solves adam to adam_torch optimizer issue

config/train_smollm3.py

@@ -29,7 +29,7 @@ class SmolLM3Config:
     save_interval: int = 500
     
     # Optimizer configuration
-    optimizer: str = "adamw"
+    optimizer: str = "adamw_torch"
     beta1: float = 0.9
     beta2: float = 0.95
     eps: float = 1e-8

config/train_smollm3_openhermes_fr.py

@@ -30,7 +30,7 @@ class SmolLM3ConfigOpenHermesFR(SmolLM3Config):
     save_interval: int = 500
     
     # Optimizer configuration
-    optimizer: str = "adamw"
+    optimizer: str = "adamw_torch"
     beta1: float = 0.9
     beta2: float = 0.95
     eps: float = 1e-8

config/train_smollm3_openhermes_fr_a100_large.py

@@ -30,7 +30,7 @@ class SmolLM3ConfigOpenHermesFRA100Large(SmolLM3Config):
     save_interval: int = 1000  # Less frequent saving
     
     # Optimizer configuration - optimized for large batches
-    optimizer: str = "adamw"
+    optimizer: str = "adamw_torch"
     beta1: float = 0.9
     beta2: float = 0.999  # Higher beta2 for stability with large batches
     eps: float = 1e-8

config/train_smollm3_openhermes_fr_a100_multiple_passes.py

@@ -30,7 +30,7 @@ class SmolLM3ConfigOpenHermesFRMultiplePasses(SmolLM3Config):
     save_interval: int = 2000  # Less frequent saving
     
     # Optimizer configuration - stability focused
-    optimizer: str = "adamw"
+    optimizer: str = "adamw_torch"
     beta1: float = 0.9
     beta2: float = 0.999  # Higher beta2 for stability
     eps: float = 1e-8