Fabio
c36510c21e
Some checks failed
Linux Build / Linux cpu (push) Has been cancelled
Linux Build / Linux cpu-clang (push) Has been cancelled
Linux Build / Linux vulkan (push) Has been cancelled
mac-build / macOS CPU (push) Has been cancelled
Nix Build / Nix (cpu) (push) Has been cancelled
Nix Build / Nix (vulkan) (push) Has been cancelled
windows-build / Windows CPU (push) Has been cancelled
27 lines
1 KiB
C++
27 lines
1 KiB
C++
#pragma once
|
|
|
|
#include <stdexcept>
|
|
#include <string>
|
|
|
|
namespace engine::runtime {
|
|
|
|
// A request the device cannot serve AT THIS SIZE -- e.g. a transcription
|
|
// prompt plus audio whose prefill graph does not fit in VRAM. Distinct from a
|
|
// genuine internal fault: the caller can fix it by sending less, so servers
|
|
// should surface it as a client error rather than an opaque 500.
|
|
class CapacityError : public std::runtime_error {
|
|
public:
|
|
explicit CapacityError(const std::string & message) : std::runtime_error(message) {}
|
|
};
|
|
|
|
// A request the session turns away as sent: an input or option it does not
|
|
// take, or one that does not read, such as a malformed input artifact. The
|
|
// caller can fix it, so servers answer it as a client error, as they do
|
|
// CapacityError. A fault of the model or the server stays a plain
|
|
// runtime_error.
|
|
class InvalidRequestError : public std::runtime_error {
|
|
public:
|
|
explicit InvalidRequestError(const std::string & message) : std::runtime_error(message) {}
|
|
};
|
|
|
|
} // namespace engine::runtime
|