v0.7 (#26)

* update from scratch configs

* update gym pretraining configs - use fewer epochs

* update robomimic pretraining configs - use fewer epochs

* allow trajectory plotting in eval agent

* add simple vit unet

* update avoid pretraining configs - use fewer epochs

* update furniture pretraining configs - use same amount of epochs as before

* add robomimic diffusion unet pretraining configs

* update robomimic finetuning configs - higher lr

* add vit unet checkpoint urls

* update pretraining and finetuning instructions as configs are updated

cfg/d3il/pretrain/avoid_m1/pre_diffusion_mlp.yaml

@@ -25,12 +25,12 @@ wandb:
   run: ${now:%H-%M-%S}_${name}
 
 train:
-  n_epochs: 15000
+  n_epochs: 5000
   batch_size: 16
   learning_rate: 1e-4
   weight_decay: 1e-6
   lr_scheduler:
-    first_cycle_steps: 15000
+    first_cycle_steps: 5000
     warmup_steps: 100
     min_lr: 1e-5
   save_model_freq: 500

cfg/d3il/pretrain/avoid_m1/pre_gaussian_mlp.yaml

@@ -24,12 +24,12 @@ wandb:
   run: ${now:%H-%M-%S}_${name}
 
 train:
-  n_epochs: 10000
+  n_epochs: 5000
   batch_size: 16
   learning_rate: 1e-4
   weight_decay: 1e-6
   lr_scheduler:
-    first_cycle_steps: 10000
+    first_cycle_steps: 5000
     warmup_steps: 100
     min_lr: 1e-5
   save_model_freq: 500

cfg/d3il/pretrain/avoid_m1/pre_gmm_mlp.yaml

@@ -25,12 +25,12 @@ wandb:
   run: ${now:%H-%M-%S}_${name}
 
 train:
-  n_epochs: 10000
-  batch_size: 32
+  n_epochs: 5000
+  batch_size: 16
   learning_rate: 1e-4
   weight_decay: 1e-6
   lr_scheduler:
-    first_cycle_steps: 10000
+    first_cycle_steps: 5000
     warmup_steps: 100
     min_lr: 1e-5
   save_model_freq: 500

cfg/d3il/pretrain/avoid_m2/pre_diffusion_mlp.yaml

@@ -25,12 +25,12 @@ wandb:
   run: ${now:%H-%M-%S}_${name}
 
 train:
-  n_epochs: 15000
+  n_epochs: 5000
   batch_size: 16
   learning_rate: 1e-4
   weight_decay: 1e-6
   lr_scheduler:
-    first_cycle_steps: 15000
+    first_cycle_steps: 5000
     warmup_steps: 100
     min_lr: 1e-5
   save_model_freq: 500

cfg/d3il/pretrain/avoid_m2/pre_gaussian_mlp.yaml

@@ -24,12 +24,12 @@ wandb:
   run: ${now:%H-%M-%S}_${name}
 
 train:
-  n_epochs: 10000
+  n_epochs: 5000
   batch_size: 16
   learning_rate: 1e-4
   weight_decay: 1e-6
   lr_scheduler:
-    first_cycle_steps: 10000
+    first_cycle_steps: 5000
     warmup_steps: 100
     min_lr: 1e-5
   save_model_freq: 500

cfg/d3il/pretrain/avoid_m2/pre_gmm_mlp.yaml

@@ -25,12 +25,12 @@ wandb:
   run: ${now:%H-%M-%S}_${name}
 
 train:
-  n_epochs: 10000
-  batch_size: 32
+  n_epochs: 5000
+  batch_size: 16
   learning_rate: 1e-4
   weight_decay: 1e-6
   lr_scheduler:
-    first_cycle_steps: 10000
+    first_cycle_steps: 5000
     warmup_steps: 100
     min_lr: 1e-5
   save_model_freq: 500

cfg/d3il/pretrain/avoid_m3/pre_diffusion_mlp.yaml

@@ -25,12 +25,12 @@ wandb:
   run: ${now:%H-%M-%S}_${name}
 
 train:
-  n_epochs: 15000
+  n_epochs: 5000
   batch_size: 16
   learning_rate: 1e-4
   weight_decay: 1e-6
   lr_scheduler:
-    first_cycle_steps: 15000
+    first_cycle_steps: 5000
     warmup_steps: 100
     min_lr: 1e-5
   save_model_freq: 500

cfg/d3il/pretrain/avoid_m3/pre_gaussian_mlp.yaml

@@ -24,12 +24,12 @@ wandb:
   run: ${now:%H-%M-%S}_${name}
 
 train:
-  n_epochs: 10000
+  n_epochs: 5000
   batch_size: 16
   learning_rate: 1e-4
   weight_decay: 1e-6
   lr_scheduler:
-    first_cycle_steps: 10000
+    first_cycle_steps: 5000
     warmup_steps: 100
     min_lr: 1e-5
   save_model_freq: 500

cfg/d3il/pretrain/avoid_m3/pre_gmm_mlp.yaml

@@ -25,12 +25,12 @@ wandb:
   run: ${now:%H-%M-%S}_${name}
 
 train:
-  n_epochs: 10000
+  n_epochs: 5000
   batch_size: 32
   learning_rate: 1e-4
   weight_decay: 1e-6
   lr_scheduler:
-    first_cycle_steps: 10000
+    first_cycle_steps: 5000
     warmup_steps: 100
     min_lr: 1e-5
   save_model_freq: 500