Shortcuts

Struct CTCLossOptions

Page Contents

Struct Documentation

struct CTCLossOptions

Options for the CTCLoss module.

Example:

CTCLoss
model(CTCLossOptions().blank(42).zero_infinity(false).reduction(torch::kSum));

Public Types

typedef std::variant<enumtype::kNone, enumtype::kMean, enumtype::kSum> reduction_t

Public Functions

inline auto blank(const int64_t &new_blank) -> decltype(*this)

blank label. Default 0.

inline auto blank(int64_t &&new_blank) -> decltype(*this)
inline const int64_t &blank() const noexcept
inline int64_t &blank() noexcept
inline auto reduction(const reduction_t &new_reduction) -> decltype(*this)

Specifies the reduction to apply to the output. Default: Mean.

inline auto reduction(reduction_t &&new_reduction) -> decltype(*this)
inline const reduction_t &reduction() const noexcept
inline reduction_t &reduction() noexcept
inline auto zero_infinity(const bool &new_zero_infinity) -> decltype(*this)

Whether to zero infinite losses and the associated gradients.

Default: false. Infinite losses mainly occur when the inputs are too short to be aligned to the targets.

inline auto zero_infinity(bool &&new_zero_infinity) -> decltype(*this)
inline const bool &zero_infinity() const noexcept
inline bool &zero_infinity() noexcept

Docs

Access comprehensive developer documentation for PyTorch

View Docs

Tutorials

Get in-depth tutorials for beginners and advanced developers

View Tutorials

Resources

Find development resources and get your questions answered

View Resources