lightning/pytorch_lightning/core/memory.py

# Copyright The PyTorch Lightning team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import os
import shutil
import subprocess
from collections import OrderedDict
from typing import Any, Dict, List, Optional, Tuple, Union

import numpy as np
import torch
import torch.nn as nn
from torch.utils.hooks import RemovableHandle

from pytorch_lightning.utilities import AMPType, DeviceType

PARAMETER_NUM_UNITS = [" ", "K", "M", "B", "T"]
UNKNOWN_SIZE = "?"


class LayerSummary(object):
    """
    Summary class for a single layer in a :class:`~pytorch_lightning.core.lightning.LightningModule`.
    It collects the following information:

    - Type of the layer (e.g. Linear, BatchNorm1d, ...)
    - Input shape
    - Output shape
    - Number of parameters

    The input and output shapes are only known after the example input array was
    passed through the model.

    Example::

        >>> model = torch.nn.Conv2d(3, 8, 3)
        >>> summary = LayerSummary(model)
        >>> summary.num_parameters
        224
        >>> summary.layer_type
        'Conv2d'
        >>> output = model(torch.rand(1, 3, 5, 5))
        >>> summary.in_size
        [1, 3, 5, 5]
        >>> summary.out_size
        [1, 8, 3, 3]

    Args:
        module: A module to summarize

    """

    def __init__(self, module: nn.Module):
        super().__init__()
        self._module = module
        self._hook_handle = self._register_hook()
        self._in_size = None
        self._out_size = None

    def __del__(self):
        self.detach_hook()

    def _register_hook(self) -> Optional[RemovableHandle]:
        """
        Registers a hook on the module that computes the input- and output size(s) on the first forward pass.
        If the hook is called, it will remove itself from the from the module, meaning that
        recursive models will only record their input- and output shapes once.
        Registering hooks on :class:`~torch.jit.ScriptModule` is not supported.

        Return:
            A handle for the installed hook, or ``None`` if registering the hook is not possible.
        """

        def hook(module, inp, out):
            if len(inp) == 1:
                inp = inp[0]
            self._in_size = parse_batch_shape(inp)
            self._out_size = parse_batch_shape(out)
            self._hook_handle.remove()

        handle = None
        if not isinstance(self._module, torch.jit.ScriptModule):
            handle = self._module.register_forward_hook(hook)
        return handle

    def detach_hook(self):
        """
        Removes the forward hook if it was not already removed in the forward pass.
        Will be called after the summary is created.
        """
        if self._hook_handle is not None:
            self._hook_handle.remove()

    @property
    def in_size(self) -> Union[str, List]:
        return self._in_size or UNKNOWN_SIZE

    @property
    def out_size(self) -> Union[str, List]:
        return self._out_size or UNKNOWN_SIZE

    @property
    def layer_type(self) -> str:
        """ Returns the class name of the module. """
        return str(self._module.__class__.__name__)

    @property
    def num_parameters(self) -> int:
        """ Returns the number of parameters in this module. """
        return sum(np.prod(p.shape) for p in self._module.parameters())


class ModelSummary(object):
    """
    Generates a summary of all layers in a :class:`~pytorch_lightning.core.lightning.LightningModule`.

    Args:
        model: The model to summarize (also referred to as the root module)
        mode: Can be one of

             - `top` (default): only the top-level modules will be recorded (the children of the root module)
             - `full`: summarizes all layers and their submodules in the root module

    The string representation of this summary prints a table with columns containing
    the name, type and number of parameters for each layer.

    The root module may also have an attribute ``example_input_array`` as shown in the example below.
    If present, the root module will be called with it as input to determine the
    intermediate input- and output shapes of all layers. Supported are tensors and
    nested lists and tuples of tensors. All other types of inputs will be skipped and show as `?`
    in the summary table. The summary will also display `?` for layers not used in the forward pass.

    Example::

        >>> import pytorch_lightning as pl
        >>> class LitModel(pl.LightningModule):
        ...
        ...     def __init__(self):
        ...         super().__init__()
        ...         self.net = nn.Sequential(nn.Linear(256, 512), nn.BatchNorm1d(512))
        ...         self.example_input_array = torch.zeros(10, 256)  # optional
        ...
        ...     def forward(self, x):
        ...         return self.net(x)
        ...
        >>> model = LitModel()
        >>> ModelSummary(model, mode='top')  # doctest: +NORMALIZE_WHITESPACE
          | Name | Type       | Params | In sizes  | Out sizes
        ------------------------------------------------------------
        0 | net  | Sequential | 132 K  | [10, 256] | [10, 512]
        ------------------------------------------------------------
        132 K     Trainable params
        0         Non-trainable params
        132 K     Total params
        0.530     Total estimated model params size (MB)
        >>> ModelSummary(model, mode='full')  # doctest: +NORMALIZE_WHITESPACE
          | Name  | Type        | Params | In sizes  | Out sizes
        --------------------------------------------------------------
        0 | net   | Sequential  | 132 K  | [10, 256] | [10, 512]
        1 | net.0 | Linear      | 131 K  | [10, 256] | [10, 512]
        2 | net.1 | BatchNorm1d | 1.0 K    | [10, 512] | [10, 512]
        --------------------------------------------------------------
        132 K     Trainable params
        0         Non-trainable params
        132 K     Total params
        0.530     Total estimated model params size (MB)
    """

    MODE_TOP = "top"
    MODE_FULL = "full"
    MODE_DEFAULT = MODE_TOP
    MODES = [MODE_FULL, MODE_TOP]

    def __init__(self, model, mode: str = MODE_DEFAULT):
        self._model = model
        self._mode = mode
        self._layer_summary = self.summarize()
        # 1 byte -> 8 bits
        # TODO: how do we compute precisin_megabytes in case of mixed precision?
        precision = self._model.precision if isinstance(self._model.precision, int) else 32
        self._precision_megabytes = (precision / 8.0) * 1e-6

    @property
    def named_modules(self) -> List[Tuple[str, nn.Module]]:
        if self._mode == ModelSummary.MODE_FULL:
            mods = self._model.named_modules()
            mods = list(mods)[1:]  # do not include root module (LightningModule)
        elif self._mode == ModelSummary.MODE_TOP:
            # the children are the top-level modules
            mods = self._model.named_children()
        else:
            mods = []
        return list(mods)

    @property
    def layer_names(self) -> List[str]:
        return list(self._layer_summary.keys())

    @property
    def layer_types(self) -> List[str]:
        return [layer.layer_type for layer in self._layer_summary.values()]

    @property
    def in_sizes(self) -> List:
        return [layer.in_size for layer in self._layer_summary.values()]

    @property
    def out_sizes(self) -> List:
        return [layer.out_size for layer in self._layer_summary.values()]

    @property
    def param_nums(self) -> List[int]:
        return [layer.num_parameters for layer in self._layer_summary.values()]

    @property
    def total_parameters(self) -> int:
        return sum(p.numel() for p in self._model.parameters())

    @property
    def trainable_parameters(self) -> int:
        return sum(p.numel() for p in self._model.parameters() if p.requires_grad)

    @property
    def model_size(self) -> float:
        # todo: seems it does not work with quantized models - it returns 0.0
        return self.total_parameters * self._precision_megabytes

    def summarize(self) -> Dict[str, LayerSummary]:
        summary = OrderedDict((name, LayerSummary(module)) for name, module in self.named_modules)
        if self._model.example_input_array is not None:
            self._forward_example_input()
        for layer in summary.values():
            layer.detach_hook()
        return summary

    def _forward_example_input(self) -> None:
        """ Run the example input through each layer to get input- and output sizes. """
        model = self._model
        trainer = self._model.trainer

        input_ = model.example_input_array
        input_ = model._apply_batch_transfer_handler(input_, model.device)

        if trainer is not None and trainer.amp_backend == AMPType.NATIVE and trainer._device_type != DeviceType.TPU:
            model.forward = torch.cuda.amp.autocast()(model.forward)

        mode = model.training
        model.eval()
        with torch.no_grad():
            # let the model hooks collect the input- and output shapes
            if isinstance(input_, (list, tuple)):
                model(*input_)
            elif isinstance(input_, dict):
                model(**input_)
            else:
                model(input_)
        model.train(mode)  # restore mode of module

    def __str__(self):
        """
        Makes a summary listing with:

        Layer Name, Layer Type, Number of Parameters, Input Sizes, Output Sizes, Model Size
        """
        arrays = [
            [" ", list(map(str, range(len(self._layer_summary))))],
            ["Name", self.layer_names],
            ["Type", self.layer_types],
            ["Params", list(map(get_human_readable_count, self.param_nums))],
        ]
        if self._model.example_input_array is not None:
            arrays.append(["In sizes", self.in_sizes])
            arrays.append(["Out sizes", self.out_sizes])
        total_parameters = self.total_parameters
        trainable_parameters = self.trainable_parameters
        model_size = self.model_size

        return _format_summary_table(total_parameters, trainable_parameters, model_size, *arrays)

    def __repr__(self):
        return str(self)


def parse_batch_shape(batch: Any) -> Union[str, List]:
    if hasattr(batch, "shape"):
        return list(batch.shape)

    if isinstance(batch, (list, tuple)):
        shape = [parse_batch_shape(el) for el in batch]
        return shape

    return UNKNOWN_SIZE


def _format_summary_table(total_parameters: int, trainable_parameters: int, model_size: float, *cols) -> str:
    """
    Takes in a number of arrays, each specifying a column in
    the summary table, and combines them all into one big
    string defining the summary table that are nicely formatted.
    """
    n_rows = len(cols[0][1])
    n_cols = 1 + len(cols)

    # Get formatting width of each column
    col_widths = []
    for c in cols:
        col_width = max(len(str(a)) for a in c[1]) if n_rows else 0
        col_width = max(col_width, len(c[0]))  # minimum length is header length
        col_widths.append(col_width)

    # Formatting
    s = "{:<{}}"
    total_width = sum(col_widths) + 3 * n_cols
    header = [s.format(c[0], l) for c, l in zip(cols, col_widths)]

    # Summary = header + divider + Rest of table
    summary = " | ".join(header) + "\n" + "-" * total_width
    for i in range(n_rows):
        line = []
        for c, l in zip(cols, col_widths):
            line.append(s.format(str(c[1][i]), l))
        summary += "\n" + " | ".join(line)
    summary += "\n" + "-" * total_width

    summary += "\n" + s.format(get_human_readable_count(trainable_parameters), 10)
    summary += "Trainable params"
    summary += "\n" + s.format(get_human_readable_count(total_parameters - trainable_parameters), 10)
    summary += "Non-trainable params"
    summary += "\n" + s.format(get_human_readable_count(total_parameters), 10)
    summary += "Total params"
    summary += "\n" + s.format(get_formatted_model_size(model_size), 10)
    summary += "Total estimated model params size (MB)"

    return summary


def get_memory_profile(mode: str) -> Union[Dict[str, int], Dict[int, int]]:
    """ Get a profile of the current memory usage.

    Args:
        mode: There are two modes:

            - 'all' means return memory for all gpus
            - 'min_max' means return memory for max and min

    Return:
        A dictionary in which the keys are device ids as integers and
        values are memory usage as integers in MB.
        If mode is 'min_max', the dictionary will also contain two additional keys:

        - 'min_gpu_mem': the minimum memory usage in MB
        - 'max_gpu_mem': the maximum memory usage in MB
    """
    memory_map = get_gpu_memory_map()

    if mode == "min_max":
        min_index, min_memory = min(memory_map.items(), key=lambda item: item[1])
        max_index, max_memory = max(memory_map.items(), key=lambda item: item[1])

        memory_map = {"min_gpu_mem": min_memory, "max_gpu_mem": max_memory}

    return memory_map


def get_gpu_memory_map() -> Dict[str, int]:
    """
    Get the current gpu usage.

    Return:
        A dictionary in which the keys are device ids as integers and
        values are memory usage as integers in MB.
    """
    result = subprocess.run(
        [shutil.which("nvidia-smi"), "--query-gpu=memory.used", "--format=csv,nounits,noheader"],
        encoding="utf-8",
        # capture_output=True,          # valid for python version >=3.7
        stdout=subprocess.PIPE,
        stderr=subprocess.PIPE,  # for backward compatibility with python version 3.6
        check=True,
    )

    # Convert lines into a dictionary
    gpu_memory = [float(x) for x in result.stdout.strip().split(os.linesep)]
    gpu_memory_map = {f"gpu_id: {gpu_id}/memory.used (MB)": memory for gpu_id, memory in enumerate(gpu_memory)}
    return gpu_memory_map


def get_formatted_model_size(total_model_size: float) -> float:
    return f"{total_model_size:,.3f}"


def get_human_readable_count(number: int) -> str:
    """
    Abbreviates an integer number with K, M, B, T for thousands, millions,
    billions and trillions, respectively.

    Examples:
        >>> get_human_readable_count(123)
        '123  '
        >>> get_human_readable_count(1234)  # (one thousand)
        '1.2 K'
        >>> get_human_readable_count(2e6)   # (two million)
        '2.0 M'
        >>> get_human_readable_count(3e9)   # (three billion)
        '3.0 B'
        >>> get_human_readable_count(4e14)  # (four hundred trillion)
        '400 T'
        >>> get_human_readable_count(5e15)  # (more than trillion)
        '5,000 T'

    Args:
        number: a positive integer number

    Return:
        A string formatted according to the pattern described above.

    """
    assert number >= 0
    labels = PARAMETER_NUM_UNITS
    num_digits = int(np.floor(np.log10(number)) + 1 if number > 0 else 1)
    num_groups = int(np.ceil(num_digits / 3))
    num_groups = min(num_groups, len(labels))  # don't abbreviate beyond trillions
    shift = -3 * (num_groups - 1)
    number = number * (10**shift)
    index = num_groups - 1
    if index < 1 or number >= 100:
        return f"{int(number):,d} {labels[index]}"

    return f"{number:,.1f} {labels[index]}"