refactor: convert sweep config to split format per model #16

configs/cdae_sweep_config.yaml

@@ -0,0 +1,24 @@
+# CDAE
+sweep_count: 100
+method:     grid # grid, random, bayes
+name:       cdae_grid_sweep
+metric: 
+  goal:     minimize # minimize, maximize
+  name:     valid_loss # valid_MAP@K
+parameters:
+  # batch_size: ##########[COMMON]###############
+  #   values: [16, 32, 64]
+  # lr:
+  #   values: [1e-1, 1e-3, 1e-4, 1e-5]
+  # optimizer:
+  #   values: [adam, adamw]
+  # weight_decay:
+  #   values: [1e-1, 1e-3, 1e-5]
+  neg_times: ###########[MODEL]###############
+    values: [1, 5, 10]
+  hidden_size:
+    values: [32, 64, 128, 256, 512, 1024]
+  corruption_level:
+    values: [.1, .3, .4, .5, .6, .9]
+  hidden_activation:
+    values: [sigmoid, identity]

configs/mf_sweep_config.yaml

@@ -0,0 +1,18 @@
+# CDAE
+sweep_count: 100
+method:     grid # grid, random, bayes
+name:       mf_grid_sweep
+metric: 
+  goal:     minimize
+  name:     valid_loss
+parameters:
+  # batch_size: ##########[COMMON]###############
+  #   values: [16, 32, 64]
+  # lr:
+  #   values: [1e-1, 1e-3, 1e-4, 1e-5]
+  optimizer:
+    values: [adam, adamw]
+  # weight_decay:
+  #   values: [1e-1, 1e-3, 1e-5]
+  embed_size: ###########[MODEL]###############
+    values: [32, 64, 128, 256, 512, 1024]

train.py

@@ -151,7 +151,7 @@ def main(cfg: OmegaConf):
     })
 
     if cfg.wandb and cfg.sweep:
-        sweep_cfg = OmegaConf.load('configs/sweep_config.yaml')
+        sweep_cfg = OmegaConf.load(f'configs/{cfg.model_name.lower()}_sweep_config.yaml')
         merge_cfg = OmegaConf.create({})
         merge_cfg.update(cfg)
         merge_cfg.update(sweep_cfg)