↓ 2 callersFunctionlagrange_solve(f_grad, g_val, g_grad, x0, lr=0.01,
lr_lambda=0.01, steps=5000)
phases/01-math-foundations/18-convex-optimization/code/convex.py:106
↓ 2 callersFunctionloss_scaled_for_accum(logits: torch.Tensor, target: torch.Tensor, accum_steps: int, loss_fn: Callable[[torch.Tensor, torch.Tensor],
phases/19-capstone-projects/46-gradient-accumulation/code/main.py:91