Added my annotator/preprocessor and improved network jitter on reference trainer

.gitmodules

@@ -4,3 +4,6 @@
 [submodule "repositories/leco"]
 	path = repositories/leco
 	url = https://github.com/p1atdev/LECO
+[submodule "repositories/batch_annotator"]
+	path = repositories/batch_annotator
+	url = https://github.com/ostris/batch-annotator

extensions_built_in/image_reference_slider_trainer/ImageReferenceSliderTrainerProcess.py

@@ -16,6 +16,7 @@ from toolkit import train_tools
 import torch
 from jobs.process import BaseSDTrainProcess
 import random
+from toolkit.basic import value_map
 
 
 def flush():
@@ -91,11 +92,18 @@ class ImageReferenceSliderTrainerProcess(BaseSDTrainProcess):
                 network_neg_weight = network_neg_weight.item()
 
             # get an array of random floats between -weight_jitter and weight_jitter
+            loss_jitter_multiplier = 1.0
             weight_jitter = self.slider_config.weight_jitter
             if weight_jitter > 0.0:
                 jitter_list = random.uniform(-weight_jitter, weight_jitter)
+                orig_network_pos_weight = network_pos_weight
                 network_pos_weight += jitter_list
                 network_neg_weight += (jitter_list * -1.0)
+                # penalize the loss for its distance from network_pos_weight
+                # a jitter_list of abs(3.0) on a weight of 5.0 is a 60% jitter
+                # so the loss_jitter_multiplier needs to be 0.4
+                loss_jitter_multiplier = value_map(abs(jitter_list), 0.0, weight_jitter, 1.0, 0.0)
+
 
             # if items in network_weight list are tensors, convert them to floats
 
@@ -210,7 +218,7 @@ class ImageReferenceSliderTrainerProcess(BaseSDTrainProcess):
                     # add min_snr_gamma
                     loss = apply_snr_weight(loss, timesteps, noise_scheduler, self.train_config.min_snr_gamma)
 
-                loss = loss.mean()
+                loss = loss.mean() * loss_jitter_multiplier
                 loss_slide_float = loss.item()
 
                 loss_float = loss.item()