trainer ymls and fix main

configs/stable-diffusion/v1-finetune_everydream.yaml

@@ -1,5 +1,5 @@
 model:
-  base_learning_rate: 9.0e-07
+  base_learning_rate: 1.0e-06
   target: ldm.models.diffusion.ddpm.LatentDiffusion
   params:
     reg_weight: 1.0
@@ -20,15 +20,6 @@ model:
     embedding_reg_weight: 0.0
     unfreeze_model: True
     model_lr: 5.0e-7
-    # scheduler_config:
-    #   target: ldm.lr_scheduler.LambdaLinearScheduler
-    #   params:
-    #     verbosity_interval: 200
-    #     warm_up_steps: 5
-    #     max_decay_steps: 100
-    #     lr_start: 6.0e-7
-    #     lr_max: 8.0e-7 
-    #     lr_min: 1.0e-7 
 
     personalization_config:
       target: ldm.modules.embedding_manager.EmbeddingManager
@@ -93,7 +84,7 @@ data:
       params:
         size: 512
         set: train
-        repeats: 5
+        repeats: 3
     validation:
       target: ldm.data.personalized.PersonalizedBase
       params:
@@ -116,6 +107,6 @@ lightning:
   trainer:
     benchmark: True
     max_epochs: 3
-    #precision: 16 # need lightning 1.6+
+    #precision: 16 # need lightning 1.6+ ??
     #num_nodes: 2 # for multigpu
-    #check_val_every_n_epoch: 1
+    #check_val_every_n_epoch: 2

configs/stable-diffusion/v1-finetune_unfrozen.yaml

@@ -17,9 +17,9 @@ model:
     monitor: val/loss_simple_ema
     scale_factor: 0.18215
     use_ema: False
-    embedding_reg_weight: 0.0
+    #embedding_reg_weight: 0.0
     unfreeze_model: True
-    model_lr: 5.0e-7
+    model_lr: 6.0e-7
     # scheduler_config:
     #   target: ldm.lr_scheduler.LambdaLinearScheduler
     #   params:
@@ -30,14 +30,14 @@ model:
     #     lr_max: 8.0e-7 
     #     lr_min: 1.0e-7 
 
-    personalization_config:
-      target: ldm.modules.embedding_manager.EmbeddingManager
-      params:
-        placeholder_strings: ["*"]
-        initializer_words: ["sculpture"]
-        per_image_tokens: false
-        num_vectors_per_token: 1
-        progressive_words: False
+    # personalization_config:
+    #   target: ldm.modules.embedding_manager.EmbeddingManager
+    #   params:
+    #     placeholder_strings: ["*"]
+    #     initializer_words: ["sculpture"]
+    #     per_image_tokens: false
+    #     num_vectors_per_token: 1
+    #     progressive_words: False
 
     unet_config:
       target: ldm.modules.diffusionmodules.openaimodel.UNetModel
@@ -109,13 +109,13 @@ lightning:
     image_logger:
       target: main.ImageLogger
       params:
-        batch_frequency: 300
-        max_images: 16
+        batch_frequency: 500
+        max_images: 12
         increase_log_steps: False
 
   trainer:
     benchmark: True
-    max_epochs: 3
+    max_epochs: 5
     #precision: 16 # need lightning 1.6+
     #num_nodes: 2 # for multigpu
     #check_val_every_n_epoch: 1

main.py

@@ -153,7 +153,7 @@ def get_parser(**parser_kwargs):
         "--max_training_steps",
         type=int,
         required=False,
-        default=9400,
+        default=35000,
         help="Number of iterations to run")
 
     parser.add_argument("--actual_resume", 
@@ -605,21 +605,22 @@ if __name__ == "__main__":
         logger_cfg = OmegaConf.merge(default_logger_cfg, logger_cfg)
         trainer_kwargs["logger"] = instantiate_from_config(logger_cfg)
 
-        # modelcheckpoint - use TrainResult/EvalResult(checkpoint_on=metric) to
-        # specify which metric is used to determine best models
-        # default_modelckpt_cfg = {
-        #     "target": "pytorch_lightning.callbacks.ModelCheckpoint",
-        #     "params": {
-        #         "dirpath": ckptdir,
-        #         "filename": "{epoch:03}",
-        #         "verbose": True,
-        #         "save_last": True,
-        #     }
-        # }
+        #modelcheckpoint - use TrainResult/EvalResult(checkpoint_on=metric) to
+        #specify which metric is used to determine best models
+        default_modelckpt_cfg = {
+            "target": "pytorch_lightning.callbacks.ModelCheckpoint",
+            "params": {
+                "dirpath": ckptdir,
+                "filename": "{epoch:03}",
+                "verbose": True,
+                "save_last": True,
+            }
+        }
+
         if hasattr(model, "monitor"):
             print(f"Monitoring {model.monitor} as checkpoint metric.")
             default_modelckpt_cfg["params"]["monitor"] = model.monitor
-            default_modelckpt_cfg["params"]["save_top_k"] = 1
+            default_modelckpt_cfg["params"]["save_top_k"] = 3
 
         if "modelcheckpoint" in lightning_config:
             modelckpt_cfg = lightning_config.modelcheckpoint