Fix cantains contains (PaddlePaddle#61066)

paddle/cinn/backends/codegen_cuda_util.h

@@ -34,7 +34,7 @@ namespace backends {
 #define TENSOR_SHAPE_ARGS "tensor_shape_args"
 
 /**
- * Split a CINN Module into two separate modules, one cantains the host
+ * Split a CINN Module into two separate modules, one contains the host
  * functions, the other contains the device kernels.
  *
  * This contains some process:

paddle/cinn/hlir/dialect/operator/transforms/group_merge/op_with_group_merge_util.h

@@ -196,7 +196,7 @@ inline bool is_horizontal_relation(::pir::Operation* producer,
         if (!consumer->ops_set.count(tmp_op)) {
           continue;
         }
-        // recored visited op.
+        // recorded visited op.
         if (!visited_set.count(tmp_op)) {
           visited_set.insert(tmp_op);
           candidates.push(tmp_op);

paddle/cinn/hlir/pass/op_fusion_pass_util.h

@@ -134,15 +134,15 @@ CONDITION_FUNC(is_horizontal_relation) {
       candidates.pop();
       // visit all producer node
       for (auto tmp_node : helper->GetProducerNode(candidate)) {
-        // check depency.
+        // check dependency.
         if (producer == tmp_node) {
           return true;
         }
         // check node is in region.
         if (!consumer->nodes_set.count(tmp_node)) {
           continue;
         }
-        // recored visited node.
+        // recorded visited node.
         if (!visited_set.count(tmp_node)) {
           visited_set.insert(tmp_node);
           candidates.push(tmp_node);

paddle/cinn/ir/group_schedule/tactic/optimize_reduction_tactic.cc

@@ -119,7 +119,7 @@ void OptimizeReductionTactic::Apply(ir::IRSchedule* sch,
   std::vector<ir::Expr> loops = sch->GetLoops(block_id);
   int first_reduce_loop_idx = context_->iter_space_info.sp_space.size();
   CHECK_LT(first_reduce_loop_idx, loops.size())
-      << "first_reduce_loop_idx shoud be less than number of loop.";
+      << "first_reduce_loop_idx should be less than number of loop.";
   ir::Expr block = sch->GetBlock(block_id);
   ir::Tensor reduce_tensor = analyzer::GetStoreTensorOfSBlock(block);
   int non_reduce_memory_space_rank =

paddle/cinn/lang/README.md

@@ -12,7 +12,7 @@ One can pass a `Shared` object by passing a pointer and the consumer object shou
 
 ## Tensor
 
-The input or the temporary ouptut node.
+The input or the temporary output node.
 
 Every `Compute` will output a Tensor, the tensor can be sliced.
 

paddle/fluid/distributed/ps/table/depends/dense.h

@@ -28,7 +28,7 @@
 namespace paddle {
 namespace distributed {
 
-// dense optimzier
+// dense optimizer
 // TODO(tangwei12) integrate with sparse optimzer later.
 class DenseOptimizer {
  public:

paddle/fluid/framework/fleet/box_wrapper.h

@@ -107,7 +107,7 @@ class BasicAucCalculator {
   double size() const { return _size; }
   double rmse() const { return _rmse; }
   std::vector<double>& get_negative() { return _table[0]; }
-  std::vector<double>& get_postive() { return _table[1]; }
+  std::vector<double>& get_positive() { return _table[1]; }
   double& local_abserr() { return _local_abserr; }
   double& local_sqrerr() { return _local_sqrerr; }
   double& local_pred() { return _local_pred; }
@@ -453,7 +453,7 @@ class BoxWrapper {
       p_agent_ = boxps::PSAgentBase::GetIns(feedpass_thread_num_);
       p_agent_->Init();
       for (const auto& slot_name : slot_omit_in_feedpass) {
-        slot_name_omited_in_feedpass_.insert(slot_name);
+        slot_name_omitted_in_feedpass_.insert(slot_name);
       }
       slot_vector_ = slot_vector;
       keys_tensor.resize(platform::GetGPUDeviceCount());
@@ -548,8 +548,8 @@ class BoxWrapper {
 
   bool UseAfsApi() const { return use_afs_api_; }
 
-  const std::unordered_set<std::string>& GetOmitedSlot() const {
-    return slot_name_omited_in_feedpass_;
+  const std::unordered_set<std::string>& GetOmittedSlot() const {
+    return slot_name_omitted_in_feedpass_;
   }
 
   class MetricMsg {
@@ -914,7 +914,7 @@ class BoxWrapper {
   // TODO(hutuxian): magic number, will add a config to specify
   const int feedpass_thread_num_ = 30;  // magic number
   static std::shared_ptr<BoxWrapper> s_instance_;
-  std::unordered_set<std::string> slot_name_omited_in_feedpass_;
+  std::unordered_set<std::string> slot_name_omitted_in_feedpass_;
   // EMBEDX_DIM and EXPAND_EMBED_DIM
   static int embedx_dim_;
   static int expand_embed_dim_;
@@ -1098,19 +1098,19 @@ class BoxHelper {
     const std::deque<Record>& pass_data = input_channel_->GetData();
 
     // get feasigns that FeedPass doesn't need
-    const std::unordered_set<std::string>& slot_name_omited_in_feedpass_ =
-        box_ptr->GetOmitedSlot();
-    std::unordered_set<int> slot_id_omited_in_feedpass_;
+    const std::unordered_set<std::string>& slot_name_omitted_in_feedpass_ =
+        box_ptr->GetOmittedSlot();
+    std::unordered_set<int> slot_id_omitted_in_feedpass_;
     const auto& all_readers = dataset_->GetReaders();
     PADDLE_ENFORCE_GT(all_readers.size(),
                       0,
                       platform::errors::PreconditionNotMet(
                           "Readers number must be greater than 0."));
     const auto& all_slots_name = all_readers[0]->GetAllSlotAlias();
     for (size_t i = 0; i < all_slots_name.size(); ++i) {
-      if (slot_name_omited_in_feedpass_.find(all_slots_name[i]) !=
-          slot_name_omited_in_feedpass_.end()) {
-        slot_id_omited_in_feedpass_.insert(i);
+      if (slot_name_omitted_in_feedpass_.find(all_slots_name[i]) !=
+          slot_name_omitted_in_feedpass_.end()) {
+        slot_id_omitted_in_feedpass_.insert(i);
       }
     }
     const size_t tnum = box_ptr->GetFeedpassThreadNum();
@@ -1130,7 +1130,7 @@ class BoxHelper {
                       begin,
                       begin + len_per_thread + (i < remain ? 1 : 0),
                       p_agent,
-                      std::ref(slot_id_omited_in_feedpass_),
+                      std::ref(slot_id_omitted_in_feedpass_),
                       i));
       begin += len_per_thread + (i < remain ? 1 : 0);
     }

paddle/fluid/framework/fleet/ps_gpu_wrapper.h

@@ -1015,7 +1015,7 @@ class PSGPUWrapper {
 #ifdef PADDLE_WITH_CUDA
   std::vector<MemoryPool*> mem_pools_;
   std::vector<HBMMemoryPoolFix*> hbm_pools_;  // in multi mfdim, one table need
-                                              // hbm pools of totol dims number
+                                              // hbm pools of total dims number
 #endif
 
   std::shared_ptr<

paddle/fluid/framework/new_executor/executor_statistics.cc

@@ -454,9 +454,9 @@ int StatisticsEngine::Stat(const platform::NodeTrees& trees) {
     }
   }
   auto& python_end = statistics_[name2idx_["PythonEnd"]];
-  const auto& totol = statistics_[name2idx_["Total"]];
+  const auto& total = statistics_[name2idx_["Total"]];
   const auto& cplusplus_end = statistics_[name2idx_["CplusplusEnd"]];
-  python_end.total_time = totol.total_time - cplusplus_end.total_time;
+  python_end.total_time = total.total_time - cplusplus_end.total_time;
   python_end.count = cplusplus_end.count + 1;
 
   auto& launch_kernel = statistics_[name2idx_["LaunchKernel"]];

paddle/fluid/framework/ps_gpu_worker.cc

@@ -432,7 +432,7 @@ void PSGPUWorker::TrainFiles() {
         {
           std::lock_guard<std::mutex> lock(mutex);
           VLOG(0) << "worker " << thread_id_ << ": " << var_name
-                  << " cantains inf or nan";
+                  << " contains inf or nan";
           auto all_vars = thread_scope->LocalVarNames();
           std::stringstream ss;
           ss << "====== worker " << thread_id_ << "======\n";
@@ -556,7 +556,7 @@ void PSGPUWorker::TrainFilesWithProfiler() {
   for (size_t i = 0; i < op_name.size(); ++i) {
     VLOG(0) << "card:" << thread_id_ << ", op: " << op_name[i]
             << ", mean time: " << op_total_time[i] / total_ins_num
-            << "s, totol time:" << op_total_time[i] << "sec";
+            << "s, total time:" << op_total_time[i] << "sec";
   }
   VLOG(0) << "card: " << thread_id_ << " read time: " << read_time
           << ", percent: " << read_time / total_time * 100;