d4f144581d
Co-authored-by: leo <yixingwang@YIXINGs-MacBook-Pro.local> Co-authored-by: aryan5v <email.aryan1005@gmail.com> Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com>
18 lines
852 B
YAML
18 lines
852 B
YAML
transformers:
|
|
fastwan2.1-t2v-1.3b-vsa-dit:
|
|
name: FastWan2.1 T2V 1.3B VSA transformer
|
|
# Sum of tensor shapes in the FastWan checkpoint header at the pinned revision below.
|
|
# Includes 70_824_960 VSA to_gate_compress parameters across the 30 blocks.
|
|
# block_parameter_count includes only blocks.* tensors; the total also includes embeddings and output heads.
|
|
parameter_count: 1_489_821_760
|
|
block_count: 30
|
|
block_parameter_count: 1_464_046_080
|
|
geometry:
|
|
hidden_size: 1536
|
|
ffn_size: 8960
|
|
patch_size: [1, 2, 2]
|
|
gated_ffn: false
|
|
sources:
|
|
- https://huggingface.co/FastVideo/FastWan2.1-T2V-1.3B-Diffusers/tree/25e7ed7f41fd8ce2fdd108688c65e8caf0ce3aef/transformer
|
|
- https://github.com/hao-ai-lab/FastVideo/blob/81b73f6b88b18803a24510a336b0bae1b0c2c4ef/fastvideo/models/wan/transformer.py
|