000_4.jpg
- text: >-
/@step 1600 weights:/HST style autochrome photo of a drugged God
Christ-insect in wooden pillbox being pinned needle-poked by a giant
Marina Tsvetaeva in a long red gown, arcane symbolism, colorful HD
realistic film in Yury Annenkov & David Lynch styles, detailed face
worrying, full-height red-robed woman roams European city circa 1930
output:
url: samples/1729700659095__000001600_4.jpg
- text: >-
/@step 1800 weights:/HST style autochrome photo of a drugged God
Christ-insect in wooden pillbox being pinned needle-poked by a giant
Marina Tsvetaeva in a long red gown, arcane symbolism, colorful HD
realistic film in Yury Annenkov & David Lynch styles, detailed face
worrying, full-height red-robed woman roams European city circa 1930
output:
url: samples/1729701571749__000001800_4.jpg
- text: >-
/@step 800 weights:/HHST style autochrome photo of a drugged God
Christ-insect in wooden pillbox being pinned needle-poked by a giant
Marina Tsvetaeva in a long red gown, arcane symbolism, colorful HD
realistic film in Yury Annenkov & David Lynch styles, detailed face
worrying, full-height red-robed woman roams European city circa 1930
output:
url: samples/1729696985671__000000800_4.jpg
- text: >-
/@step 1800 weights:/HST style communist poster with text "JOIN RCA!",
over autochrome color photo of Vladimir Lenin at a Dada cabaret in 1916
Zurich, dancing with red feathered drunken dinosaur, an early conceptual
artist. Lenin is full of contageous awe, his blemished skin flushing with
anxious excitement, his famous bald spot sweatily glistening under warm
lights. In the back, Krupskaya and Inessa Armand laugh.
output:
url: samples/1729701488064__000001800_0.jpg
- text: >-
/@step 4000 weights:/HHST style communist poster with text "JOIN RCA!",
over autochrome color photo of Vladimir Lenin at a Dada cabaret in 1916
Zurich, dancing with red feathered drunken dinosaur, an early conceptual
artist. Lenin is full of contageous awe, his blemished skin flushing with
anxious excitement, his famous bald spot sweatily glistening under warm
lights. In the back, Krupskaya and Inessa Armand laugh.
output:
url: samples/1729711540947__000004000_0.jpg
- text: >-
/@step 1600 weights:/ HST style communist poster with text "JOIN RCA!",
over autochrome color photo of Vladimir Lenin at a Dada cabaret in 1916
Zurich, dancing with red feathered drunken dinosaur, an early conceptual
artist. Lenin is full of contageous awe, his blemished skin flushing with
anxious excitement, his famous bald spot sweatily glistening under warm
lights. In the back, Krupskaya and Inessa Armand laugh.
output:
url: samples/1729700575417__000001600_0.jpg
- text: >-
/@step 1600 weights:/ HST autochrome style analog dslr award-winning 8k
art photo showing a nurse battling a giant cell phone, above text caption
of \ONCE WE WERE ALL HARRY GARDNER!\, in a highly realistic modern
American medical hospital
output:
url: samples/1729700596338__000001600_1.jpg
- text: >-
/@step 1000 weights:/ HST autochrome style analog dslr award-winning 8k
art photo showing a nurse battling a giant cell phone, above text caption
of \ONCE WE WERE ALL HARRY GARDNER!\, in a highly realistic modern
American medical hospital
output:
url: samples/1729697845170__000001000_1.jpg
- text: >-
/@step 3600 weights:/ HST style autochrome photograph of a dark CIA agent
Koala leaping at the throat of goth freedom-fighter Julian Assange, tragic
atmosphere, award-winning art photo
output:
url: samples/1729709764758__000003600_2.jpg
- text: >-
/@step 1200 weights:/ HST style autochrome photograph of a dark CIA agent
Koala leaping at the throat of goth freedom-fighter Julian Assange, tragic
atmosphere, award-winning art photo
output:
url: samples/1729698783304__000001200_2.jpg
- text: >-
/@step 1000 weights:/ HST style autochrome photograph of a dark CIA agent
Koala leaping at the throat of goth freedom-fighter Julian Assange, tragic
atmosphere, award-winning art photo
output:
url: samples/1729697866093__000001000_2.jpg
- text: >-
/@step 4000 weights:/ HST style autochrome photo of realistic green-eyed
black cat, with prominent regions of white fur, playing a piano and
singing, amateur 2004 photograph shot on a cell phone in a Los Angeles
apartment kitchen
output:
url: samples/1729711603742__000004000_3.jpg
- text: >-
/@step 3000 weights:/ HST style autochrome photo of realistic green-eyed
black cat, with prominent regions of white fur, playing a piano and
singing, amateur 2004 photograph shot on a cell phone in a Los Angeles
apartment kitchen
output:
url: samples/1729707029633__000003000_3.jpg
- text: >-
/@step 1800 weights:/ HST style autochrome photo of realistic green-eyed
black cat, with prominent regions of white fur, playing a piano and
singing, amateur 2004 photograph shot on a cell phone in a Los Angeles
apartment kitchen
output:
url: samples/1729701550832__000001800_3.jpg
- text: >-
/@step 1000 weights:/ HST style autochrome photo of realistic green-eyed
black cat, with prominent regions of white fur, playing a piano and
singing, amateur 2004 photograph shot on a cell phone in a Los Angeles
apartment kitchen
output:
url: samples/1729697887015__000001000_3.jpg
base_model: stabilityai/stable-diffusion-3.5-large
license: creativeml-openrail-m
language:
- en
pipeline_tag: text-to-image
library_name: diffusers
HSTsd3ii
Model trained with AI Toolkit by Ostris
Trigger words
'HST style autochrome photo'
Config Parameters
Dim:16 Alpha:32 Optimizer:Ademamix8bit LR:2e-4 More info below!
Fine-tuned using the Google Colab Notebook* of ai-toolkit.
I've used A100 via Colab Pro.
However, training SD3.5 may potentially work with Free Colab or lower VRAM in general:
Especially if one were to use:
...Say, lower rank (try 4 or 8), dataset size (in terms of caching/bucketing/pre-loading impacts), 1 batch size, Adamw8bit optimizer, 512 resolution, maybe adding the /lowvram, true/ argument, and plausibly specifying alternate quantization variants.
Generally, VRAM expenditures for fine-tuning SD3.5 tend to be lower than for Flux during training.
So, try it!
To use on Colab, modify a Flux template Notebook from here with parameters from Ostris' example config for SD3.5 here!
job: extension
config:
name: HSTsd3ii
process:
- type: sd_trainer
training_folder: /content/drive/MyDrive/HSTsd3ii
performance_log_every: 600
device: cuda:0
network:
type: lora
linear: 16
linear_alpha: 32
save:
dtype: float16
save_every: 250
push_to_hub: true
hf_repo_id: AlekseyCalvin/HSTsd3iii
hf_private: true
max_step_saves_to_keep: 16
datasets:
- folder_path: /content/dataset
caption_ext: txt
caption_dropout_rate: 0.0
shuffle_tokens: false
cache_latents_to_disk: true
resolution:
- 1024
train:
batch_size: 4
steps: 4000
gradient_accumulation_steps: 1
train_unet: true
train_text_encoder: false
gradient_checkpointing: true
noise_scheduler: flowmatch
timestep_type: linear
optimizer: ademamix8bit
lr: 0.0002
skip_first_sample: true
ema_config:
use_ema: true
ema_decay: 0.8
dtype: bf16
model:
name_or_path: stabilityai/stable-diffusion-3.5-large
is_v3: true
quantize: true
Plus validation settings:
Prompts like the above, at 1024, guidance scale 4, 25 steps, seed 42, no negatives.
Download model and use it with ComfyUI, AUTOMATIC1111, SD.Next, Invoke AI, etc.
Weights for this model are available in Safetensors format.
Download them in the Files & versions tab.
Use it with the 🧨 diffusers library
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('stabilityai/stable-diffusion-3.5-large', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('AlekseyCalvin/HSTsd3iii', weight_name='HSTsd3ii.safetensors')
image = pipeline('HST style communist poster with text "JOIN RCA!", over autochrome color photo of Vladimir Lenin at a Dada cabaret in 1916 Zurich, dancing with red feathered drunken dinosaur, an early conceptual artist. Lenin is full of contageous awe, his blemished skin flushing with anxious excitement, his famous bald spot sweatily glistening under warm lights. In the back, Krupskaya and Inessa Armand laugh. ').images[0]
image.save("my_image.png")
For more details, including weighting, merging and fusing LoRAs, check the documentation on loading LoRAs in diffusers