Training Memory Monitoring

CUDA allocator sampling and drift warnings shared by TorchLean training loops.

structure TorchLean.Trainer.Memory.State : Type

State carried by the CUDA-memory drift detector used by sustained training runs.

• firstStep : ℕ
• firstAvailableBytes : ℕ

  Driver-free bytes plus reclaimable cached bytes at the first sample.

• warned : Bool

def TorchLean.Trainer.Memory.instReprState.repr : State → ℕ → Std.Format

@[instance_reducible]

instance TorchLean.Trainer.Memory.instReprState : Repr State

def TorchLean.Trainer.Memory.cadence (options : Runtime.Autograd.Torch.Config) (steps requested : ℕ) : ℕ

Resolve an explicit CUDA-memory cadence, or enable periodic sampling for very long runs.

def TorchLean.Trainer.Memory.sample (options : Runtime.Autograd.Torch.Config) (watchEvery totalSteps done : ℕ) (state? : Option State) : IO (Option State)

Sample the CUDA allocator and warn when sustained loss of usable memory projects exhaustion before the requested run completes. Unused tensor buffers and kernel workspaces are reclaimable, so cache growth alone should not look like a loss of memory available for later allocations.