private void m_internalThread_DoWork(object sender, ActionStateArgs <T> e)
{
while (!m_internalThread.CancellationPending)
{
if (m_evtDataReady.WaitOne(10))
{
for (int i = 0; i < m_nNumBatches; i++)
{
for (int j = 0; j < m_nBatchSize; j++)
{
if (m_evtCancel.WaitOne(0))
{
return;
}
int nIdx = (i * m_nBatchSize) + j;
int nImgIdx = (int)Convert.ChangeType(m_rgInput[nIdx], typeof(int));
SimpleDatum d = m_imgdb.QueryImage(m_src.ID, nImgIdx, IMGDB_LABEL_SELECTION_METHOD.NONE, IMGDB_IMAGE_SELECTION_METHOD.NONE);
m_rgrgDataQueue[i].Push(new Datum(d));
}
}
}
}
}
private async Task<string> DownloadLatestPatcher()
{
Label = DownloadingLabel;
try
{
if (CancelEvent.WaitOne(0))
return null;
Downloader downloader = new Downloader(CancelEvent);
downloader.DownloadProgress += OnDownloadProgress;
HttpFileInfo latest = await GetLatestPatcherUrl(downloader);
if (latest == null)
throw new Exception("Не удалось найти свежую версию программы установки. Проверьте файл конфигурации.");
Maximum = latest.ContentLength;
string filePath = Path.GetTempFileName();
await downloader.Download(latest.Url, filePath);
return filePath;
}
finally
{
Label = InstallLabel;
}
}
protected override async Task DoAction()
{
Label = DownloadingLabel;
try
{
if (CancelEvent.WaitOne(0))
{
return;
}
Downloader downloader = new Downloader(CancelEvent);
downloader.DownloadProgress += OnDownloadProgress;
HttpFileInfo fileInfo = _latestTranslationInfo.Value;
Maximum = fileInfo.ContentLength;
await downloader.Download(fileInfo.Url, PatcherService.ArchiveFileName);
if (fileInfo.LastModified != null)
{
File.SetLastWriteTime(PatcherService.ArchiveFileName, fileInfo.LastModified.Value);
}
SubLabel = null;
}
finally
{
Label = DownloadLabel;
}
}
/// <summary>
/// The DownSample method reduces the number of samples per second in the resulting sample set.
/// </summary>
/// <param name="nNewSamplesPerSecond">Specifies the new (lower) samples per second - must be a factor of the original samples per second.</param>
/// <returns>A new WAVProcessor with the new downsamples samples is returned.</returns>
public WAVProcessor DownSample(int nNewSamplesPerSecond)
{
if (m_fmt.nSamplesPerSec < nNewSamplesPerSecond)
{
throw new Exception("The new sample rate (" + nNewSamplesPerSecond.ToString() + ") must be less than the old sample rate (" + m_fmt.nSamplesPerSec.ToString() + ").");
}
if (m_fmt.nSamplesPerSec % nNewSamplesPerSecond != 0)
{
throw new Exception("The new sample rate (" + nNewSamplesPerSecond.ToString() + ") must be a factor of the old sample rate (" + m_fmt.nSamplesPerSec.ToString() + ").");
}
int nStep = (int)m_fmt.nSamplesPerSec / nNewSamplesPerSecond;
List <double[]> rgrgNewSamples = new List <double[]>();
List <double> rgSamples = new List <double>();
int nIdx = 0;
int nTotal = m_rgrgSamples.Count * m_rgrgSamples[0].Length;
Stopwatch sw = new Stopwatch();
sw.Start();
for (int i = 0; i < m_rgrgSamples.Count; i++)
{
for (int j = 0; j < m_rgrgSamples[i].Length; j++)
{
if (j % nStep == 0)
{
double fVal = m_rgrgSamples[i][j];
rgSamples.Add(fVal);
if (m_evtCancel.WaitOne(0))
{
return(null);
}
if (sw.Elapsed.TotalMilliseconds > 1000)
{
double dfPct = (double)nIdx / (double)nTotal;
m_log.WriteLine("Downsampling at " + dfPct.ToString("P") + "...");
sw.Restart();
}
}
nIdx++;
}
rgrgNewSamples.Add(rgSamples.ToArray());
rgSamples = new List <double>();
}
WaveFormat fmt = m_fmt;
fmt.nSamplesPerSec = (uint)nNewSamplesPerSecond;
fmt.nAvgBytesPerSec = (uint)(nNewSamplesPerSecond * fmt.wBitsPerSample);
return(new WAVProcessor(fmt, rgrgNewSamples, m_log, m_evtCancel));
}
public bool DeleteDirectory(string strDir)
{
if (!Directory.Exists(strDir))
{
return(false);
}
Stopwatch sw = new Stopwatch();
string[] rgstrFiles = Directory.GetFiles(strDir);
sw.Start();
for (int i = 0; i < rgstrFiles.Length; i++)
{
File.Delete(rgstrFiles[i]);
//SHFILEOPSTRUCT shf = new SHFILEOPSTRUCT();
//shf.wFunc = FO_DELETE;
//shf.fFlags = FOF_ALLOWUNDO + FOF_NOCONFIRMATION;
//shf.pFrom = strDir + '\0' + '\0';
//SHFileOperation(ref shf);
if (m_evtCancel.WaitOne(0))
{
return(false);
}
if (sw.Elapsed.TotalMilliseconds > 1000)
{
m_log.Progress = (double)i / (double)rgstrFiles.Length;
m_log.WriteLine("deleting " + i.ToString("N0") + " of " + rgstrFiles.Length.ToString("N0") + "...");
sw.Restart();
}
}
rgstrFiles = Directory.GetFiles(strDir);
if (rgstrFiles.Length == 0)
{
Directory.Delete(strDir);
}
return(true);
}
/// <summary>
/// Retrieve an item from the front of the queue, but do not remove it.
/// </summary>
/// <remarks>
/// This function will wait until either data is added to the queue or the CancelEvent or AbortEvent are set.
/// </remarks>
/// <param name="t">Specifies the item removed.</param>
/// <returns>If a CancelEvent occurs, <i>false</i> is returned, otherwise if the data is successfully removed from the queue <i>true</i> is returned.</returns>
public bool Peek(ref T t)
{
while (Count == 0)
{
if (m_evtReady == null)
{
return(false);
}
if (m_evtCancel.WaitOne(0))
{
return(false);
}
Thread.Sleep(0);
}
lock (m_syncObj)
{
t = m_rgData[0];
return(true);
}
}
protected override async Task DoAction()
{
Label = PlayingLabel;
try
{
Maximum = 2;
GameLocationInfo gameLocation = PatcherService.GetGameLocation(FFXIIIGamePart.Part1);
gameLocation.Validate();
Position = 1;
if (CancelEvent.WaitOne(0))
{
return;
}
if (MusicPlayer != null && MusicPlayer.PlaybackState == NAudio.Wave.PlaybackState.Playing)
{
MusicPlayer.Pause();
}
String args = GameSettings.GetGameProcessArguments();
await Task.Factory.StartNew(() => Process.Start(gameLocation.ExecutablePath, args));
Position = 2;
if (InteractionService.LocalizatorEnvironment.Provide().ExitAfterRunGame)
{
Application.Current.MainWindow.Close();
}
}
finally
{
Label = PlayLabel;
}
}
private bool loadFile(string strImagesFile, string strSourceName, int nExtractTotal, ref int nExtractIdx, int nTotal, ref int nIdx, Log log, bool bExtractFiles, Dictionary <string, int> rgNameToLabel)
{
Stopwatch sw = new Stopwatch();
reportProgress(nIdx, nTotal, " Source: " + strSourceName);
reportProgress(nIdx, nTotal, " loading " + strImagesFile + "...");
FileStream fs = null;
try
{
int nSrcId = m_factory.AddSource(strSourceName, 3, -1, -1, false);
addLabels(nSrcId, rgNameToLabel);
m_factory.Open(nSrcId, 500, Database.FORCE_LOAD.NONE, log);
int nPos = strImagesFile.ToLower().LastIndexOf(".tar");
string strPath = strImagesFile.Substring(0, nPos);
if (!Directory.Exists(strPath))
{
Directory.CreateDirectory(strPath);
}
if (bExtractFiles)
{
log.Progress = (double)nIdx / nExtractTotal;
log.WriteLine("Extracting files from '" + strImagesFile + "'...");
if ((nExtractIdx = TarFile.ExtractTar(strImagesFile, strPath, m_evtCancel, log, nExtractTotal, nExtractIdx)) == 0)
{
log.WriteLine("Aborted.");
return(false);
}
}
// Load the annotations.
SimpleDatum.ANNOTATION_TYPE type = SimpleDatum.ANNOTATION_TYPE.BBOX;
int nResizeHeight = 0;
int nResizeWidth = 0;
// Create the training database images.
// Create the master list file.
List <Tuple <string, string> > rgFiles = createFileList(log, strPath);
sw.Start();
for (int i = 0; i < rgFiles.Count; i++)
{
SimpleDatum datum = loadDatum(log, rgFiles[i].Item1, rgFiles[i].Item2, nResizeHeight, nResizeWidth, type, rgNameToLabel);
m_factory.PutRawImageCache(nIdx, datum);
nIdx++;
if (m_evtCancel.WaitOne(0))
{
log.WriteLine("Aborted.");
return(false);
}
if (sw.Elapsed.TotalMilliseconds > 1000)
{
log.Progress = (double)nIdx / nTotal;
log.WriteLine("Loading file " + i.ToString() + " of " + rgFiles.Count.ToString() + "...");
sw.Restart();
}
}
m_factory.ClearImageCashe(true);
m_factory.Close();
}
finally
{
if (fs != null)
{
fs.Dispose();
}
}
return(true);
}
private void m_bwProcess_DoWork(object sender, DoWorkEventArgs e)
{
BackgroundWorker bw = sender as BackgroundWorker;
Log log = new Log("MyCaffe");
MyCaffeControl <float> caffe = null;
log.OnWriteLine += log_OnWriteLine1;
while (!m_evtCancel.WaitOne(0))
{
List <WaitHandle> rgWait = new List <WaitHandle>();
rgWait.AddRange(m_evtCancel.Handles);
rgWait.Add(m_evtCommandRead);
int nWait = WaitHandle.WaitAny(rgWait.ToArray());
if (nWait > 0)
{
try
{
switch (m_Cmd)
{
case COMMAND.CREATE:
SettingsCaffe settings = new SettingsCaffe();
settings.ImageDbLoadMethod = IMAGEDB_LOAD_METHOD.LOAD_ALL;
settings.EnableRandomInputSelection = true;
caffe = new MyCaffeControl <float>(settings, log, m_evtCaffeCancel);
string strSolver = System.Text.Encoding.UTF8.GetString(Properties.Resources.lenet_solver);
string strModel = System.Text.Encoding.UTF8.GetString(Properties.Resources.lenet_train_test);
caffe.Load(Phase.TRAIN, strSolver, strModel, null);
bw.ReportProgress(1, new ProgressInfo(1, 1, "MyCaffe Created.", null, true));
break;
case COMMAND.DESTROY:
m_bCaffeCreated = false;
caffe.Dispose();
caffe = null;
bw.ReportProgress(0, new ProgressInfo(0, 0, "MyCaffe Destroyed", null, false));
break;
case COMMAND.TRAIN:
caffe.Train(5000);
m_rgTrainedWeights = caffe.GetWeights();
m_sdImageMean = caffe.GetImageMean();
bw.ReportProgress(0, new ProgressInfo(0, 0, "MyCaffe Training completed.", null, true));
break;
case COMMAND.TEST:
double dfAccuracy = caffe.Test(100);
log.WriteLine("Accuracy = " + dfAccuracy.ToString("P"));
bw.ReportProgress(0, new ProgressInfo(0, 0, "MyCaffe Testing completed.", null, true));
break;
case COMMAND.DEVICEINFO:
string str1 = caffe.GetDeviceName(0);
str1 += Environment.NewLine;
str1 += caffe.Cuda.GetDeviceInfo(0, true);
bw.ReportProgress(0, new ProgressInfo(0, 0, str1, null, true));
break;
}
}
catch (Exception excpt)
{
log.WriteError(excpt);
}
}
m_evtCaffeCancel.Reset();
}
if (caffe != null)
{
caffe.Dispose();
m_evtCancel.Reset();
m_bCaffeCreated = false;
}
m_evtThreadDone.Set();
}
private bool loadFile(DatasetFactory factory, string strImagesFile, string strLabelsFile, string strSourceName, string strExportPath)
{
if (strExportPath != null)
{
strExportPath += strSourceName;
if (!Directory.Exists(strExportPath))
{
Directory.CreateDirectory(strExportPath);
}
}
Stopwatch sw = new Stopwatch();
reportProgress(0, 0, " Source: " + strSourceName);
reportProgress(0, 0, " loading " + strImagesFile + "...");
BinaryFile image_file = new BinaryFile(strImagesFile);
BinaryFile label_file = new BinaryFile(strLabelsFile);
try
{
// Verify the files
uint magicImg = image_file.ReadUInt32();
uint magicLbl = label_file.ReadUInt32();
if (magicImg != 2051)
{
throw new Exception("Incorrect image file magic.");
}
if (magicLbl != 2049)
{
throw new Exception("Incorrect label file magic.");
}
uint num_items = image_file.ReadUInt32();
uint num_labels = label_file.ReadUInt32();
if (num_items != num_labels)
{
throw new Exception("The number of items must be equal to the number of labels!");
}
// Add the data source to the database.
uint rows = image_file.ReadUInt32();
uint cols = image_file.ReadUInt32();
int nChannels = 1; // black and white
if (factory != null)
{
int nSrcId = factory.AddSource(strSourceName, nChannels, (int)cols, (int)rows, false, 0, true);
factory.Open(nSrcId, 500, Database.FORCE_LOAD.NONE, m_log);
factory.DeleteSourceData();
}
// Storing to database;
byte[] rgLabel;
byte[] rgPixels;
Datum datum = new Datum(false, nChannels, (int)cols, (int)rows, -1, DateTime.MinValue, new List <byte>(), 0, false, -1);
string strAction = (m_param.ExportToFile) ? "exporing" : "loading";
reportProgress(0, (int)num_items, " " + strAction + " a total of " + num_items.ToString() + " items.");
reportProgress(0, (int)num_items, " (with rows: " + rows.ToString() + ", cols: " + cols.ToString() + ")");
sw.Start();
List <SimpleDatum> rgImg = new List <SimpleDatum>();
FileStream fsFileDesc = null;
StreamWriter swFileDesc = null;
if (m_param.ExportToFile)
{
string strFile = strExportPath + "\\file_list.txt";
fsFileDesc = File.OpenWrite(strFile);
swFileDesc = new StreamWriter(fsFileDesc);
}
for (int i = 0; i < num_items; i++)
{
rgPixels = image_file.ReadBytes((int)(rows * cols));
rgLabel = label_file.ReadBytes(1);
if (sw.Elapsed.TotalMilliseconds > 1000)
{
reportProgress(i, (int)num_items, " " + strAction + " data...");
sw.Restart();
}
datum.SetData(rgPixels.ToList(), (int)rgLabel[0]);
if (factory != null)
{
factory.PutRawImageCache(i, datum);
}
else if (strExportPath != null)
{
saveToFile(strExportPath, i, datum, swFileDesc);
}
rgImg.Add(new SimpleDatum(datum));
if (m_evtCancel.WaitOne(0))
{
return(false);
}
}
if (swFileDesc != null)
{
swFileDesc.Flush();
swFileDesc.Close();
swFileDesc.Dispose();
fsFileDesc.Close();
fsFileDesc.Dispose();
}
if (factory != null)
{
factory.ClearImageCashe(true);
factory.UpdateSourceCounts();
factory.SaveImageMean(SimpleDatum.CalculateMean(m_log, rgImg.ToArray(), new WaitHandle[] { new ManualResetEvent(false) }), true);
}
reportProgress((int)num_items, (int)num_items, " " + strAction + " completed.");
}
finally
{
image_file.Dispose();
label_file.Dispose();
}
return(true);
}
/// <summary>
/// Renders the deep draw image(s) depending on the Octave's installed.
/// </summary>
/// <param name="bmpInput">Specifies the input image.</param>
/// <param name="nFocusLabel">Specifies a label to focus on (use this when running on classifying layers).</param>
/// <param name="dfDetailPercentageToOutput">Optionally, specifies the amount of detail to apply to the original image when producing the final image (Default = 0.25 for 25%).</param>
/// <param name="strOutputDir">Optionally, specifies the output directory wheren images are to be output. When <i>null</i>, no images are output, but are instead set in each Octave.</param>
/// <param name="bVisualizeEachStep">Optionally, specifies to create an image at each step of the process which can be useful when making a video of the evolution (default = <i>false</i>).</param>
/// <param name="rgDirectInputs">Optionally, specifies the direct inputs used to set each output. When not <i>null</i> the direct inputs are used instead of the <i>nFocusLabel</i> whereby the
/// network outputs are set to the direct input values and the <i>nFocusLabel</i> is used to index the image and should therefore be unique for each set of direct inputs.
/// By default, this value is set to <i>null</i>.
/// </param>
/// <returns>Upon completing the render, this method returns <i>true</i>, otherwise if cancelled it returns <i>false</i>.</returns>
public bool Render(Bitmap bmpInput, int nFocusLabel = -1, double dfDetailPercentageToOutput = 0.25, string strOutputDir = null, bool bVisualizeEachStep = false, float[] rgDirectInputs = null)
{
if (rgDirectInputs != null && nFocusLabel < 0)
{
throw new Exception("The focus label must be set to a unique value >= 0 that corresponds to this specific direct input set.");
}
// get the input dimensions from net
Blob <T> blobSrc = m_net.blob_by_name("data");
int nW = blobSrc.width;
int nH = blobSrc.height;
m_log.WriteLine("Starting drawing...");
blobSrc.Reshape(1, 3, nH, nW); // resize the networks input.
// Set the base data.
if (strOutputDir != null)
{
bmpInput.Save(strOutputDir + "\\input_image.png");
}
Datum d = ImageData.GetImageData(bmpInput, 3, false, -1);
m_blobBase.mutable_cpu_data = m_transformer.Transform(d);
m_blobDetail.SetData(0.0);
m_blobBlur.SetData(0);
for (int i = 0; i < m_rgOctaves.Count; i++)
{
Octaves o = m_rgOctaves[i];
// Select layer.
string strLayer = o.LayerName;
// Add changed details to the image.
if (nFocusLabel < 0)
{
m_cuda.add(blobSrc.count(), m_blobBase.gpu_data, m_blobDetail.gpu_data, blobSrc.mutable_gpu_data, o.PercentageOfPreviousOctaveDetailsToApply);
}
for (int j = 0; j < o.IterationN; j++)
{
if (m_evtCancel.WaitOne(0))
{
return(false);
}
if (nFocusLabel >= 0)
{
blobSrc.CopyFrom(m_blobBase);
}
double dfSigma = o.StartSigma + ((o.EndSigma - o.StartSigma) * j) / o.IterationN;
double dfStepSize = o.StartStepSize + ((o.EndStepSize - o.StartStepSize) * j) / o.IterationN;
make_step(strLayer, dfSigma, dfStepSize, nFocusLabel, rgDirectInputs);
if ((bVisualizeEachStep || (j == o.IterationN - 1 && o.Save)))
{
// Get the detail.
m_cuda.sub(m_blobDetail.count(), blobSrc.gpu_data, m_blobBase.gpu_data, m_blobDetail.mutable_gpu_data);
if (dfDetailPercentageToOutput < 1.0)
{
// reuse blob blur memory.
m_cuda.add(m_blobBlur.count(), m_blobBase.gpu_data, m_blobDetail.gpu_data, m_blobBlur.mutable_gpu_data, dfDetailPercentageToOutput);
}
else
{
m_blobBlur.CopyFrom(blobSrc);
}
Image bmp = getImage(m_blobBlur);
if (nFocusLabel < 0)
{
Bitmap bmp1 = AdjustContrast(bmp, 0.9f, 1.6f, 1.2f);
bmp.Dispose();
bmp = bmp1;
}
if (strOutputDir != null)
{
string strFile = strOutputDir + "\\" + o.UniqueName + "_" + j.ToString();
if (nFocusLabel >= 0)
{
if (rgDirectInputs != null)
{
strFile += "_idx_" + nFocusLabel.ToString();
}
else
{
strFile += "_class_" + nFocusLabel.ToString();
}
}
bmp.Save(strFile + ".png");
}
if (j == o.IterationN - 1)
{
o.Images.Add(nFocusLabel, bmp);
}
else
{
bmp.Dispose();
}
}
m_log.Progress = (double)j / (double)o.IterationN;
m_log.WriteLine("Focus Label: " + nFocusLabel.ToString() + " Octave: '" + o.LayerName + "' - " + j.ToString() + " of " + o.IterationN.ToString() + " " + m_log.Progress.ToString("P"));
if (nFocusLabel >= 0)
{
m_blobBase.CopyFrom(blobSrc);
}
}
// Extract details produced on the current octave.
if (nFocusLabel < 0)
{
m_cuda.sub(m_blobDetail.count(), blobSrc.gpu_data, m_blobBase.gpu_data, m_blobDetail.mutable_gpu_data);
}
}
m_log.WriteLine("Rendering completed!");
return(true);
}
/// <summary>
/// The consoldiate thread synchronized all data queries using their synchronization field (field #0) to make sure
/// that all data items line up.
/// </summary>
private void consolidateThread()
{
int nAllDataReady = ((int)Math.Pow(2, m_colDataQuery.Count)) - 1;
int nWait = 0;
while (!m_evtCancel.WaitOne(nWait))
{
if (!m_evtEnabled.WaitOne(0))
{
nWait = 250;
m_evtPaused.Set();
continue;
}
nWait = 0;
m_evtPaused.Reset();
int nDataReady = 0;
int nDataDone = 0;
for (int i = 0; i < m_colDataQuery.Count; i++)
{
DataQuery dq = m_colDataQuery[i];
if (dq.DataReady(1))
{
nDataReady |= (0x0001 << i);
}
if (dq.DataDone())
{
nDataDone |= (0x0001 << i);
}
}
if (nDataDone != 0)
{
m_colData.QueryEnd.Set();
}
if (nDataReady != nAllDataReady)
{
continue;
}
DataItem di = new DataItem(m_nFieldCount);
int nLocalFieldCount = m_colDataQuery[0].FieldCount;
double[] rg = m_colDataQuery[0].GetNextData();
if (rg == null)
{
continue;
}
DateTime dtSync = Utility.ConvertTimeFromMinutes(rg[0]);
bool bSkip = false;
for (int i = 0; i < m_nSegmentSize; i++)
{
int nFieldIdx = di.Add(0, i, rg, nLocalFieldCount);
for (int j = 1; j < m_colDataQuery.Count; j++)
{
double[] rg1 = m_colDataQuery[j].GetNextData();
if (rg1 == null)
{
bSkip = true;
break;
}
DateTime dtSync1 = Utility.ConvertTimeFromMinutes(rg1[0]);
while (dtSync1 < dtSync)
{
rg1 = m_colDataQuery[j].GetNextData();
if (rg1 == null)
{
bSkip = true;
break;
}
dtSync1 = Utility.ConvertTimeFromMinutes(rg1[0]);
}
if (bSkip)
{
break;
}
nLocalFieldCount = m_colDataQuery[j].FieldCount;
nFieldIdx = di.Add(nFieldIdx, i, rg1, nLocalFieldCount);
}
if (bSkip)
{
break;
}
}
if (!bSkip)
{
m_colData.Add(di);
}
}
}
/// <summary>
/// Run the trained model. When run each hand-written image is fed in sequence (by label,
/// e.g. 0,1,2,...,9 through the model, yet images within each label are selected at random.
/// </summary>
/// <param name="mycaffe">Specifies the mycaffe instance running the sequence run model.</param>
/// <param name="bw">Specifies the background worker.</param>
private void runModel(MyCaffeControl <float> mycaffe, BackgroundWorker bw)
{
Random random = new Random((int)DateTime.Now.Ticks);
// Get the internal RUN net and associated blobs.
Net <float> net = m_mycaffe.GetInternalNet(Phase.RUN);
Blob <float> blobData = net.FindBlob("data");
Blob <float> blobClip1 = net.FindBlob("clip1");
Blob <float> blobIp1 = net.FindBlob("ip1");
List <float> rgPrediction = new List <float>();
List <float> rgTarget = new List <float>();
List <float> rgT = new List <float>();
m_mycaffeInput.UpdateRunWeights();
blobClip1.SetData(0);
bool bForcedError = false;
for (int i = 0; i < 100; i++)
{
if (m_evtCancel.WaitOne(0))
{
return;
}
int nLabelSeq = m_nLabelSeq;
if (m_evtForceError.WaitOne(0))
{
nLabelSeq = random.Next(10);
bForcedError = true;
}
else
{
bForcedError = false;
}
// Get images one number at a time, in order by label, but randomly selected.
SimpleDatum sd = m_imgDb.QueryImage(m_ds.TrainingSource.ID, 0, null, IMGDB_IMAGE_SELECTION_METHOD.RANDOM, nLabelSeq);
ResultCollection res = m_mycaffeInput.Run(sd);
Net <float> inputNet = m_mycaffeInput.GetInternalNet(Phase.RUN);
Blob <float> input_ip = inputNet.FindBlob(m_strInputOutputBlobName);
Dictionary <string, float[]> data = Signal.GenerateSample(1, m_nLabelSeq / 10.0f, 1, m_model.InputLabel, m_model.TimeSteps);
float[] rgFY1 = data["FY"];
// Run the model.
blobClip1.SetData(1);
blobData.mutable_cpu_data = input_ip.mutable_cpu_data;
net.Forward();
rgPrediction.AddRange(blobIp1.mutable_cpu_data);
// Graph and show the results.
float[] rgFT = data["FT"];
float[] rgFY = data["FY"];
for (int j = 0; j < rgFT.Length; j++)
{
rgT.Add(rgFT[j]);
rgTarget.Add(rgFY[j]);
}
while (rgTarget.Count * 5 > pbImage.Width)
{
rgTarget.RemoveAt(0);
rgPrediction.RemoveAt(0);
}
// Plot the graph.
PlotCollection plotsTarget = createPlots("Target", rgT.ToArray(), new List <float[]>()
{
rgTarget.ToArray()
}, 0);
PlotCollection plotsPrediction = createPlots("Predicted", rgT.ToArray(), new List <float[]>()
{
rgPrediction.ToArray()
}, 0);
PlotCollection plotsAvePrediction = createPlotsAve("Predicted SMA", plotsPrediction, 10);
PlotCollectionSet set = new PlotCollectionSet(new List <PlotCollection>()
{
plotsTarget, plotsPrediction, plotsAvePrediction
});
// Create the graph image and display
Image img = SimpleGraphingControl.QuickRender(set, pbImage.Width, pbImage.Height);
img = drawInput(img, sd, res.DetectedLabel, bForcedError);
bw.ReportProgress(0, img);
Thread.Sleep(1000);
m_nLabelSeq++;
if (m_nLabelSeq == 10)
{
m_nLabelSeq = 0;
}
}
}
public void Create(DatasetConfiguration config, IXDatasetCreatorProgress progress)
{
string strTrainingFile1 = Properties.Settings.Default.vocTrainingFile1;
string strTrainingFile2 = Properties.Settings.Default.vocTrainingFile2;
string strTestingFile1 = Properties.Settings.Default.vocTestingFile1;
bool bExtract = Properties.Settings.Default.ExpandFiles;
m_evtCancel.Reset();
DataConfigSetting dsTrainingFile1 = config.Settings.Find("Training Data File 2007");
strTrainingFile1 = dsTrainingFile1.Value.ToString();
DataConfigSetting dsTrainingFile2 = config.Settings.Find("Training Data File 2012");
strTrainingFile2 = dsTrainingFile2.Value.ToString();
DataConfigSetting dsTestingFile1 = config.Settings.Find("Testing Data File 2007");
strTestingFile1 = dsTestingFile1.Value.ToString();
DataConfigSetting dsExtract = config.Settings.Find("Extract Data Files");
OptionItem extractOption = dsExtract.Value as OptionItem;
bExtract = (extractOption.Index == 0) ? false : true;
DataConfigSetting dsName = config.Settings.Find("Output Dataset Name");
string strDsName = dsName.Value.ToString();
string strTrainingSrc = strDsName + ".training";
string strTestingSrc = strDsName + ".testing";
m_iprogress = progress;
m_factory.DeleteSources(strTrainingSrc, strTestingSrc);
Log log = new Log("VOC0712 Dataset Creator");
log.OnWriteLine += new EventHandler <LogArg>(log_OnWriteLine);
try
{
VOCDataParameters param = new VOCDataParameters(strTrainingFile1, strTrainingFile2, strTestingFile1, bExtract);
VOCDataLoader loader = new VOCDataLoader(param, log, m_evtCancel);
loader.OnProgress += Loader_OnProgress;
loader.OnError += Loader_OnError;
loader.OnCompleted += Loader_OnCompleted;
if (!loader.LoadDatabase(config.ID))
{
return;
}
using (DNNEntities entities = EntitiesConnection.CreateEntities())
{
List <Dataset> rgDs = entities.Datasets.Where(p => p.Name == strDsName).ToList();
List <Source> rgSrcTraining = entities.Sources.Where(p => p.Name == strTrainingSrc).ToList();
List <Source> rgSrcTesting = entities.Sources.Where(p => p.Name == strTestingSrc).ToList();
if (rgSrcTraining.Count == 0)
{
throw new Exception("Could not find the training source '" + strTrainingSrc + "'.");
}
if (rgSrcTesting.Count == 0)
{
throw new Exception("Could not find the tesing source '" + strTestingSrc + "'.");
}
int nSrcTestingCount = rgSrcTesting[0].ImageCount.GetValueOrDefault();
int nSrcTrainingCount = rgSrcTraining[0].ImageCount.GetValueOrDefault();
int nSrcTotalCount = nSrcTestingCount + nSrcTrainingCount;
double dfTestingPct = (nSrcTrainingCount == 0) ? 0.0 : nSrcTestingCount / (double)nSrcTotalCount;
Dataset ds = null;
if (rgDs.Count == 0)
{
ds = new Dataset();
ds.Name = strDsName;
}
else
{
ds = rgDs[0];
}
ds.ImageEncoded = rgSrcTesting[0].ImageEncoded;
ds.ImageChannels = rgSrcTesting[0].ImageChannels;
ds.ImageHeight = rgSrcTraining[0].ImageHeight;
ds.ImageWidth = rgSrcTraining[0].ImageWidth;
ds.TestingPercent = (decimal)dfTestingPct;
ds.TestingSourceID = rgSrcTesting[0].ID;
ds.TestingTotal = rgSrcTesting[0].ImageCount;
ds.TrainingSourceID = rgSrcTraining[0].ID;
ds.TrainingTotal = rgSrcTraining[0].ImageCount;
ds.DatasetCreatorID = config.ID;
ds.DatasetGroupID = 0;
ds.ModelGroupID = 0;
if (rgDs.Count == 0)
{
entities.Datasets.Add(ds);
}
entities.SaveChanges();
}
}
catch (Exception excpt)
{
log.WriteLine("ERROR: " + excpt.Message);
}
finally
{
if (m_evtCancel.WaitOne(0))
{
log.WriteLine("ABORTED converting VOC0712 data files.");
m_iprogress.OnCompleted(new CreateProgressArgs(1, "ABORTED!", null, true));
}
else
{
log.WriteLine("Done converting VOC0712 data files.");
m_iprogress.OnCompleted(new CreateProgressArgs(1, "COMPLETED."));
}
Properties.Settings.Default.vocTestingFile1 = strTestingFile1;
Properties.Settings.Default.vocTrainingFile1 = strTrainingFile1;
Properties.Settings.Default.vocTrainingFile2 = strTrainingFile2;
Properties.Settings.Default.ExpandFiles = bExtract;
Properties.Settings.Default.Save();
}
}
/// <summary>
/// Step the gym one step in the data.
/// </summary>
/// <param name="nAction">Specifies the action to run on the gym.</param>
/// <param name="bGetLabel">Not used.</param>
/// <param name="extraProp">Optionally, specifies extra properties.</param>
/// <returns>A tuple containing state data, the reward, and the done state is returned.</returns>
public Tuple <State, double, bool> Step(int nAction, bool bGetLabel = false, PropertySet extraProp = null)
{
DataState data = new DataState();
ScoreCollection scores = null;
if (ActivePhase == Phase.RUN)
{
if (extraProp == null)
{
throw new Exception("The extra properties are needed when querying data during the RUN phase.");
}
int nDataCount = extraProp.GetPropertyAsInt("DataCountRequested");
string strStartTime = extraProp.GetProperty("SeedTime");
int nStartIdx = m_scores.Count - nDataCount;
DateTime dt;
if (DateTime.TryParse(strStartTime, out dt))
{
nStartIdx = m_scores.FindIndexAt(dt, nDataCount);
}
scores = m_scores.CopyFrom(nStartIdx, nDataCount);
}
else
{
int nCount = 0;
m_scores = load(out m_nDim, out m_nWidth);
if (m_bRecreateData || m_scores.Count != m_ds.TrainingSource.ImageCount)
{
Stopwatch sw = new Stopwatch();
sw.Start();
m_scores = new ScoreCollection();
while (m_nCurrentIdx < m_ds.TrainingSource.ImageCount)
{
// Query images sequentially by index in batches
List <SimpleDatum> rgSd = new List <SimpleDatum>();
for (int i = 0; i < m_nBatchSize; i++)
{
SimpleDatum sd = m_imgdb.QueryImage(m_ds.TrainingSource.ID, m_nCurrentIdx + i, IMGDB_LABEL_SELECTION_METHOD.NONE, IMGDB_IMAGE_SELECTION_METHOD.NONE);
rgSd.Add(sd);
nCount++;
if (nCount == m_ds.TrainingSource.ImageCount)
{
break;
}
}
List <ResultCollection> rgRes = m_mycaffe.Run(rgSd, ref m_blobWork);
if (m_nWidth == 0)
{
m_nWidth = rgRes[0].ResultsOriginal.Count;
m_nDim = rgRes[0].ResultsOriginal.Count * 2;
}
// Fill SimpleDatum with the ordered label,score pairs starting with the detected label.
for (int i = 0; i < rgRes.Count; i++)
{
m_scores.Add(new Score(rgSd[i].TimeStamp, rgSd[i].Index, rgRes[i]));
m_nCurrentIdx++;
}
if (sw.Elapsed.TotalMilliseconds > 1000)
{
m_log.Progress = (double)m_nCurrentIdx / (double)m_ds.TrainingSource.ImageCount;
m_log.WriteLine("Running model on image " + m_nCurrentIdx.ToString() + " of " + m_ds.TrainingSource.ImageCount.ToString() + " of '" + m_strDataset + "' dataset.");
if (m_evtCancel.WaitOne(0))
{
return(null);
}
}
}
save(m_nDim, m_nWidth, m_scores);
}
else
{
m_nCurrentIdx = m_scores.Count;
}
scores = m_scores;
}
float[] rgfRes = scores.Data;
SimpleDatum sdRes = new SimpleDatum(scores.Count, m_nWidth, 2, rgfRes, 0, rgfRes.Length);
data.SetData(sdRes);
m_nCurrentIdx = 0;
return(new Tuple <State, double, bool>(data, 0, false));
}
private bool loadFile(string strImagesFile, string strSourceName, int nTotal, ref int nIdx, Log log)
{
Stopwatch sw = new Stopwatch();
int nStart = nIdx;
reportProgress(nIdx, nTotal, " Source: " + strSourceName);
reportProgress(nIdx, nTotal, " loading " + strImagesFile + "...");
FileStream fs = null;
try
{
fs = new FileStream(strImagesFile, FileMode.Open, FileAccess.Read);
using (BinaryReader br = new BinaryReader(fs))
{
fs = null;
int nSrcId = m_factory.AddSource(strSourceName, 3, 32, 32, false, 0, true);
m_factory.Open(nSrcId, 500, Database.FORCE_LOAD.NONE, log);
if (nIdx == 0)
{
m_factory.DeleteSourceData();
}
sw.Start();
for (int i = 0; i < 10000; i++)
{
int nLabel = (int)br.ReadByte();
byte[] rgImgBytes = br.ReadBytes(3072);
Bitmap img = createImage(rgImgBytes);
Datum d = ImageData.GetImageDataD(img, 3, false, nLabel);
m_factory.PutRawImageCache(nIdx, d, 5);
m_rgImg.Add(new SimpleDatum(d));
nIdx++;
if (sw.ElapsedMilliseconds > 1000)
{
reportProgress(nStart + i, nTotal, "loading " + strImagesFile + " " + i.ToString("N0") + " of 10,000...");
sw.Restart();
}
if (m_evtCancel.WaitOne(0))
{
return(false);
}
}
m_factory.ClearImageCache(true);
if (nIdx == nTotal)
{
m_factory.UpdateSourceCounts();
}
}
}
finally
{
if (fs != null)
{
fs.Dispose();
}
}
return(true);
}
/// <summary>
/// The query thread is where all data is collected from the underlying custom query managed.
/// </summary>
private void queryThread()
{
try
{
int nWait = 0;
m_iquery.Open();
while (!m_evtCancel.WaitOne(nWait))
{
if (!m_evtQueryEnabled.WaitOne(0))
{
m_evtPaused.Set();
nWait = 250;
continue;
}
m_evtPaused.Reset();
if (m_rgDataQueue.Count >= m_nMaxCount)
{
nWait = 10;
continue;
}
double[] rgData = m_iquery.QueryByTime(m_dt, m_tsInc, m_nSegmentSize);
if (rgData == null)
{
m_bQueryEnd = true;
nWait = 10;
continue;
}
nWait = 0;
m_bQueryEnd = false;
int nItemCount = rgData.Length / m_nSegmentSize;
int nSrcIdx = 0;
lock (m_objSync)
{
for (int i = 0; i < m_nSegmentSize; i++)
{
double[] rgItem = new double[nItemCount];
Array.Copy(rgData, nSrcIdx, rgItem, 0, nItemCount);
nSrcIdx += nItemCount;
m_rgDataQueue.Enqueue(rgItem);
}
}
m_dt += TimeSpan.FromMilliseconds(m_nSegmentSize * m_tsInc.TotalMilliseconds);
}
}
catch (Exception excpt)
{
throw excpt;
}
finally
{
m_iquery.Close();
}
}
private bool loadFile(DatasetFactory factory, List <Tuple <byte[], int> > rgData, int nC, int nH, int nW, string strSourceName, string strExportPath)
{
if (strExportPath != null)
{
strExportPath += strSourceName;
if (!Directory.Exists(strExportPath))
{
Directory.CreateDirectory(strExportPath);
}
}
Stopwatch sw = new Stopwatch();
reportProgress(0, 0, " Source: " + strSourceName);
try
{
if (factory != null)
{
int nSrcId = factory.AddSource(strSourceName, nC, nW, nH, false, 0, true);
factory.Open(nSrcId, 500, Database.FORCE_LOAD.NONE, m_log);
factory.DeleteSourceData();
}
// Storing to database;
int nLabel;
byte[] rgPixels;
Datum datum = new Datum(false, nC, nW, nH, -1, DateTime.MinValue, new List <byte>(), 0, false, -1);
string strAction = (m_param.ExportToFile) ? "exporing" : "loading";
reportProgress(0, rgData.Count, " " + strAction + " a total of " + rgData.Count.ToString() + " items.");
reportProgress(0, rgData.Count, " (with rows: " + nH.ToString() + ", cols: " + nW.ToString() + ")");
sw.Start();
List <SimpleDatum> rgImg = new List <SimpleDatum>();
FileStream fsFileDesc = null;
StreamWriter swFileDesc = null;
if (m_param.ExportToFile)
{
string strFile = strExportPath + "\\file_list.txt";
fsFileDesc = File.OpenWrite(strFile);
swFileDesc = new StreamWriter(fsFileDesc);
}
for (int i = 0; i < rgData.Count; i++)
{
rgPixels = rgData[i].Item1;
nLabel = rgData[i].Item2;
if (sw.Elapsed.TotalMilliseconds > 1000)
{
reportProgress(i, rgData.Count, " " + strAction + " data...");
sw.Restart();
}
datum.SetData(rgPixels, nLabel);
if (factory != null)
{
factory.PutRawImageCache(i, datum, 5);
}
else if (strExportPath != null)
{
saveToFile(strExportPath, i, datum, swFileDesc);
}
rgImg.Add(new SimpleDatum(datum));
if (m_evtCancel.WaitOne(0))
{
return(false);
}
}
if (swFileDesc != null)
{
swFileDesc.Flush();
swFileDesc.Close();
swFileDesc.Dispose();
fsFileDesc.Close();
fsFileDesc.Dispose();
}
if (factory != null)
{
factory.ClearImageCache(true);
factory.UpdateSourceCounts();
factory.SaveImageMean(SimpleDatum.CalculateMean(m_log, rgImg.ToArray(), new WaitHandle[] { new ManualResetEvent(false) }), true);
}
reportProgress(rgData.Count, rgData.Count, " " + strAction + " completed.");
}
finally
{
}
return(true);
}
/// <summary>
/// Process the content image by applying the style to it that was learned from the style image.
/// </summary>
/// <param name="bmpStyle">Specifies the image used to train the what style to apply to the content.</param>
/// <param name="bmpContent">Specifies the content image to which the style is to be applied.</param>
/// <param name="nIterations">Specifies the number of training iterations.</param>
/// <param name="strResultDir">Optionally, specifies an output directory where intermediate images are stored.</param>
/// <param name="nIntermediateOutput">Optionally, specifies how often to output an intermediate image.</param>
/// <param name="dfTvLoss">Optionally, specifies the TV-Loss weight for smoothing (default = 0, which disables this loss).</param>
/// <returns>The resulting image is returned.</returns>
public Bitmap Process(Bitmap bmpStyle, Bitmap bmpContent, int nIterations, string strResultDir = null, int nIntermediateOutput = -1, double dfTvLoss = 0)
{
Solver <T> solver = null;
Net <T> net = null;
BlobCollection <T> colContentActivations = new BlobCollection <T>();
BlobCollection <T> colGramActivations = new BlobCollection <T>();
double dfLoss;
try
{
m_dfTVLossWeight = dfTvLoss;
m_nIterations = nIterations;
if (bmpStyle.Width != bmpContent.Width ||
bmpStyle.Height != bmpContent.Height)
{
bmpStyle = ImageTools.ResizeImage(bmpStyle, bmpContent.Width, bmpContent.Height);
}
m_log.WriteLine("Creating input network...");
m_log.Enable = false;
net = new Net <T>(m_cuda, m_log, m_param, m_evtCancel, null, Phase.TEST);
m_log.Enable = true;
if (m_rgWeights != null)
{
net.LoadWeights(m_rgWeights, m_persist);
}
//-----------------------------------------
// Get style and content activations.
//-----------------------------------------
prepare_data_blob(net, bmpStyle);
net.Forward(out dfLoss);
foreach (KeyValuePair <string, double> kvGram in m_rgLayers["gram"])
{
string strGram = kvGram.Key;
Blob <T> blobGram = net.blob_by_name(strGram);
colGramActivations.Add(blobGram.Clone());
}
prepare_data_blob(net, bmpContent);
net.Forward(out dfLoss);
foreach (KeyValuePair <string, double> kvContent in m_rgLayers["content"])
{
string strContent = kvContent.Key;
Blob <T> blobContent = net.blob_by_name(strContent);
colContentActivations.Add(blobContent.Clone());
}
//-----------------------------------------
// Prepare the network by adding new layers.
//-----------------------------------------
NetParameter net_param = m_param;
foreach (KeyValuePair <string, double> kvInput in m_rgLayers["input"])
{
string strName = kvInput.Key;
LayerParameter p = new LayerParameter(LayerParameter.LayerType.INPUT);
p.name = "input_" + strName;
p.top.Add(p.name);
Blob <T> blob = net.blob_by_name(strName);
p.input_param.shape.Add(new BlobShape(blob.shape()));
net_param.layer.Add(p);
}
foreach (KeyValuePair <string, double> kvContent in m_rgLayers["content"])
{
string strName = kvContent.Key;
string strScale1 = "input_" + strName;
string strScale2 = strName;
if (m_dfContentDataScale != 1.0)
{
strScale1 += "b";
LayerParameter ps1 = new LayerParameter(LayerParameter.LayerType.SCALAR);
ps1.scalar_param.value = m_dfContentDataScale;
ps1.scalar_param.operation = ScalarParameter.ScalarOp.MUL;
ps1.scalar_param.passthrough_gradient = true;
ps1.bottom.Add("input_" + strName);
ps1.top.Add(strScale1);
net_param.layer.Add(ps1);
strScale2 += "b";
LayerParameter ps2 = new LayerParameter(LayerParameter.LayerType.SCALAR);
ps2.scalar_param.value = m_dfContentDataScale;
ps2.scalar_param.operation = ScalarParameter.ScalarOp.MUL;
ps2.scalar_param.passthrough_gradient = true;
ps2.bottom.Add(strName);
ps2.top.Add(strScale2);
net_param.layer.Add(ps2);
}
LayerParameter event_param = new LayerParameter(LayerParameter.LayerType.EVENT);
event_param.name = "event_" + strName;
event_param.bottom.Add(strScale2);
event_param.bottom.Add(strScale1);
event_param.top.Add("event_" + strName);
net_param.layer.Add(event_param);
LayerParameter p = new LayerParameter(LayerParameter.LayerType.EUCLIDEAN_LOSS);
p.name = "loss_" + strName;
Blob <T> blobContent = colContentActivations[strName];
double dfScale = get_content_scale(blobContent);
p.loss_weight.Add(kvContent.Value * dfScale);
p.bottom.Add("event_" + strName);
p.bottom.Add(strScale1);
p.top.Add("loss_" + strName);
net_param.layer.Add(p);
}
foreach (KeyValuePair <string, double> kvGram in m_rgLayers["gram"].ToList())
{
string strGramName = kvGram.Key;
LayerParameter event_param = new LayerParameter(LayerParameter.LayerType.EVENT);
event_param.name = "event_" + strGramName;
event_param.bottom.Add(strGramName);
event_param.bottom.Add("input_" + strGramName);
event_param.top.Add("event_" + strGramName);
net_param.layer.Add(event_param);
LayerParameter p = new LayerParameter(LayerParameter.LayerType.EUCLIDEAN_LOSS);
p.name = "loss_" + strGramName;
Blob <T> blobGram = colGramActivations[strGramName];
double dfScale = get_style_scale(blobGram);
p.loss_weight.Add(kvGram.Value * dfScale);
p.bottom.Add("input_" + strGramName);
p.bottom.Add("event_" + strGramName);
p.top.Add("loss_" + strGramName);
net_param.layer.Add(p);
}
// Add TV Loss;
if (m_dfTVLossWeight != 0)
{
LayerParameter p = new LayerParameter(LayerParameter.LayerType.TV_LOSS);
p.name = "loss_tv";
double dfWeight = m_dfTVLossWeight;
p.loss_weight.Add(dfWeight);
p.bottom.Add("data");
p.top.Add("loss_tv");
net_param.layer.Add(p);
}
// Replace InputLayer with ParameterLayer,
// so that we'll be able to backprop into the image.
Blob <T> data = net.blob_by_name("data");
for (int i = 0; i < net_param.layer.Count; i++)
{
LayerParameter p = net_param.layer[i];
if (p.name == "input1")
{
net_param.layer[i].SetType(LayerParameter.LayerType.PARAMETER);
net_param.layer[i].parameter_param.shape = new BlobShape(data.shape());
break;
}
}
// Disable weights learning.
List <LayerParameter.LayerType> rgTypes = new List <LayerParameter.LayerType>();
rgTypes.Add(LayerParameter.LayerType.CONVOLUTION);
rgTypes.Add(LayerParameter.LayerType.DECONVOLUTION);
rgTypes.Add(LayerParameter.LayerType.INNERPRODUCT);
rgTypes.Add(LayerParameter.LayerType.PRELU);
rgTypes.Add(LayerParameter.LayerType.BIAS);
rgTypes.Add(LayerParameter.LayerType.EMBED);
rgTypes.Add(LayerParameter.LayerType.LSTM);
rgTypes.Add(LayerParameter.LayerType.LSTM_SIMPLE);
rgTypes.Add(LayerParameter.LayerType.RNN);
foreach (LayerParameter layer in net_param.layer)
{
if (rgTypes.Contains(layer.type))
{
layer.parameters = new List <ParamSpec>();
layer.parameters.Add(new ParamSpec(0, 0));
layer.parameters.Add(new ParamSpec(0, 0));
}
}
net.Dispose();
net = null;
//-----------------------------------------
// Create solver and assign inputs.
//-----------------------------------------
RawProto proto1 = net_param.ToProto("root");
string str = proto1.ToString();
SolverParameter solver_param = new SolverParameter();
solver_param.display = m_nDisplayEvery;
solver_param.train_net_param = net_param;
solver_param.test_iter.Clear();
solver_param.test_interval = 0;
solver_param.test_initialization = false;
solver_param.base_lr = m_dfLearningRate;
solver_param.type = m_solverType;
m_log.WriteLine("Creating " + m_solverType.ToString() + " solver with learning rate = " + m_dfLearningRate.ToString() + "...");
m_log.Enable = false;
if (m_solverType == SolverParameter.SolverType.LBFGS)
{
solver = new LBFGSSolver <T>(m_cuda, m_log, solver_param, m_evtCancel, null, null, null, m_persist);
}
else
{
solver = Solver <T> .Create(m_cuda, m_log, solver_param, m_evtCancel, null, null, null, m_persist);
}
m_log.Enable = true;
solver.OnSnapshot += Solver_OnSnapshot;
solver.OnTrainingIteration += Solver_OnTrainingIteration;
foreach (Layer <T> layer in solver.net.layers)
{
if (layer.type == LayerParameter.LayerType.EVENT)
{
EventLayer <T> eventLayer = layer as EventLayer <T>;
eventLayer.OnBackward += EventLayer_OnBackward;
}
}
prepare_input_param(solver.net, bmpContent);
foreach (KeyValuePair <string, double> kvContent in m_rgLayers["content"])
{
string strName = kvContent.Key;
Blob <T> blobDst = solver.net.blob_by_name("input_" + strName);
Blob <T> blobSrc = colContentActivations[strName];
blobDst.CopyFrom(blobSrc);
}
foreach (KeyValuePair <string, double> kvGram in m_rgLayers["gram"])
{
string strName = kvGram.Key;
Blob <T> blobDst = solver.net.blob_by_name("input_" + strName);
Blob <T> blobSrc = colGramActivations[strName];
blobDst.CopyFrom(blobSrc);
}
//-----------------------------------------
// Optimize.
//-----------------------------------------
int nIterations1 = m_nIterations;
if (strResultDir != null && nIntermediateOutput > 0)
{
nIterations1 /= nIntermediateOutput;
}
if (m_rgWeights != null)
{
Blob <T> blobInput = solver.net.learnable_parameters[0];
solver.net.learnable_parameters.RemoveAt(0);
solver.net.LoadWeights(m_rgWeights, m_persist);
solver.net.learnable_parameters.Insert(0, blobInput);
}
if (strResultDir != null)
{
strResultDir = strResultDir.TrimEnd('\\');
strResultDir += "\\";
}
for (int i = 0; i < nIterations1; i++)
{
if (m_evtCancel.WaitOne(0))
{
break;
}
solver.Step(nIntermediateOutput, TRAIN_STEP.NONE, true, true, true);
if (strResultDir != null)
{
Bitmap bmpTemp = save(solver.net);
string strFile = strResultDir + i.ToString() + "_temp.png";
if (File.Exists(strFile))
{
File.Delete(strFile);
}
bmpTemp.Save(strFile);
}
}
Bitmap bmpOutput = save(solver.net);
return(bmpOutput);
}
catch (Exception excpt)
{
throw excpt;
}
finally
{
if (net != null)
{
net.Dispose();
}
if (solver != null)
{
solver.Dispose();
}
colGramActivations.Dispose();
colContentActivations.Dispose();
}
}
