|
| subroutine, public | torch_api::torch_tensor_expand_dim (tensor, dim, extent, result) |
| | Creates an expanded tensor view along one singleton dimension.
|
| subroutine, public | torch_api::torch_tensor_narrow (tensor, dim, start_index, length, result) |
| | Creates a view of a contiguous tensor slice.
|
| subroutine, public | torch_api::torch_tensor_backward (tensor, outer_grad) |
| | Runs autograd on a Torch tensor.
|
| subroutine, public | torch_api::torch_tensor_backward_scalar (tensor) |
| | Runs autograd on a scalar Torch tensor.
|
| subroutine, public | torch_api::torch_tensor_to_device_leaf (tensor, requires_grad) |
| | Moves a tensor to the active Torch device and makes it an autograd leaf.
|
| subroutine, public | torch_api::torch_use_cuda (use_cuda) |
| | Select whether Torch wrappers should use CUDA when available.
|
| subroutine, public | torch_api::torch_tensor_grad (tensor, grad) |
| | Returns the gradient of a Torch tensor which was computed by autograd.
|
| subroutine, public | torch_api::torch_tensor_grad_batch3 (tensor1, tensor2, tensor3, grad1, grad2, grad3) |
| | Copies three autograd gradients to CPU memory.
|
| subroutine, public | torch_api::torch_tensor_weighted_sum (values, weights, result) |
| | Returns the weighted sum of two Torch tensors.
|
| real(kind=dp) function, public | torch_api::torch_tensor_item_double (tensor) |
| | Returns a scalar double value from a Torch tensor.
|
| subroutine, public | torch_api::torch_tensor_release (tensor) |
| | Releases a Torch tensor and all its ressources.
|
| subroutine, public | torch_api::torch_dict_create (dict) |
| | Creates an empty Torch dictionary.
|
| subroutine, public | torch_api::torch_dict_clone (source, target) |
| | Clones a Torch dictionary.
|
| subroutine, public | torch_api::torch_dict_insert (dict, key, tensor) |
| | Inserts a Torch tensor into a Torch dictionary.
|
| subroutine, public | torch_api::torch_dict_get (dict, key, tensor) |
| | Retrieves a Torch tensor from a Torch dictionary.
|
| subroutine, public | torch_api::torch_dict_release (dict) |
| | Releases a Torch dictionary and all its ressources.
|
| subroutine, public | torch_api::torch_model_load (model, filename) |
| | Loads a Torch model from given "*.pth" file. (In Torch lingo models are called modules).
|
| subroutine, public | torch_api::torch_model_load_with_metadata (model, filename, key1, value1, key2, value2) |
| | Loads a Torch model and reads two metadata entries in the same archive pass.
|
| subroutine, public | torch_api::torch_model_remap_device_constants (model) |
| | Maps serialized TorchScript device constants to the active Torch device.
|
| subroutine, public | torch_api::torch_model_disable_parameter_gradients (model) |
| | Disable gradients for inference-only model parameters.
|
| subroutine, public | torch_api::torch_model_forward (model, inputs, outputs) |
| | Evaluates the given Torch model.
|
| subroutine, public | torch_api::torch_model_forward_mol_tensor (model, method_name, inputs, output) |
| | Evaluates a TorchScript model method expecting keyword argument "mol".
|
| subroutine, public | torch_api::torch_model_release (model) |
| | Releases a Torch model and all its ressources.
|
| character(:) function, allocatable, public | torch_api::torch_model_read_metadata (filename, key) |
| | Reads metadata entry from given "*.pth" file. (In Torch lingo they are called extra files).
|
| logical function, public | torch_api::torch_cuda_is_available () |
| | Returns true iff the Torch CUDA backend is available.
|
| integer function, public | torch_api::torch_cuda_device_count () |
| | Return the number of CUDA devices visible to Torch.
|
| subroutine, public | torch_api::torch_allow_tf32 (allow_tf32) |
| | Set whether to allow the use of TF32. Needed due to changes in defaults from pytorch 1.7 to 1.11 to >=1.12 See https://pytorch.org/docs/stable/notes/cuda.html.
|
| subroutine, public | torch_api::torch_model_freeze (model) |
| | Freeze the given Torch model: applies generic optimization that speed up model. See https://pytorch.org/docs/stable/generated/torch.jit.freeze.html.
|
| subroutine, public | torch_api::torch_model_freeze_preserving_method (model, method_name) |
| | Freeze a Torch model while preserving one exported method.
|
1.15.0