Early 2000's Rebecca Gayheart.
Trained using 147, 512x512, 48 frames@24fps videos and 16, 1024px images on a 4090.
config.toml:
# This configuration should allow you to train Wan 14b t2v on 512x512x81 sized videos (or varying aspect ratios of the same size), with 24GB VRAM.
# change this
output_dir = '/mnt/d/Projects/video-training/gayheart/output'
# and this
dataset = '/mnt/d/Projects/video-training/gayheart/dataset.toml'
# training settings
epochs = 100
micro_batch_size_per_gpu = 1
pipeline_stages = 1
gradient_accumulation_steps = 1
gradient_clipping = 1
warmup_steps = 10
# eval settings
eval_every_n_epochs = 1
eval_before_first_step = true
eval_micro_batch_size_per_gpu = 1
eval_gradient_accumulation_steps = 1
# misc settings
save_every_n_epochs = 5
#checkpoint_every_n_minutes = 30
checkpoint_every_n_epochs = 5
activation_checkpointing = 'unsloth'
partition_method = 'parameters'
save_dtype = 'bfloat16'
caching_batch_size = 1
steps_per_print = 1
video_clip_mode = 'single_beginning'
blocks_to_swap = 24
[model]
type = 'wan'
ckpt_path = '/mnt/d/software_tools/diffusion-pipe/models/wan/Wan2.1-T2V-14B'
transformer_path = '/mnt/d/software_tools/diffusion-pipe/models/wan/Wan2_1-T2V-14B_fp8_e5m2.safetensors' #kijai
vae_path = '/mnt/d/software_tools/diffusion-pipe/models/wan/Wan_2_1_VAE_bf16.safetensors' #kijai
llm_path = '/mnt/d/software_tools/diffusion-pipe/models/wan/umt5-xxl-enc-bf16.safetensors'
dtype = 'bfloat16'
transformer_dtype = 'float8'
timestep_sample_method = 'logit_normal'
[adapter]
type = 'lora'
rank = 32
dtype = 'bfloat16'
[optimizer]
type = 'AdamW8bitKahan'
lr = 2e-5
betas = [0.9, 0.99]
weight_decay = 0.01
stabilize = false
dataset.toml
# Resolution settings.
# Can adjust this to 1024 for image training, especially on 24gb cards.
resolutions = [512]
#Aspect ratio bucketing settings
enable_ar_bucket = true
min_ar = 0.5
max_ar = 2.0
num_ar_buckets = 7
# Frame buckets (1 is for images)
frame_buckets = [1, 33]
[[directory]]
# Set this to where your dataset is
path = '/mnt/d/Projects/video-training/gayheart/512px-clips/'
# Reduce as necessary
num_repeats = 1
[[directory]]
path = '/mnt/d/Projects/video-training/gayheart/1024px-images/'
resolutions=[1024]
# Frame buckets (1 is for images)
frame_buckets = [1]
# Reduce as necessary
num_repeats = 5