Fix misleading WandB error when WANDB_DISABLED is set (#39891)

When users set `report_to="wandb"` but also have `WANDB_DISABLED=true` in their environment,
the previous error message was misleading: "WandbCallback requires wandb to be installed. Run pip install wandb."

This was confusing because wandb was actually installed, just disabled via the environment variable.

The fix detects this specific case and provides a clear, actionable error message explaining
the conflict and how to resolve it.

Author: notkisk

src/transformers/integrations/integration_utils.py

@@ -806,6 +806,19 @@ class WandbCallback(TrainerCallback):
     def __init__(self):
         has_wandb = is_wandb_available()
         if not has_wandb:
+            # Check if wandb is actually installed but disabled via WANDB_DISABLED
+            if importlib.util.find_spec("wandb") is not None:
+                # wandb is installed but disabled
+                wandb_disabled = os.getenv("WANDB_DISABLED", "").upper() in ENV_VARS_TRUE_VALUES
+                if wandb_disabled:
+                    raise RuntimeError(
+                        "You specified `report_to='wandb'` but also set the `WANDB_DISABLED` environment variable.\n"
+                        "This disables wandb logging, even though it was explicitly requested.\n\n"
+                        "- To enable wandb logging: unset `WANDB_DISABLED`.\n"
+                        "- To disable logging: use `report_to='none'`.\n\n"
+                        "Note: WANDB_DISABLED is deprecated and will be removed in v5."
+                    )
+            # If wandb is not installed at all, use the original error message
             raise RuntimeError("WandbCallback requires wandb to be installed. Run `pip install wandb`.")
         if has_wandb:
             import wandb