From f308cac6df712da3619fb05b14f3345f0ec07b9a Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Adri=C3=A0=20Arrufat?= <adria.arrufat@outlook.com>
Date: Sun, 26 Sep 2021 20:47:00 +0900
Subject: [PATCH] try to fix it reusing the code... not sure though

---
 dlib/dnn/trainer.h | 46 +++++++---------------------------------------
 1 file changed, 7 insertions(+), 39 deletions(-)

diff --git a/dlib/dnn/trainer.h b/dlib/dnn/trainer.h
index bf67ce33d..da714cb20 100644
--- a/dlib/dnn/trainer.h
+++ b/dlib/dnn/trainer.h
@@ -1194,6 +1194,9 @@ namespace dlib
             job.have_data.resize(devs);
             job.test_only = test_only;
 
+            // check if the iterator points to anything
+            const bool have_labels = sizeof(*lbegin) != 1;
+
             // chop the data into devs blocks, each of about block_size elements.
             const double block_size = num / static_cast<double>(devs);
 
@@ -1211,7 +1214,8 @@ namespace dlib
                 if (start < stop)
                 {
                     devices[i]->net.to_tensor(dbegin+start, dbegin+stop, job.t[i]);
-                    job.labels[i].assign(lbegin+start, lbegin+stop);
+                    if (have_labels)
+                        job.labels[i].assign(lbegin+start, lbegin+stop);
                     job.have_data[i] = true;
                 }
                 else
@@ -1237,44 +1241,8 @@ namespace dlib
             data_iterator dend
         )
         {
-            propagate_exception();
-            size_t num = std::distance(dbegin, dend);
-            size_t devs = devices.size();
-            job.t.resize(devs);
-            job.have_data.resize(devs);
-            job.test_only = test_only;
-
-            // chop the data into devs blocks, each of about block_size elements.
-            const double block_size = num / static_cast<double>(devs);
-
-            const auto prev_dev = dlib::cuda::get_device();
-
-            double j = 0;
-
-            for (size_t i = 0; i < devs; ++i)
-            {
-                dlib::cuda::set_device(devices[i]->device_id);
-
-                const size_t start = static_cast<size_t>(std::round(j));
-                const size_t stop  = static_cast<size_t>(std::round(j + block_size));
-
-                if (start < stop)
-                {
-                    devices[i]->net.to_tensor(dbegin+start, dbegin+stop, job.t[i]);
-                    job.have_data[i] = true;
-                }
-                else
-                {
-                    job.have_data[i] = false;
-                }
-
-                j += block_size;
-            }
-
-            DLIB_ASSERT(std::fabs(j - num) < 1e-10);
-
-            dlib::cuda::set_device(prev_dev);
-            job_pipe.enqueue(job);
+            typename std::vector<training_label_type>::iterator nothing;
+            send_job(test_only, dbegin, dend, nothing);
         }
 
         void print_progress()