Sign In

Frieren Flux 3.79Mo

12
121
0
0
Verified:
SafeTensor
Type
LoRA
Stats
48
0
Reviews
Published
Feb 11, 2025
Base Model
Flux.1 D
Training
Steps: 2,000
Epochs: 10
Usage Tips
Clip Skip: 1
Strength: 1
Trigger Words
Frieren
Hash
AutoV2
B8741096C8
Comfy Workflow Badge
tosave's Avatar
tosave
The FLUX.1 [dev] Model is licensed by Black Forest Labs. Inc. under the FLUX.1 [dev] Non-Commercial License. Copyright Black Forest Labs. Inc.
IN NO EVENT SHALL BLACK FOREST LABS, INC. BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.

Frieren (Sousou no Frieren)

Trigger word : Frieren (no frieren ^^)

Little experience of handling single and double blocks, little report:

unet_lr = 0.0004

lr_scheduler = cosine with 5 restarts

network_args = [ "train_double_block_indices=1,4,6,17,18", "train_single_block_indices=0,3,7,9,12",]

network_dim = 2

network_alpha = 16

resolution = "768,768"

50 images x 16 = 800 repeats / batch 2 x 10 epochs = 4000 steps

I selected the 8th epoch, the following ones were overcooked.

It's hard to estimate which blocks influence what, but little by little I'm finding clues.

A general remark, the simple blocks from 17 to 37 don't seem to do much.

If you have any leads on the blocks I'm interested :)

Have fun !