Scorer model loading (#860)

* Add MMAP as an option
* Use io::isBin
* Allow getYamlFromModel from an Item vector
* ScorerWrapper can now load on to a graph from Item vector
The interface IEncoderDecoder can now call graph loads directly from an
Item Vector.
* Translator loads model before creating scorers
Scorers are created from an Item vector
* Replace model-config try-catch with check using IsNull
* Prefer empty vs size
* load by items should be pure virtual
* Stepwise forward load to encdec
* nematus can load from items
* amun can load from items
* loadItems in TranslateService
* Remove logging
* Remove by filename scorer functions
* Replace by filename createScorer
* Explicitly provide default value for get model-mmap
* CLI option for model-mmap only for translation and CPU compile
* Ensure model-mmap option is CPU only
* Remove move on temporary object
* Reinstate log messages for model loading in Amun / Nematus
* Add log messages for model loading in scorers

Co-authored-by: Roman Grundkiewicz <rgrundkiewicz@gmail.com>

CHANGELOG.md

@@ -27,6 +27,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 - Dynamic gradient-scaling with `--dynamic-gradient-scaling`.
 - Add unit tests for binary files.
 - Fix compilation with OMP
+- Added `--model-mmap` option to enable mmap loading for CPU-based translation
 - Compute aligned memory sizes using exact sizing
 - Support for loading lexical shortlist from a binary blob
 - Integrate a shortlist converter (which can convert a text lexical shortlist to a binary shortlist) into marian-conv with --shortlist option

src/common/config_parser.cpp

@@ -183,7 +183,12 @@ void ConfigParser::addOptionsModel(cli::CLIWrapper& cli) {
         "Path prefix for pre-trained model to initialize model weights");
     }
   }
-
+#ifdef COMPILE_CPU
+  if(mode_ == cli::mode::translation) {
+    cli.add<bool>("--model-mmap",
+      "Use memory-mapping when loading model (CPU only)");
+  }
+#endif
   cli.add<bool>("--ignore-model-config",
       "Ignore the model configuration saved in npz file");
   cli.add<std::string>("--type",

src/common/config_validator.cpp

@@ -54,6 +54,9 @@ void ConfigValidator::validateOptionsTranslation() const {
   ABORT_IF(models.empty() && configs.empty(),
            "You need to provide at least one model file or a config file");
 
+  ABORT_IF(get<bool>("model-mmap") && get<size_t>("cpu-threads") == 0,
+           "Model MMAP is CPU-only, please use --cpu-threads");
+
   for(const auto& modelFile : models) {
     filesystem::Path modelPath(modelFile);
     ABORT_IF(!filesystem::exists(modelPath), "Model file does not exist: " + modelFile);

src/common/io.cpp

@@ -56,6 +56,18 @@ void getYamlFromModel(YAML::Node& yaml,
     yaml = YAML::Load(item.data());
 }
 
+// Load YAML from item
+void getYamlFromModel(YAML::Node& yaml,
+                      const std::string& varName,
+                      const std::vector<Item>& items) {
+    for(auto& item : items) {
+      if(item.name == varName) {
+        yaml = YAML::Load(item.data());
+        return;
+      }
+    }
+}
+
 void addMetaToItems(const std::string& meta,
                     const std::string& varName,
                     std::vector<io::Item>& items) {

src/common/io.h

@@ -21,6 +21,7 @@ bool isBin(const std::string& fileName);
 
 void getYamlFromModel(YAML::Node& yaml, const std::string& varName, const std::string& fileName);
 void getYamlFromModel(YAML::Node& yaml, const std::string& varName, const void* ptr);
+void getYamlFromModel(YAML::Node& yaml, const std::string& varName, const std::vector<Item>& items);
 
 void addMetaToItems(const std::string& meta,
                     const std::string& varName,

src/graph/expression_graph.h

@@ -739,7 +739,7 @@ public:
 
 public:
   /** Load model (mainly parameter objects) from array of io::Items */
-  void load(std::vector<io::Item>& ioItems, bool markReloaded = true) {
+  void load(const std::vector<io::Item>& ioItems, bool markReloaded = true) {
     setReloaded(false);
     for(auto& item : ioItems) {
       std::string pName = item.name;

src/models/amun.h

@@ -36,7 +36,7 @@ public:
   }
 
   void load(Ptr<ExpressionGraph> graph,
-            const std::string& name,
+            const std::vector<io::Item>& items,
             bool /*markedReloaded*/ = true) override {
     std::map<std::string, std::string> nameMap
         = {{"decoder_U", "decoder_cell1_U"},
@@ -89,9 +89,7 @@ public:
     if(opt<bool>("tied-embeddings-src") || opt<bool>("tied-embeddings-all"))
       nameMap["Wemb"] = "Wemb";
 
-    LOG(info, "Loading model from {}", name);
-    // load items from .npz file
-    auto ioItems = io::loadItems(name);
+    auto ioItems = items;
     // map names and remove a dummy matrices
     for(auto it = ioItems.begin(); it != ioItems.end();) {
       // for backwards compatibility, turn one-dimensional vector into two dimensional matrix with first dimension being 1 and second dimension of the original size
@@ -120,6 +118,14 @@ public:
     graph->load(ioItems);
   }
 
+  void load(Ptr<ExpressionGraph> graph,
+            const std::string& name,
+            bool /*markReloaded*/ = true) override {
+    LOG(info, "Loading model from {}", name);
+    auto ioItems = io::loadItems(name);
+    load(graph, ioItems);
+  }
+
   void save(Ptr<ExpressionGraph> graph,
             const std::string& name,
             bool saveTranslatorConfig = false) override {

src/models/costs.h

@@ -325,6 +325,12 @@ protected:
 public:
   Stepwise(Ptr<IEncoderDecoder> encdec, Ptr<ILogProbStep> cost) : encdec_(encdec), cost_(cost) {}
 
+  virtual void load(Ptr<ExpressionGraph> graph,
+                    const std::vector<io::Item>& items,
+                    bool markedReloaded = true) override {
+    encdec_->load(graph, items, markedReloaded);
+  }
+
   virtual void load(Ptr<ExpressionGraph> graph,
                     const std::string& name,
                     bool markedReloaded = true) override {

src/models/encoder_decoder.cpp

@@ -144,6 +144,12 @@ std::string EncoderDecoder::getModelParametersAsString() {
   return std::string(out.c_str());
 }
 
+void EncoderDecoder::load(Ptr<ExpressionGraph> graph,
+                          const std::vector<io::Item>& items,
+                          bool markedReloaded) {
+  graph->load(items, markedReloaded && !opt<bool>("ignore-model-config", false));
+}
+
 void EncoderDecoder::load(Ptr<ExpressionGraph> graph,
                           const std::string& name,
                           bool markedReloaded) {

src/models/encoder_decoder.h

@@ -12,6 +12,11 @@ namespace marian {
 class IEncoderDecoder : public models::IModel {
 public:
   virtual ~IEncoderDecoder() {}
+
+  virtual void load(Ptr<ExpressionGraph> graph,
+                    const std::vector<io::Item>& items,
+                    bool markedReloaded = true) = 0;
+
   virtual void load(Ptr<ExpressionGraph> graph,
                     const std::string& name,
                     bool markedReloaded = true) override
@@ -91,6 +96,10 @@ public:
 
   void push_back(Ptr<DecoderBase> decoder);
 
+  virtual void load(Ptr<ExpressionGraph> graph,
+                    const std::vector<io::Item>& items,
+                    bool markedReloaded = true) override;
+
   virtual void load(Ptr<ExpressionGraph> graph,
                     const std::string& name,
                     bool markedReloaded = true) override;

src/models/nematus.h

@@ -26,11 +26,9 @@ public:
   }
 
   void load(Ptr<ExpressionGraph> graph,
-            const std::string& name,
+            const std::vector<io::Item>& items,
             bool /*markReloaded*/ = true) override {
-    LOG(info, "Loading model from {}", name);
-    // load items from .npz file
-    auto ioItems = io::loadItems(name);
+    auto ioItems = items;
     // map names and remove a dummy matrix 'decoder_c_tt' from items to avoid creating isolated node
     for(auto it = ioItems.begin(); it != ioItems.end();) {
       // for backwards compatibility, turn one-dimensional vector into two dimensional matrix with first dimension being 1 and second dimension of the original size
@@ -41,7 +39,7 @@ public:
         it->shape.set(0, 1);
         it->shape.set(1, dim);
       }
-      
+
       if(it->name == "decoder_c_tt") {
         it = ioItems.erase(it);
       } else if(it->name == "uidx") {
@@ -59,6 +57,14 @@ public:
     graph->load(ioItems);
   }
 
+  void load(Ptr<ExpressionGraph> graph,
+            const std::string& name,
+            bool /*markReloaded*/ = true) override {
+    LOG(info, "Loading model from {}", name);
+    auto ioItems = io::loadItems(name);
+    load(graph, ioItems);
+  }
+
   void save(Ptr<ExpressionGraph> graph,
             const std::string& name,
             bool saveTranslatorConfig = false) override {

src/translator/scorers.cpp

@@ -5,7 +5,7 @@ namespace marian {
 
 Ptr<Scorer> scorerByType(const std::string& fname,
                          float weight,
-                         const std::string& model,
+                         std::vector<io::Item> items,
                          Ptr<Options> options) {
   options->set("inference", true);
   std::string type = options->get<std::string>("type");
@@ -22,7 +22,7 @@ Ptr<Scorer> scorerByType(const std::string& fname,
 
   LOG(info, "Loading scorer of type {} as feature {}", type, fname);
 
-  return New<ScorerWrapper>(encdec, fname, weight, model);
+  return New<ScorerWrapper>(encdec, fname, weight, items);
 }
 
 Ptr<Scorer> scorerByType(const std::string& fname,
@@ -47,30 +47,30 @@ Ptr<Scorer> scorerByType(const std::string& fname,
   return New<ScorerWrapper>(encdec, fname, weight, ptr);
 }
 
-std::vector<Ptr<Scorer>> createScorers(Ptr<Options> options) {
+std::vector<Ptr<Scorer>> createScorers(Ptr<Options> options, const std::vector<std::vector<io::Item>> models) {
   std::vector<Ptr<Scorer>> scorers;
 
-  auto models = options->get<std::vector<std::string>>("models");
-
   std::vector<float> weights(models.size(), 1.f);
   if(options->hasAndNotEmpty("weights"))
     weights = options->get<std::vector<float>>("weights");
 
   bool isPrevRightLeft = false;  // if the previous model was a right-to-left model
   size_t i = 0;
-  for(auto model : models) {
+  for(auto items : models) {
     std::string fname = "F" + std::to_string(i);
 
     // load options specific for the scorer
     auto modelOptions = New<Options>(options->clone());
-    try {
-      if(!options->get<bool>("ignore-model-config")) {
-        YAML::Node modelYaml;
-        io::getYamlFromModel(modelYaml, "special:model.yml", model);
+    if(!options->get<bool>("ignore-model-config")) {
+      YAML::Node modelYaml;
+      io::getYamlFromModel(modelYaml, "special:model.yml", items);
+      if(!modelYaml.IsNull()) {
+        LOG(info, "Loaded model config");
         modelOptions->merge(modelYaml, true);
       }
-    } catch(std::runtime_error&) {
-      LOG(warn, "No model settings found in model file");
+      else {
+        LOG(warn, "No model settings found in model file");
+      }
     }
 
     // l2r and r2l cannot be used in the same ensemble
@@ -85,13 +85,24 @@ std::vector<Ptr<Scorer>> createScorers(Ptr<Options> options) {
       }
     }
 
-    scorers.push_back(scorerByType(fname, weights[i], model, modelOptions));
+    scorers.push_back(scorerByType(fname, weights[i], items, modelOptions));
     i++;
   }
 
   return scorers;
 }
 
+std::vector<Ptr<Scorer>> createScorers(Ptr<Options> options) {
+  std::vector<std::vector<io::Item>> model_items;
+  auto models = options->get<std::vector<std::string>>("models");
+  for(auto model : models) {
+    auto items = io::loadItems(model);
+    model_items.push_back(std::move(items));
+  }
+
+  return createScorers(options, model_items);
+}
+
 std::vector<Ptr<Scorer>> createScorers(Ptr<Options> options, const std::vector<const void*>& ptrs) {
   std::vector<Ptr<Scorer>> scorers;
 
@@ -105,14 +116,16 @@ std::vector<Ptr<Scorer>> createScorers(Ptr<Options> options, const std::vector<c
 
     // load options specific for the scorer
     auto modelOptions = New<Options>(options->clone());
-    try {
-      if(!options->get<bool>("ignore-model-config")) {
-        YAML::Node modelYaml;
-        io::getYamlFromModel(modelYaml, "special:model.yml", ptr);
+    if(!options->get<bool>("ignore-model-config")) {
+      YAML::Node modelYaml;
+      io::getYamlFromModel(modelYaml, "special:model.yml", ptr);
+      if(!modelYaml.IsNull()) {
+        LOG(info, "Loaded model config");
         modelOptions->merge(modelYaml, true);
       }
-    } catch(std::runtime_error&) {
-      LOG(warn, "No model settings found in model file");
+      else {
+        LOG(warn, "No model settings found in model file");
+      }
     }
 
     scorers.push_back(scorerByType(fname, weights[i], ptr, modelOptions));

src/translator/scorers.h

@@ -73,9 +73,19 @@ class ScorerWrapper : public Scorer {
 private:
   Ptr<IEncoderDecoder> encdec_;
   std::string fname_;
+  std::vector<io::Item> items_;
   const void* ptr_;
 
 public:
+  ScorerWrapper(Ptr<models::IModel> encdec,
+                const std::string& name,
+                float weight,
+                std::vector<io::Item>& items)
+      : Scorer(name, weight),
+        encdec_(std::static_pointer_cast<IEncoderDecoder>(encdec)),
+        items_(items),
+        ptr_{0} {}
+
   ScorerWrapper(Ptr<models::IModel> encdec,
                 const std::string& name,
                 float weight,
@@ -97,7 +107,9 @@ public:
 
   virtual void init(Ptr<ExpressionGraph> graph) override {
     graph->switchParams(getName());
-    if(ptr_)
+    if(!items_.empty())
+      encdec_->load(graph, items_);
+    else if(ptr_)
       encdec_->mmap(graph, ptr_);
     else
       encdec_->load(graph, fname_);
@@ -142,12 +154,19 @@ public:
   }
 };
 
+Ptr<Scorer> scorerByType(const std::string& fname,
+                        float weight,
+                        std::vector<io::Item> items,
+                        Ptr<Options> options);
+
 Ptr<Scorer> scorerByType(const std::string& fname,
                          float weight,
                          const std::string& model,
                          Ptr<Options> config);
 
+
 std::vector<Ptr<Scorer>> createScorers(Ptr<Options> options);
+std::vector<Ptr<Scorer>> createScorers(Ptr<Options> options, const std::vector<std::vector<io::Item>> models);
 
 Ptr<Scorer> scorerByType(const std::string& fname,
                          float weight,

src/translator/translator.h

@@ -20,12 +20,7 @@
 #include "translator/scorers.h"
 
 // currently for diagnostics only, will try to mmap files ending in *.bin suffix when enabled.
-// @TODO: add this as an actual feature.
-#define MMAP 0
-
-#if MMAP
 #include "3rd_party/mio/mio.hpp"
-#endif
 
 namespace marian {
 
@@ -42,9 +37,8 @@ private:
 
   size_t numDevices_;
 
-#if MMAP
-  std::vector<mio::mmap_source> mmaps_;
-#endif
+  std::vector<mio::mmap_source> model_mmaps_; // map
+  std::vector<std::vector<io::Item>> model_items_; // non-mmap
 
 public:
   Translate(Ptr<Options> options)
@@ -76,15 +70,21 @@ public:
     scorers_.resize(numDevices_);
     graphs_.resize(numDevices_);
 
-#if MMAP
     auto models = options->get<std::vector<std::string>>("models");
-    for(auto model : models) {
-      marian::filesystem::Path modelPath(model);
-      ABORT_IF(modelPath.extension() != marian::filesystem::Path(".bin"),
-              "Non-binarized models cannot be mmapped");
-      mmaps_.push_back(std::move(mio::mmap_source(model)));
+    if(options_->get<bool>("model-mmap", false)) {
+      for(auto model : models) {
+        ABORT_IF(!io::isBin(model), "Non-binarized models cannot be mmapped");
+        LOG(info, "Loading model from {}", model);
+        model_mmaps_.push_back(mio::mmap_source(model));
+      }
+    }
+    else {
+      for(auto model : models) {
+        LOG(info, "Loading model from {}", model);
+        auto items = io::loadItems(model);
+        model_items_.push_back(std::move(items));
+      }
     }
-#endif
 
     size_t id = 0;
     for(auto device : devices) {
@@ -101,11 +101,14 @@ public:
         graph->reserveWorkspaceMB(options_->get<size_t>("workspace"));
         graphs_[id] = graph;
 
-#if MMAP
-        auto scorers = createScorers(options_, mmaps_);
-#else
-        auto scorers = createScorers(options_);
-#endif
+        std::vector<Ptr<Scorer>> scorers;
+        if(options_->get<bool>("model-mmap", false)) {
+          scorers = createScorers(options_, model_mmaps_);
+        }
+        else {
+          scorers = createScorers(options_, model_items_);
+        }
+
         for(auto scorer : scorers) {
           scorer->init(graph);
           if(shortlistGenerator_)
@@ -146,11 +149,11 @@ public:
     std::mutex syncCounts;
 
     // timer and counters for total elapsed time and statistics
-    std::unique_ptr<timer::Timer> totTimer(new timer::Timer()); 
+    std::unique_ptr<timer::Timer> totTimer(new timer::Timer());
     size_t totBatches      = 0;
     size_t totLines        = 0;
     size_t totSourceTokens = 0;
-    
+
     // timer and counters for elapsed time and statistics between updates
     std::unique_ptr<timer::Timer> curTimer(new timer::Timer());
     size_t curBatches      = 0;
@@ -176,7 +179,7 @@ public:
     bg.prepare();
     for(auto batch : bg) {
       auto task = [=, &syncCounts,
-                      &totBatches, &totLines, &totSourceTokens, &totTimer, 
+                      &totBatches, &totLines, &totSourceTokens, &totTimer,
                       &curBatches, &curLines, &curSourceTokens, &curTimer](size_t id) {
         thread_local Ptr<ExpressionGraph> graph;
         thread_local std::vector<Ptr<Scorer>> scorers;
@@ -200,12 +203,12 @@ public:
         }
 
         // if we asked for speed information display this
-        if(statFreq.n > 0) { 
+        if(statFreq.n > 0) {
           std::lock_guard<std::mutex> lock(syncCounts);
-          totBatches++; 
+          totBatches++;
           totLines        += batch->size();
           totSourceTokens += batch->front()->batchWords();
-        
+
           curBatches++;
           curLines        += batch->size();
           curSourceTokens += batch->front()->batchWords();
@@ -214,10 +217,10 @@ public:
             double totTime = totTimer->elapsed();
             double curTime = curTimer->elapsed();
 
-            LOG(info, 
-                "Processed {} batches, {} lines, {} source tokens in {:.2f}s - Speed (since last): {:.2f} batches/s - {:.2f} lines/s - {:.2f} tokens/s", 
+            LOG(info,
+                "Processed {} batches, {} lines, {} source tokens in {:.2f}s - Speed (since last): {:.2f} batches/s - {:.2f} lines/s - {:.2f} tokens/s",
                 totBatches, totLines, totSourceTokens, totTime, curBatches / curTime, curLines / curTime, curSourceTokens / curTime);
-            
+
             // reset stats between updates
             curBatches = curLines = curSourceTokens = 0;
             curTimer.reset(new timer::Timer());
@@ -230,12 +233,12 @@ public:
 
     // make sure threads are joined before other local variables get de-allocated
     threadPool.join_all();
-    
+
     // display final speed numbers over total translation if intermediate displays were requested
     if(statFreq.n > 0) {
       double totTime = totTimer->elapsed();
-      LOG(info, 
-          "Processed {} batches, {} lines, {} source tokens in {:.2f}s - Speed (total): {:.2f} batches/s - {:.2f} lines/s - {:.2f} tokens/s", 
+      LOG(info,
+          "Processed {} batches, {} lines, {} source tokens in {:.2f}s - Speed (total): {:.2f} batches/s - {:.2f} lines/s - {:.2f} tokens/s",
           totBatches, totLines, totSourceTokens, totTime, totBatches / totTime, totLines / totTime, totSourceTokens / totTime);
     }
   }
@@ -288,6 +291,14 @@ public:
     auto devices = Config::getDevices(options_);
     numDevices_ = devices.size();
 
+    // preload models
+    std::vector<std::vector<io::Item>> model_items_;
+    auto models = options->get<std::vector<std::string>>("models");
+    for(auto model : models) {
+      auto items = io::loadItems(model);
+      model_items_.push_back(std::move(items));
+    }
+
     // initialize scorers
     for(auto device : devices) {
       auto graph = New<ExpressionGraph>(true);
@@ -303,7 +314,7 @@ public:
       graph->reserveWorkspaceMB(options_->get<size_t>("workspace"));
       graphs_.push_back(graph);
 
-      auto scorers = createScorers(options_);
+      auto scorers = createScorers(options_, model_items_);
       for(auto scorer : scorers) {
         scorer->init(graph);
         if(shortlistGenerator_)