You can find detailed information about the versions in these articles:<ul><li><a target="_blank" rel="ugc" href="https://civarchive.com/articles/18479/hollow-knight-wan-22-ti2v-5b">[Wan 2.2 TI2V-5B]</a></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/articles/18481">[Wan 2.2 T2V-A14B]</a></li></ul>

Hollow knight [Wan 2.2]

Training Detailsdataset.toml<pre><code>resolutions = [[ 1280, 704]]
enable_ar_bucket = true
min_ar = 0.5
max_ar = 2.0
num_ar_buckets = 7
ar_buckets = [[1280, 704]]
frame_buckets = [ 1, 33]
[[directory]]
path = "/home/user/hollow_knight_dataset"
num_repeats = 5</code></pre>train.toml<pre><code>output_dir = "/home/user/hollow_knight/A14B"
dataset = "/home/user/dataset_A14B.toml"
epochs = 80
micro_batch_size_per_gpu = 1
pipeline_stages = 1
gradient_accumulation_steps = 1
gradient_clipping = 1
warmup_steps = 10
eval_every_n_epochs = 1
eval_before_first_step = true
eval_micro_batch_size_per_gpu = 1
eval_gradient_accumulation_steps = 1
save_every_n_epochs = 4
activation_checkpointing = "unsloth"
partition_method = "parameters"
save_dtype = "bfloat16"
caching_batch_size = 1
steps_per_print = 10
video_clip_mode = "single_beginning"
blocks_to_swap = 32

[model]
type = "wan"
ckpt_path = "/home/user/Wan2.2-T2V-A14B"
transformer_path = "/home/user/Wan2.2-T2V-A14B/low_noise_model"
dtype = "bfloat16"
transformer_dtype = "float8"
min_t = 0
max_t = 0.875

[adapter]
type = "lora"
rank = 32
dtype = "bfloat16"

[optimizer]
type = "AdamW8bitKahan"
lr = 5e-5
betas = [0.9, 0.99]
weight_decay = 0.01
stabilize = false</code></pre>

hollow_knight_LOW_wanA14B.safetensors

t2v-A14B-LOW

t2v-A14B-HIGH

ti2v-5B

Wan 2.2 14BI didn't really want to upload the results to the gallery with only the "high" or only the "low" LoRA. So all you see in the gallery is LOW + HIGH LoRA. The only difference is that a "simple" wan is used for the HIGH section. And for the LOW section lightning-4-step. Maybe it's a little confusing, but I wanted to show both launch options for this LoRA.

Files with SHA256: c2746bda235f2bd7e219dd2e42bd1924fa0660ff3128c3812485108a5678a9cf