frankleeeee's picture
updated to v1.1
5613724
raw
history blame
908 Bytes
# Define dataset
dataset = dict(
type="VideoTextDataset",
data_path=None,
num_frames=16,
frame_interval=3,
image_size=(256, 256),
)
# Define acceleration
num_workers = 4
dtype = "bf16"
grad_checkpoint = True
plugin = "zero2"
sp_size = 1
# Define model
model = dict(
type="STDiT-XL/2",
space_scale=0.5,
time_scale=1.0,
from_pretrained="PixArt-XL-2-512x512.pth",
enable_flashattn=True,
enable_layernorm_kernel=True,
)
vae = dict(
type="VideoAutoencoderKL",
from_pretrained="stabilityai/sd-vae-ft-ema",
)
text_encoder = dict(
type="t5",
from_pretrained="DeepFloyd/t5-v1_1-xxl",
model_max_length=120,
shardformer=True,
)
scheduler = dict(
type="iddpm",
timestep_respacing="",
)
# Others
seed = 42
outputs = "outputs"
wandb = False
epochs = 1000
log_every = 10
ckpt_every = 1000
load = None
batch_size = 8
lr = 2e-5
grad_clip = 1.0