metadata
license: other
license_name: circlestone-labs-non-commercial-license
license_link: https://huggingface.co/circlestone-labs/Anima/blob/main/LICENSE.md
language:
- en
pipeline_tag: text-to-image
base_model:
- circlestone-labs/Anima-Base-v1.0-Diffusers
base_model_relation: quantized
library_name: diffusers
For more information (including how to compress models yourself), check out https://huggingface.co/DFloat11 and https://github.com/LeanModels/DFloat11
Feel free to request for other models for compression as well (for either the diffusers library, ComfyUI, or any other model), although models that use architectures which are unfamiliar to me might be more difficult.
How to Use
diffusers
import torch
from diffusers import ModularPipeline, CosmosTransformer3DModel
from dfloat11 import DFloat11Model
from transformers.initialization import no_init_weights
with no_init_weights():
transformer = CosmosTransformer3DModel.from_config(
CosmosTransformer3DModel.load_config(
"circlestone-labs/Anima-Base-v1.0-Diffusers",
subfolder="transformer"
),
torch_dtype=torch.bfloat16
).to(torch.bfloat16)
DFloat11Model.from_pretrained("mingyi456/Anima-Base-v1.0-DF11-Diffusers", device="cpu", bfloat16_model=transformer)
text_encoder = DFloat11Model.from_pretrained("mingyi456/Qwen3-0.6B-Base-DF11", device = "cpu")
pipe = ModularPipeline.from_pretrained("circlestone-labs/Anima-Base-v1.0-Diffusers")
pipe.update_components(text_encoder=text_encoder.model)
pipe.update_components(transformer=transformer)
pipe.load_components(dtype=torch.bfloat16)
DFloat11Model.from_pretrained(
"mingyi456/Anima-Base-v1.0-text_conditioner-DF11-Diffusers",
device = "cpu",
bfloat16_model=pipe.text_conditioner
)
pipe.to("cuda")
image = pipe(
prompt="masterpiece, best quality, 1girl, solo, city lights",
generator=torch.Generator("cpu").manual_seed(114514),
).images[0]
image.save("image anima")
ComfyUI
Refer to this model instead.
Compression details
This is the pattern_dict for compression:
pattern_dict = {
r"transformer_blocks\.\d+": (
"norm1.linear_1",
"norm1.linear_2",
"attn1.to_q",
"attn1.to_k",
"attn1.to_v",
"attn1.to_out.0",
"norm2.linear_1",
"norm2.linear_2",
"attn2.to_q",
"attn2.to_k",
"attn2.to_v",
"attn2.to_out.0",
"norm3.linear_1",
"norm3.linear_2",
"ff.net.0.proj",
"ff.net.2"
),
r"time_embed\.t_embedder": (
"linear_1",
"linear_2",
)
}