neural_dream.lua

require 'torch'
require 'nn'
require 'image'
require 'optim'

local iteration_strength = 1
local save_frequency = 1

local loadcaffe_wrap = require 'loadcaffe_wrapper'

--------------------------------------------------------------------------------

local cmd = torch.CmdLine()

-- Basic options
cmd:option('-content_image', 'examples/inputs/tubingen.jpg',
           'Content target image')
cmd:option('-image_size', 512, 'Maximum height / width of generated image')
cmd:option('-gpu', -1, 'Zero-indexed ID of the GPU to use; for CPU mode set -gpu = -1')

-- Optimization options
cmd:option('-num_iterations', 1000)
cmd:option('-normalize_gradients', false)
cmd:option('-learning_rate', 1e1)

-- Output options
cmd:option('-print_iter', 50)
cmd:option('-save_iter', 100)
cmd:option('-output_image', 'out.png')

-- Other options
cmd:option('-pooling', 'max', 'max|avg')
cmd:option('-proto_file', 'models/VGG_ILSVRC_19_layers_deploy.prototxt')
cmd:option('-model_file', 'models/VGG_ILSVRC_19_layers.caffemodel')
cmd:option('-backend', 'nn', 'nn|cudnn')
cmd:option('-seed', -1)

cmd:option('-style_layers', 'relu1_1,relu2_1,relu3_1,relu4_1,relu5_1', 'layers for style')

cmd:option('-acid_dose', 1000)

function nn.SpatialConvolutionMM:accGradParameters()
  -- nop.  not needed by our net
end

local function main(params)
  if params.gpu >= 0 then
    require 'cutorch'
    require 'cunn'
    cutorch.setDevice(params.gpu + 1)
  else
    params.backend = 'nn-cpu'
  end

  if params.backend == 'cudnn' then
    require 'cudnn'
    cudnn.SpatialConvolution.accGradParameters = nn.SpatialConvolutionMM.accGradParameters -- ie: nop
  end
  
  local cnn = loadcaffe_wrap.load(params.proto_file, params.model_file, params.backend):float()
  if params.gpu >= 0 then
    cnn:cuda()
  end
  
  local content_image = image.load(params.content_image, 3)
  content_image = image.scale(content_image, params.image_size, 'bilinear')
  local content_image_caffe = preprocess(content_image):float()

  if params.gpu >= 0 then
    content_image_caffe = content_image_caffe:cuda()
  end
  
  local style_layers = params.style_layers:split(",")

  -- Set up the network, inserting style and content loss modules
  local style_losses = {}
  local next_style_idx = 1
  local net = nn.Sequential()
  for i = 1, #cnn do
    if next_style_idx <= #style_layers then
      local layer = cnn:get(i)
      local name = layer.name
      local layer_type = torch.type(layer)
      local is_pooling = (layer_type == 'cudnn.SpatialMaxPooling' or layer_type == 'nn.SpatialMaxPooling')
      if is_pooling and params.pooling == 'avg' then
        assert(layer.padW == 0 and layer.padH == 0)
        local kW, kH = layer.kW, layer.kH
        local dW, dH = layer.dW, layer.dH
        local avg_pool_layer = nn.SpatialAveragePooling(kW, kH, dW, dH):float()
        if params.gpu >= 0 then avg_pool_layer:cuda() end
        local msg = 'Replacing max pooling at layer %d with average pooling'
        print(string.format(msg, i))
        net:add(avg_pool_layer)
      else
        net:add(layer)
      end
      if name == style_layers[next_style_idx] then
        print("Setting up style layer  ", i, ":", layer.name)
        local gram = GramMatrix():float()
        if params.gpu >= 0 then
          gram = gram:cuda()
        end
        local target_features = net:forward(content_image_caffe):clone()
        local target = gram:forward(target_features):clone()
        target:div(target_features:nElement())
        target:mul(params.acid_dose)
        local norm = params.normalize_gradients
        local loss_module = nn.StyleLoss(iteration_strength, target, norm):float()
        if params.gpu >= 0 then
          loss_module:cuda()
        end
        net:add(loss_module)
        table.insert(style_losses, loss_module)
        next_style_idx = next_style_idx + 1
      end
    end
  end

  -- We don't need the base CNN anymore, so clean it up to save memory.
  cnn = nil
  for i=1,#net.modules do
    local module = net.modules[i]
    if torch.type(module) == 'nn.SpatialConvolutionMM' then
        -- remote these, not used, but uses gpu memory
        module.gradWeight = nil
        module.gradBias = nil
    end
  end
  collectgarbage()
  
  -- Initialize the image
  if params.seed >= 0 then
    torch.manualSeed(params.seed)
  end
  local img = content_image_caffe:clone():float()
  if params.gpu >= 0 then
    img = img:cuda()
  end
  
  -- Run it through the network once to get the proper size for the gradient
  -- All the gradients will come from the extra loss modules, so we just pass
  -- zeros into the top of the net on the backward pass.
  local y = net:forward(img)
  local dy = img.new(#y):zero()

  -- Declaring this here lets us access it in maybe_print
  local optim_state = {
      learningRate = params.learning_rate,
    }

  local function maybe_print(t, loss)
    local verbose = (params.print_iter > 0 and t % params.print_iter == 0)
    if verbose then
      print(string.format('Iteration %d / %d', t, params.num_iterations))
      for i, loss_module in ipairs(style_losses) do
        print(string.format('  Style %d loss: %f', i, loss_module.loss))
      end
      print(string.format('  Total loss: %f', loss))
    end
  end

  local function maybe_save(t)
    local should_save = params.save_iter > 0 and t % params.save_iter == 0
    should_save = should_save or t == params.num_iterations
    if should_save then
      local disp = deprocess(img:double())
      disp = image.minmax{tensor=disp, min=0, max=1}

      local filename = build_filename(params.output_image, t)
      if t == params.num_iterations then
        filename = params.output_image
      end
      image.save(filename, disp)
    end
  end

  -- Function to evaluate loss and gradient. We run the net forward and
  -- backward to get the gradient, and sum up losses from the loss modules.
  -- optim.lbfgs internally handles iteration and calls this fucntion many
  -- times, so we manually count the number of iterations to handle printing
  -- and saving intermediate results.
  local num_calls = 0
  local function feval(x)
    num_calls = num_calls + 1
    net:forward(x)
    local grad = net:backward(x, dy)
    local loss = 0
    for _, mod in ipairs(style_losses) do
      loss = loss + mod.loss
    end
    maybe_print(num_calls, loss)
    maybe_save(num_calls)

    collectgarbage()
    -- optim.lbfgs expects a vector for gradients
    return loss, grad:view(grad:nElement())
  end

  -- Run optimization.

  for t = 1, params.num_iterations do
    local x, losses = optim.adam(feval, img, optim_state)
  end

end
  

function build_filename(output_image, iteration)
  local ext = paths.extname(output_image)
  local basename = paths.basename(output_image, ext)
  if iteration < 10 then
    iteration = '0' .. iteration
  end
  return string.format('%s_%s.%s', basename, iteration, ext)
end


-- Preprocess an image before passing it to a Caffe model.
-- We need to rescale from [0, 1] to [0, 255], convert from RGB to BGR,
-- and subtract the mean pixel.
function preprocess(img)
  local mean_pixel = torch.DoubleTensor({103.939, 116.779, 123.68})
  local perm = torch.LongTensor{3, 2, 1}
  img = img:index(1, perm):mul(256.0)
  mean_pixel = mean_pixel:view(3, 1, 1):expandAs(img)
  img:add(-1, mean_pixel)
  return img
end


-- Undo the above preprocessing.
function deprocess(img)
  local mean_pixel = torch.DoubleTensor({103.939, 116.779, 123.68})
  mean_pixel = mean_pixel:view(3, 1, 1):expandAs(img)
  img = img + mean_pixel
  local perm = torch.LongTensor{3, 2, 1}
  img = img:index(1, perm):div(256.0)
  return img
end


-- Returns a network that computes the CxC Gram matrix from inputs
-- of size C x H x W
function GramMatrix()
  local net = nn.Sequential()
  net:add(nn.View(-1):setNumInputDims(2))
  local concat = nn.ConcatTable()
  concat:add(nn.Identity())
  concat:add(nn.Identity())
  net:add(concat)
  net:add(nn.MM(false, true))
  return net
end


-- Define an nn Module to compute style loss in-place
local StyleLoss, parent = torch.class('nn.StyleLoss', 'nn.Module')

function StyleLoss:__init(strength, target, normalize)
  parent.__init(self)
  self.normalize = normalize or false
  self.strength = strength
  self.target = target
  self.loss = 0
  
  self.gram = GramMatrix()
  self.G = nil
  self.crit = nn.MSECriterion()
end

function StyleLoss:updateOutput(input)
  self.G = self.gram:forward(input)
  self.G:div(input:nElement())
  self.loss = self.crit:forward(self.G, self.target)
--  self.loss = self.loss * self.strength
  self.output = input
  return self.output
end

function StyleLoss:updateGradInput(input, gradOutput)
  local dG = self.crit:backward(self.G, self.target)
  dG:div(input:nElement())
  self.gradInput = self.gram:backward(input, dG)
  if self.normalize then
    self.gradInput:div(torch.norm(self.gradInput, 1) + 1e-8)
  end
  self.gradInput:mul(self.strength)
  self.gradInput:add(gradOutput)
  return self.gradInput
end



local params = cmd:parse(arg)
main(params)