Merge pull request intel#120 from sguada/images_layer

Images layer: A data provider layer directly from images

include/caffe/vision_layers.hpp

@@ -377,6 +377,44 @@ class DataLayer : public Layer<Dtype> {
   Blob<Dtype> data_mean_;
 };
 
+// This function is used to create a pthread that prefetches the data.
+template <typename Dtype>
+void* ImagesLayerPrefetch(void* layer_pointer);
+
+template <typename Dtype>
+class ImagesLayer : public Layer<Dtype> {
+  // The function used to perform prefetching.
+ friend void* ImagesLayerPrefetch<Dtype>(void* layer_pointer);
+
+ public:
+  explicit ImagesLayer(const LayerParameter& param)
+      : Layer<Dtype>(param) {}
+  virtual ~ImagesLayer();
+  virtual void SetUp(const vector<Blob<Dtype>*>& bottom,
+      vector<Blob<Dtype>*>* top);
+
+ protected:
+  virtual void Forward_cpu(const vector<Blob<Dtype>*>& bottom,
+      vector<Blob<Dtype>*>* top);
+  virtual void Forward_gpu(const vector<Blob<Dtype>*>& bottom,
+      vector<Blob<Dtype>*>* top);
+  virtual Dtype Backward_cpu(const vector<Blob<Dtype>*>& top,
+      const bool propagate_down, vector<Blob<Dtype>*>* bottom);
+  virtual Dtype Backward_gpu(const vector<Blob<Dtype>*>& top,
+      const bool propagate_down, vector<Blob<Dtype>*>* bottom);
+
+  vector<std::pair<std::string, int> > lines_;
+  int lines_id_;
+  int datum_channels_;
+  int datum_height_;
+  int datum_width_;
+  int datum_size_;
+  pthread_t thread_;
+  shared_ptr<Blob<Dtype> > prefetch_data_;
+  shared_ptr<Blob<Dtype> > prefetch_label_;
+  Blob<Dtype> data_mean_;
+};
+
 
 template <typename Dtype>
 class SoftmaxLayer : public Layer<Dtype> {

src/caffe/layer_factory.cpp

@@ -27,6 +27,8 @@ Layer<Dtype>* GetLayer(const LayerParameter& param) {
     return new ConvolutionLayer<Dtype>(param);
   } else if (type == "data") {
     return new DataLayer<Dtype>(param);
+  } else if (type == "images") {
+    return new ImagesLayer<Dtype>(param);
   } else if (type == "dropout") {
     return new DropoutLayer<Dtype>(param);
   } else if (type == "euclidean_loss") {

src/caffe/layers/images_layer.cpp

@@ -0,0 +1,274 @@
+// Copyright 2013 Yangqing Jia
+
+#include <stdint.h>
+#include <leveldb/db.h>
+#include <pthread.h>
+
+#include <string>
+#include <vector>
+#include <iostream>
+#include <fstream>
+
+#include "caffe/layer.hpp"
+#include "caffe/util/io.hpp"
+#include "caffe/vision_layers.hpp"
+
+using std::string;
+using std::pair;
+
+namespace caffe {
+
+template <typename Dtype>
+void* ImagesLayerPrefetch(void* layer_pointer) {
+  CHECK(layer_pointer);
+  ImagesLayer<Dtype>* layer = reinterpret_cast<ImagesLayer<Dtype>*>(layer_pointer);
+  CHECK(layer);
+  Datum datum;
+  CHECK(layer->prefetch_data_);
+  Dtype* top_data = layer->prefetch_data_->mutable_cpu_data();
+  Dtype* top_label = layer->prefetch_label_->mutable_cpu_data();
+  const Dtype scale = layer->layer_param_.scale();
+  const int batchsize = layer->layer_param_.batchsize();
+  const int cropsize = layer->layer_param_.cropsize();
+  const bool mirror = layer->layer_param_.mirror();
+  const int new_height  = layer->layer_param_.new_height();
+  const int new_width  = layer->layer_param_.new_height();
+
+  if (mirror && cropsize == 0) {
+    LOG(FATAL) << "Current implementation requires mirror and cropsize to be "
+        << "set at the same time.";
+  }
+  // datum scales
+  const int channels = layer->datum_channels_;
+  const int height = layer->datum_height_;
+  const int width = layer->datum_width_;
+  const int size = layer->datum_size_;  
+  const int lines_size = layer->lines_.size();
+  const Dtype* mean = layer->data_mean_.cpu_data();
+  for (int itemid = 0; itemid < batchsize; ++itemid) {
+    // get a blob
+    CHECK_GT(lines_size,layer->lines_id_);
+    if (!ReadImageToDatum(layer->lines_[layer->lines_id_].first, layer->lines_[layer->lines_id_].second, 
+      new_height, new_width, &datum)) {
+      continue;
+    };    
+    const string& data = datum.data();
+    if (cropsize) {
+      CHECK(data.size()) << "Image cropping only support uint8 data";
+      int h_off, w_off;
+      // We only do random crop when we do training.
+      if (Caffe::phase() == Caffe::TRAIN) {
+        h_off = rand() % (height - cropsize);
+        w_off = rand() % (width - cropsize);
+      } else {
+        h_off = (height - cropsize) / 2;
+        w_off = (width - cropsize) / 2;
+      }
+      if (mirror && rand() % 2) {
+        // Copy mirrored version
+        for (int c = 0; c < channels; ++c) {
+          for (int h = 0; h < cropsize; ++h) {
+            for (int w = 0; w < cropsize; ++w) {
+              top_data[((itemid * channels + c) * cropsize + h) * cropsize
+                       + cropsize - 1 - w] =
+                  (static_cast<Dtype>(
+                      (uint8_t)data[(c * height + h + h_off) * width
+                                    + w + w_off])
+                    - mean[(c * height + h + h_off) * width + w + w_off])
+                  * scale;
+            }
+          }
+        }
+      } else {
+        // Normal copy
+        for (int c = 0; c < channels; ++c) {
+          for (int h = 0; h < cropsize; ++h) {
+            for (int w = 0; w < cropsize; ++w) {
+              top_data[((itemid * channels + c) * cropsize + h) * cropsize + w]
+                  = (static_cast<Dtype>(
+                      (uint8_t)data[(c * height + h + h_off) * width
+                                    + w + w_off])
+                     - mean[(c * height + h + h_off) * width + w + w_off])
+                  * scale;
+            }
+          }
+        }
+      }
+    } else {
+      // Just copy the whole data
+      if (data.size()) {
+        for (int j = 0; j < size; ++j) {
+          top_data[itemid * size + j] =
+              (static_cast<Dtype>((uint8_t)data[j]) - mean[j]) * scale;
+        }
+      } else {
+        for (int j = 0; j < size; ++j) {
+          top_data[itemid * size + j] =
+              (datum.float_data(j) - mean[j]) * scale;
+        }
+      }
+    }
+
+    top_label[itemid] = datum.label();
+    // go to the next iter
+    layer->lines_id_++;
+    if (layer->lines_id_ >= lines_size) {
+      // We have reached the end. Restart from the first.
+      DLOG(INFO) << "Restarting data prefetching from start.";
+      layer->lines_id_=0;
+      if (layer->layer_param_.shuffle_images()) {
+        std::random_shuffle(layer->lines_.begin(), layer->lines_.end());
+      }
+    }
+  }
+
+  return (void*)NULL;
+}
+
+template <typename Dtype>
+ImagesLayer<Dtype>::~ImagesLayer<Dtype>() {
+  // Finally, join the thread
+  CHECK(!pthread_join(thread_, NULL)) << "Pthread joining failed.";
+}
+
+template <typename Dtype>
+void ImagesLayer<Dtype>::SetUp(const vector<Blob<Dtype>*>& bottom,
+      vector<Blob<Dtype>*>* top) {
+  CHECK_EQ(bottom.size(), 0) << "Input Layer takes no input blobs.";
+  CHECK_EQ(top->size(), 2) << "Input Layer takes two blobs as output.";
+  const int new_height  = this->layer_param_.new_height();
+  const int new_width  = this->layer_param_.new_height();
+  CHECK((new_height==0 && new_width==0)||(new_height>0 && new_width > 0)) << 
+  "Current implementation requires new_height and new_width to be set at the same time.";
+  // Read the file with filenames and labels
+  LOG(INFO) << "Opening file " << this->layer_param_.source();
+  std::ifstream infile(this->layer_param_.source().c_str());  
+  string filename;
+  int label;
+  while (infile >> filename >> label) {
+    lines_.push_back(std::make_pair(filename, label));
+  }
+
+  if (this->layer_param_.shuffle_images()) {
+    // randomly shuffle data
+    LOG(INFO) << "Shuffling data";
+    std::random_shuffle(lines_.begin(), lines_.end());
+  }
+  LOG(INFO) << "A total of " << lines_.size() << " images.";
+
+  lines_id_ = 0;
+  // Check if we would need to randomly skip a few data points
+  if (this->layer_param_.rand_skip()) {
+    unsigned int skip = rand() % this->layer_param_.rand_skip();
+    LOG(INFO) << "Skipping first " << skip << " data points.";
+    CHECK_GT(lines_.size(),skip) << "Not enought points to skip";
+    lines_id_ = skip;
+  }
+  // Read a data point, and use it to initialize the top blob.
+  Datum datum;
+  CHECK(ReadImageToDatum(lines_[lines_id_].first, lines_[lines_id_].second,
+                         new_height,new_width,&datum));
+  // image
+  int cropsize = this->layer_param_.cropsize();
+  if (cropsize > 0) {
+    (*top)[0]->Reshape(
+        this->layer_param_.batchsize(), datum.channels(), cropsize, cropsize);
+    prefetch_data_.reset(new Blob<Dtype>(
+        this->layer_param_.batchsize(), datum.channels(), cropsize, cropsize));
+  } else {
+    (*top)[0]->Reshape(
+        this->layer_param_.batchsize(), datum.channels(), datum.height(),
+        datum.width());
+    prefetch_data_.reset(new Blob<Dtype>(
+        this->layer_param_.batchsize(), datum.channels(), datum.height(),
+        datum.width()));
+  }
+  LOG(INFO) << "output data size: " << (*top)[0]->num() << ","
+      << (*top)[0]->channels() << "," << (*top)[0]->height() << ","
+      << (*top)[0]->width();
+  // label
+  (*top)[1]->Reshape(this->layer_param_.batchsize(), 1, 1, 1);
+  prefetch_label_.reset(
+      new Blob<Dtype>(this->layer_param_.batchsize(), 1, 1, 1));
+  // datum size
+  datum_channels_ = datum.channels();
+  datum_height_ = datum.height();
+  datum_width_ = datum.width();
+  datum_size_ = datum.channels() * datum.height() * datum.width();
+  CHECK_GT(datum_height_, cropsize);
+  CHECK_GT(datum_width_, cropsize);
+  // check if we want to have mean
+  if (this->layer_param_.has_meanfile()) {
+    BlobProto blob_proto;
+    LOG(INFO) << "Loading mean file from" << this->layer_param_.meanfile();
+    ReadProtoFromBinaryFile(this->layer_param_.meanfile().c_str(), &blob_proto);
+    data_mean_.FromProto(blob_proto);
+    CHECK_EQ(data_mean_.num(), 1);
+    CHECK_EQ(data_mean_.channels(), datum_channels_);
+    CHECK_EQ(data_mean_.height(), datum_height_);
+    CHECK_EQ(data_mean_.width(), datum_width_);
+  } else {
+    // Simply initialize an all-empty mean.
+    data_mean_.Reshape(1, datum_channels_, datum_height_, datum_width_);
+  }
+  // Now, start the prefetch thread. Before calling prefetch, we make two
+  // cpu_data calls so that the prefetch thread does not accidentally make
+  // simultaneous cudaMalloc calls when the main thread is running. In some
+  // GPUs this seems to cause failures if we do not so.
+  prefetch_data_->mutable_cpu_data();
+  prefetch_label_->mutable_cpu_data();
+  data_mean_.cpu_data();
+  DLOG(INFO) << "Initializing prefetch";
+  CHECK(!pthread_create(&thread_, NULL, ImagesLayerPrefetch<Dtype>,
+      reinterpret_cast<void*>(this))) << "Pthread execution failed.";
+  DLOG(INFO) << "Prefetch initialized.";
+}
+
+template <typename Dtype>
+void ImagesLayer<Dtype>::Forward_cpu(const vector<Blob<Dtype>*>& bottom,
+      vector<Blob<Dtype>*>* top) {
+  // First, join the thread
+  CHECK(!pthread_join(thread_, NULL)) << "Pthread joining failed.";
+  // Copy the data
+  memcpy((*top)[0]->mutable_cpu_data(), prefetch_data_->cpu_data(),
+      sizeof(Dtype) * prefetch_data_->count());
+  memcpy((*top)[1]->mutable_cpu_data(), prefetch_label_->cpu_data(),
+      sizeof(Dtype) * prefetch_label_->count());
+  // Start a new prefetch thread
+  CHECK(!pthread_create(&thread_, NULL, ImagesLayerPrefetch<Dtype>,
+      reinterpret_cast<void*>(this))) << "Pthread execution failed.";
+}
+
+template <typename Dtype>
+void ImagesLayer<Dtype>::Forward_gpu(const vector<Blob<Dtype>*>& bottom,
+      vector<Blob<Dtype>*>* top) {
+  // First, join the thread
+  CHECK(!pthread_join(thread_, NULL)) << "Pthread joining failed.";
+  // Copy the data
+  CUDA_CHECK(cudaMemcpy((*top)[0]->mutable_gpu_data(),
+      prefetch_data_->cpu_data(), sizeof(Dtype) * prefetch_data_->count(),
+      cudaMemcpyHostToDevice));
+  CUDA_CHECK(cudaMemcpy((*top)[1]->mutable_gpu_data(),
+      prefetch_label_->cpu_data(), sizeof(Dtype) * prefetch_label_->count(),
+      cudaMemcpyHostToDevice));
+  // Start a new prefetch thread
+  CHECK(!pthread_create(&thread_, NULL, ImagesLayerPrefetch<Dtype>,
+      reinterpret_cast<void*>(this))) << "Pthread execution failed.";
+}
+
+// The backward operations are dummy - they do not carry any computation.
+template <typename Dtype>
+Dtype ImagesLayer<Dtype>::Backward_cpu(const vector<Blob<Dtype>*>& top,
+      const bool propagate_down, vector<Blob<Dtype>*>* bottom) {
+  return Dtype(0.);
+}
+
+template <typename Dtype>
+Dtype ImagesLayer<Dtype>::Backward_gpu(const vector<Blob<Dtype>*>& top,
+      const bool propagate_down, vector<Blob<Dtype>*>* bottom) {
+  return Dtype(0.);
+}
+
+INSTANTIATE_CLASS(ImagesLayer);
+
+}  // namespace caffe

src/caffe/proto/caffe.proto

@@ -91,6 +91,17 @@ message LayerParameter {
   // point would be set as rand_skip * rand(0,1). Note that rand_skip should not
   // be larger than the number of keys in the leveldb.
   optional uint32 rand_skip = 53 [ default = 0 ];
+
+  // For the Reshape Layer one need to specify the new dimensions
+  optional int32 new_num = 60 [default = 0];
+  optional int32 new_channels = 61 [default = 0];
+  optional int32 new_height = 62 [default = 0];
+  optional int32 new_width = 63 [default = 0];
+
+  // Used by ImageLayer to shuffle the list of files at every epoch it will also
+  // resize images if new_height or new_width are not zero
+  optional bool shuffle_images = 64 [default = false];
+
 }
 
 message LayerConnection {