tensorflow_manager

Bindings for basic::tensorflow_manager namespace

class pyrosetta.rosetta.basic.tensorflow_manager.RosettaTensorflowManager

Bases: pyrosetta.rosetta.utility.SingletonBase_basic_tensorflow_manager_RosettaTensorflowManager_t

A manager class for loading Tensorflow sessions and controlling mapping to the CPU or GPU.

static get_instance() basic::tensorflow_manager::RosettaTensorflowManager

C++: utility::SingletonBase<basic::tensorflow_manager::RosettaTensorflowManager>::get_instance() –> class basic::tensorflow_manager::RosettaTensorflowManager *

class pyrosetta.rosetta.basic.tensorflow_manager.RosettaTensorflowProtocolBase

Bases: pybind11_builtins.pybind11_object

A pure virtual base class for storing Tensorflow sessions and the code for executing them and returning a result. Derived classes will be protocol-specific, and will accept a RosettaTensorflowInput and produce a RosettaTensorflowOutput.

assign(self: pyrosetta.rosetta.basic.tensorflow_manager.RosettaTensorflowProtocolBase, : pyrosetta.rosetta.basic.tensorflow_manager.RosettaTensorflowProtocolBase) pyrosetta.rosetta.basic.tensorflow_manager.RosettaTensorflowProtocolBase

C++: basic::tensorflow_manager::RosettaTensorflowProtocolBase::operator=(const class basic::tensorflow_manager::RosettaTensorflowProtocolBase &) –> class basic::tensorflow_manager::RosettaTensorflowProtocolBase &

clone(self: pyrosetta.rosetta.basic.tensorflow_manager.RosettaTensorflowProtocolBase) pyrosetta.rosetta.basic.tensorflow_manager.RosettaTensorflowProtocolBase

Clone operation.

C++: basic::tensorflow_manager::RosettaTensorflowProtocolBase::clone() const –> class std::shared_ptr<class basic::tensorflow_manager::RosettaTensorflowProtocolBase>

name(self: pyrosetta.rosetta.basic.tensorflow_manager.RosettaTensorflowProtocolBase) str

Get the name of this RosettaTensorflowProtocol.

Must be implemented by derived class.

C++: basic::tensorflow_manager::RosettaTensorflowProtocolBase::name() const –> std::string

pyrosetta.rosetta.basic.tensorflow_manager.get_tensorflow_compilation_instructions(*args, **kwargs)

Overloaded function.

  1. get_tensorflow_compilation_instructions(module_name: str) -> str

  2. get_tensorflow_compilation_instructions(module_name: str, use_single_quotations: bool) -> str

Writes the following message:

The <module_name> requires compilation with Tensorflow support. To compile with Tensorflow support… (Followed by instructions terminating in a carriage return character.)

If use_single_quotations is true, quoted strings use a single quotation (’) instead of a double (“). This is necessary in the context of, for example, the RosettaScripts XSD, where double quotations are prohibited in descriptions.

C++: basic::tensorflow_manager::get_tensorflow_compilation_instructions(const std::string &, const bool) –> std::string