Added new experimental time step weighing that should solve a lot of issues with distribution. Updated example. Removed a warning

config/examples/train_lora_flux_24gb.yaml

@@ -25,6 +25,8 @@ config:
         # datasets are a folder of images. captions need to be txt files with the same name as the image
         # for instance image2.jpg and image2.txt. Only jpg, jpeg, and png are supported currently
         # images will automatically be resized and bucketed into the resolution specified
+        # on windows, escape back slashes with another backslash so
+        # "C:\\path\\to\\images\\folder"
         - folder_path: "/path/to/images/folder"
           caption_ext: "txt"
           caption_dropout_rate: 0.05  # will drop out the caption 5% of time
@@ -33,17 +35,20 @@ config:
           resolution: [ 512, 768, 1024 ]  # flux enjoys multiple resolutions
       train:
         batch_size: 1
-        steps: 4000  # total number of steps to train 500 - 4000 is a good range
+        steps: 2000  # total number of steps to train 500 - 4000 is a good range
         gradient_accumulation_steps: 1
         train_unet: true
         train_text_encoder: false  # probably won't work with flux
-        content_or_style: balanced  # content, style, balanced
         gradient_checkpointing: true  # need the on unless you have a ton of vram
         noise_scheduler: "flowmatch" # for training only
         optimizer: "adamw8bit"
-        lr: 4e-4
+        lr: 1e-4
         # uncomment this to skip the pre training sample
 #        skip_first_sample: true
+        # uncomment to completely disable sampling
+#        disable_sampling: true
+        # uncomment to use new vell curved weighting. Experimental but may produce better results
+        linear_timesteps: true
 
         # ema will smooth out learning, but could slow it down. Recommended to leave on.
         ema_config: