nicolabs/Open-Sora
1
0
Fork 0
mirror of https://github.com/hpcaitech/Open-Sora.git synced 2026-04-10 21:01:26 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

fix typo

Browse source
This commit is contained in:
gxyes 2025-03-11 11:32:58 +08:00
parent 2ca632a2d1
commit bbc64bcb72
10 changed files with 17 additions and 23 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
configs/opensora-v1-3/inference/t2v.py
View file

@ -12,7 +12,7 @@ dtype = "bf16"
model = dict( model = dict(
type="STDiT3-XL/2", type="STDiT3-XL/2",
from_pretrained="/home/guoxinying/open_source_video_ocean_V1/OpenSora-STDiT-v4", from_pretrained="hpcaitech/OpenSora-STDiT-v4",
qk_norm=True, qk_norm=True,
enable_flash_attn=True, enable_flash_attn=True,
enable_layernorm_kernel=True, enable_layernorm_kernel=True,
@ -22,7 +22,7 @@ model = dict(
) )
vae = dict( vae = dict(
type="OpenSoraVAE_V1_3", type="OpenSoraVAE_V1_3",
from_pretrained="/home/guoxinying/open_source_video_ocean_V1/OpenSora-VAE-v1.3", from_pretrained="hpcaitech/OpenSora-VAE-v1.3",
z_channels=16, z_channels=16,
micro_batch_size=1, micro_batch_size=1,
micro_batch_size_2d=4, micro_batch_size_2d=4,
@ -35,7 +35,7 @@ vae = dict(
) )
text_encoder = dict( text_encoder = dict(
type="t5", type="t5",
from_pretrained="/mnt/jfs-hdd/sora/checkpoints/pretrained_models/t5-v1_1-xxl", from_pretrained="DeepFloyd/t5-v1_1-xxl",
model_max_length=300, model_max_length=300,
) )
scheduler = dict( scheduler = dict(

6
configs/opensora-v1-3/inference/v2v.py
View file

@ -18,7 +18,7 @@ use_oscillation_guidance_for_image = True
model = dict( model = dict(
type="STDiT3-XL/2", type="STDiT3-XL/2",
from_pretrained="/home/guoxinying/open_source_video_ocean_V1/OpenSora-STDiT-v4", from_pretrained="hpcaitech/OpenSora-STDiT-v4",
qk_norm=True, qk_norm=True,
enable_flash_attn=True, enable_flash_attn=True,
enable_layernorm_kernel=True, enable_layernorm_kernel=True,
@ -29,7 +29,7 @@ model = dict(
) )
vae = dict( vae = dict(
type="OpenSoraVAE_V1_3", type="OpenSoraVAE_V1_3",
from_pretrained="/home/guoxinying/open_source_video_ocean_V1/OpenSora-VAE-v1.3", from_pretrained="hpcaitech/OpenSora-VAE-v1.3",
z_channels=16, z_channels=16,
micro_batch_size=1, micro_batch_size=1,
micro_batch_size_2d=4, micro_batch_size_2d=4,
@ -42,7 +42,7 @@ vae = dict(
) )
text_encoder = dict( text_encoder = dict(
type="t5", type="t5",
from_pretrained="/mnt/jfs-hdd/sora/checkpoints/pretrained_models/t5-v1_1-xxl", from_pretrained="DeepFloyd/t5-v1_1-xxl",
model_max_length=300, model_max_length=300,
) )
scheduler = dict( scheduler = dict(

2
configs/opensora-v1-3/misc/eval_loss.py
View file

@ -25,7 +25,7 @@ model = dict(
) )
vae = dict( vae = dict(
type="OpenSoraVAE_V1_3", type="OpenSoraVAE_V1_3",
from_pretrained="/home/guoxinying/open_source_video_ocean_V1/OpenSora-VAE-v1.3", from_pretrained="hpcaitech/OpenSora-VAE-v1.3",
z_channels=16, z_channels=16,
micro_batch_size=1, micro_batch_size=1,
micro_batch_size_2d=4, micro_batch_size_2d=4,

10
configs/opensora-v1-3/train/adapt.py
View file

@ -26,22 +26,16 @@ plugin = "zero2"
# Model settings # Model settings
model = dict( model = dict(
type="STDiT3-XL/2", type="STDiT3-XL/2",
from_pretrained="outputs/0373-STDiT3-XL-2/epoch3-global_step36000/ema.pt", from_pretrained=None,
qk_norm=True, qk_norm=True,
enable_flash_attn=True, enable_flash_attn=True,
enable_layernorm_kernel=True, enable_layernorm_kernel=True,
adapt_16ch=True, adapt_16ch=True,
skip_temporal=True, skip_temporal=True,
) )
# vae = dict(
# type="VideoAutoencoderKL",
# from_pretrained="PixArt-alpha/pixart_sigma_sdxlvae_T5_diffusers",
# subfolder="vae",
# scaling_factor=0.13025,
# )
vae = dict( vae = dict(
type="OpenSoraVAE_V1_3", type="OpenSoraVAE_V1_3",
from_pretrained="/home/guoxinying/open_source_video_ocean_V1/OpenSora-VAE-v1.3", from_pretrained="hpcaitech/OpenSora-VAE-v1.3",
z_channels=16, z_channels=16,
micro_batch_size=1, micro_batch_size=1,
micro_batch_size_2d=4, micro_batch_size_2d=4,

4
configs/opensora-v1-3/train/adapt_i2v.py
View file

@ -63,7 +63,7 @@ model = dict(
) )
vae = dict( vae = dict(
type="OpenSoraVAE_V1_3", type="OpenSoraVAE_V1_3",
from_pretrained="/home/guoxinying/open_source_video_ocean_V1/OpenSora-VAE-v1.3", from_pretrained="hpcaitech/OpenSora-VAE-v1.3",
z_channels=16, z_channels=16,
micro_batch_size=1, micro_batch_size=1,
micro_batch_size_2d=4, micro_batch_size_2d=4,
@ -76,7 +76,7 @@ vae = dict(
) )
text_encoder = dict( text_encoder = dict(
type="t5", type="t5",
from_pretrained="pretrained_models/t5-v1_1-xxl", from_pretrained="DeepFloyd/t5-v1_1-xxl",
model_max_length=300, model_max_length=300,
shardformer=True, shardformer=True,
) )

2
configs/opensora-v1-3/train/stage1.py
View file

@ -33,7 +33,7 @@ model = dict(
) )
vae = dict( vae = dict(
type="OpenSoraVAE_V1_3", type="OpenSoraVAE_V1_3",
from_pretrained="/home/guoxinying/open_source_video_ocean_V1/OpenSora-VAE-v1.3", from_pretrained="hpcaitech/OpenSora-VAE-v1.3",
z_channels=16, z_channels=16,
micro_batch_size=1, micro_batch_size=1,
micro_batch_size_2d=4, micro_batch_size_2d=4,

2
configs/opensora-v1-3/train/stage1_hq.py
View file

@ -53,7 +53,7 @@ model = dict(
) )
vae = dict( vae = dict(
type="OpenSoraVAE_V1_3", type="OpenSoraVAE_V1_3",
from_pretrained="/home/guoxinying/open_source_video_ocean_V1/OpenSora-VAE-v1.3", from_pretrained="hpcaitech/OpenSora-VAE-v1.3",
z_channels=16, z_channels=16,
micro_batch_size=1, micro_batch_size=1,
micro_batch_size_2d=4, micro_batch_size_2d=4,

4
configs/opensora-v1-3/train/stage1_nk.py
View file

@ -24,14 +24,14 @@ plugin = "zero2"
# Model settings # Model settings
model = dict( model = dict(
type="STDiT3-XL/2", type="STDiT3-XL/2",
from_pretrained="outputs/0461-STDiT3-XL-2/epoch3-global_step40000/ema.pt", from_pretrained=None,
qk_norm=True, qk_norm=True,
enable_flash_attn=True, enable_flash_attn=True,
enable_layernorm_kernel=True, enable_layernorm_kernel=True,
) )
vae = dict( vae = dict(
type="OpenSoraVAE_V1_3", type="OpenSoraVAE_V1_3",
from_pretrained="/home/guoxinying/open_source_video_ocean_V1/OpenSora-VAE-v1.3", from_pretrained="hpcaitech/OpenSora-VAE-v1.3",
z_channels=16, z_channels=16,
micro_batch_size=1, micro_batch_size=1,
micro_batch_size_2d=4, micro_batch_size_2d=4,

2
configs/opensora-v1-3/train/stage1_noise.py
View file

@ -33,7 +33,7 @@ model = dict(
) )
vae = dict( vae = dict(
type="OpenSoraVAE_V1_3", type="OpenSoraVAE_V1_3",
from_pretrained="/home/guoxinying/open_source_video_ocean_V1/OpenSora-VAE-v1.3", from_pretrained="hpcaitech/OpenSora-VAE-v1.3",
z_channels=16, z_channels=16,
micro_batch_size=1, micro_batch_size=1,
micro_batch_size_2d=4, micro_batch_size_2d=4,

2
configs/opensora-v1-3/train/stage2.py
View file

@ -33,7 +33,7 @@ model = dict(
) )
vae = dict( vae = dict(
type="OpenSoraVAE_V1_3", type="OpenSoraVAE_V1_3",
from_pretrained="/home/guoxinying/open_source_video_ocean_V1/OpenSora-VAE-v1.3", from_pretrained="hpcaitech/OpenSora-VAE-v1.3",
z_channels=16, z_channels=16,
micro_batch_size=1, micro_batch_size=1,
micro_batch_size_2d=4, micro_batch_size_2d=4,