Merge pull request #2 from liesened/liesened-debiased

Fix apply_debiased_estimation for v-pred [SDXL]

Author: liesened

library/custom_train_functions.py

@@ -96,10 +96,13 @@ def add_v_prediction_like_loss(loss, timesteps, noise_scheduler, v_pred_like_los
     return loss
 
 
-def apply_debiased_estimation(loss, timesteps, noise_scheduler):
+def apply_debiased_estimation(loss, timesteps, noise_scheduler, v_prediction=False):
     snr_t = torch.stack([noise_scheduler.all_snr[t] for t in timesteps])  # batch_size
     snr_t = torch.minimum(snr_t, torch.ones_like(snr_t) * 1000)  # if timestep is 0, snr_t is inf, so limit it to 1000
-    weight = 1 / torch.sqrt(snr_t)
+    if v_prediction:
+        weight = 1 / (snr + 1)
+    else:
+        weight = 1 / torch.sqrt(snr_t)
     loss = weight * loss
     return loss
 

sdxl_train.py

@@ -730,7 +730,7 @@ def optimizer_hook(parameter: torch.Tensor):
                     if args.v_pred_like_loss:
                         loss = add_v_prediction_like_loss(loss, timesteps, noise_scheduler, args.v_pred_like_loss)
                     if args.debiased_estimation_loss:
-                        loss = apply_debiased_estimation(loss, timesteps, noise_scheduler)
+                        loss = apply_debiased_estimation(loss, timesteps, noise_scheduler, args.v_parameterization)
 
                     loss = loss.mean()  # mean over batch dimension
                 else:

train_network.py

@@ -998,7 +998,7 @@ def remove_model(old_ckpt_name):
                     if args.v_pred_like_loss:
                         loss = add_v_prediction_like_loss(loss, timesteps, noise_scheduler, args.v_pred_like_loss)
                     if args.debiased_estimation_loss:
-                        loss = apply_debiased_estimation(loss, timesteps, noise_scheduler)
+                        loss = apply_debiased_estimation(loss, timesteps, noise_scheduler, args.v_parameterization)
 
                     loss = loss.mean()  # 平均なのでbatch_sizeで割る必要なし