VictorMylle/Kohya-ss-sd-scripts
1
0
Fork 1
mirror of https://github.com/kohya-ss/sd-scripts.git synced 2026-04-15 00:32:25 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix: update default value for --discrete_flow_shift in anima training guide

Browse Source
This commit is contained in:
kohya-ss
2026-02-08 10:56:34 +09:00
parent 3f3e2684c7
commit 9e2b4dc8a0
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
docs/anima_train_network.md
View File

@@ -118,7 +118,7 @@ accelerate launch --num_cpu_threads_per_process 1 anima_train_network.py \
--optimizer_type="AdamW8bit" \
--lr_scheduler="constant" \
--timestep_sample_method="logit_normal" \
--discrete_flow_shift=3.0 \
--discrete_flow_shift=1.0 \
--max_train_epochs=10 \
--save_every_n_epochs=1 \
--mixed_precision="bf16" \
@@ -162,7 +162,7 @@ Besides the arguments explained in the [train_network.py guide](train_network.md
* `--timestep_sample_method=<choice>`
- Timestep sampling method. Choose from `logit_normal` (default) or `uniform`.
* `--discrete_flow_shift=<float>`
- Shift for the timestep distribution in Rectified Flow training. Default `3.0`. The shift formula is `t_shifted = (t * shift) / (1 + (shift - 1) * t)`.
- Shift for the timestep distribution in Rectified Flow training. Default `1.0`. The shift formula is `t_shifted = (t * shift) / (1 + (shift - 1) * t)`. 1.0 means no shift.
* `--sigmoid_scale=<float>`
- Scale factor for `logit_normal` timestep sampling. Default `1.0`.
* `--qwen3_max_token_length=<integer>`