huggingface/pytorch-image-models · error · ValueError

Invalid epsilon value: {eps}

Error message

Invalid epsilon value: {eps}

What it means

Raised by ADOPT's constructor when eps is negative. eps is added to denominators for numerical stability and must be >= 0.

Source

Thrown at timm/optim/adopt.py:91

            corrected_weight_decay: bool = False,
            *,
            caution: bool = False,
            foreach: Optional[bool] = False,
            maximize: bool = False,
            capturable: bool = False,
            differentiable: bool = False,
    ):
        if isinstance(lr, Tensor):
            if foreach and not capturable:
                raise ValueError(
                    "lr as a Tensor is not supported for capturable=False and foreach=True"
                )
            if lr.numel() != 1:
                raise ValueError("Tensor lr must be 1-element")
        if not 0.0 <= lr:
            raise ValueError(f"Invalid learning rate: {lr}")
        if not 0.0 <= eps:
            raise ValueError(f"Invalid epsilon value: {eps}")
        if not 0.0 <= betas[0] < 1.0:
            raise ValueError(f"Invalid beta parameter at index 0: {betas[0]}")
        if not 0.0 <= betas[1] < 1.0:
            raise ValueError(f"Invalid beta parameter at index 1: {betas[1]}")
        if not 0.0 <= weight_decay:
            raise ValueError(f"Invalid weight_decay value: {weight_decay}")

        defaults = dict(
            lr=lr,
            betas=betas,
            eps=eps,
            weight_decay=weight_decay,
            clip_exp=clip_exp,
            decoupled=decoupled,
            corrected_weight_decay=corrected_weight_decay,
            caution=caution,
            maximize=maximize,
            foreach=foreach,

View on GitHub (pinned to 9a5261e31b)

Solutions

  1. Use a small non-negative eps, typically 1e-8
  2. Fix the YAML/CLI value bound to eps

Example fix

# before
opt = timm.optim.Adopt(model.parameters(), lr=1e-3, eps=-1e-8)
# after
opt = timm.optim.Adopt(model.parameters(), lr=1e-3, eps=1e-8)
Defensive patterns

Strategy: validation

Validate before calling

assert eps >= 0.0, f'eps must be >= 0, got {eps}'

Type guard

def valid_eps(eps) -> bool:
    return isinstance(eps, (int, float)) and eps >= 0.0

Prevention

When it happens

Trigger: Constructing timm.optim.Adopt(params, eps=-1e-8).

Common situations: Config typo on the eps key, or a sweep over log-spaced eps crossing zero.

Related errors

AI-assisted analysis of huggingface/pytorch-image-models@9a5261e31b (2026-08-27). Data as JSON: /api/errors/7d02ca9a7ab74cc5. Report an issue: GitHub.