From 229acda8565b3ef0a1e980389626998cfd822513 Mon Sep 17 00:00:00 2001
From: littletomatodonkey <dazhiningsibuqu@163.com>
Date: Mon, 7 Feb 2022 12:19:25 +0000
Subject: [PATCH] fix ips info and reduce interval of metric calc

---
 ppocr/losses/rec_ctc_loss.py |  3 ++-
 tools/program.py             | 21 +++++++++++----------
 2 files changed, 13 insertions(+), 11 deletions(-)

diff --git a/ppocr/losses/rec_ctc_loss.py b/ppocr/losses/rec_ctc_loss.py
index 063d68e3..502fc8c5 100755
--- a/ppocr/losses/rec_ctc_loss.py
+++ b/ppocr/losses/rec_ctc_loss.py
@@ -31,7 +31,8 @@ class CTCLoss(nn.Layer):
             predicts = predicts[-1]
         predicts = predicts.transpose((1, 0, 2))
         N, B, _ = predicts.shape
-        preds_lengths = paddle.to_tensor([N] * B, dtype='int64')
+        preds_lengths = paddle.to_tensor(
+            [N] * B, dtype='int64', place=paddle.CPUPlace())
         labels = batch[1].astype("int32")
         label_lengths = batch[2].astype('int64')
         loss = self.loss_func(predicts, labels, preds_lengths, label_lengths)
diff --git a/tools/program.py b/tools/program.py
index 5ffb93d1..a0336916 100755
--- a/tools/program.py
+++ b/tools/program.py
@@ -146,6 +146,7 @@ def train(config,
           scaler=None):
     cal_metric_during_train = config['Global'].get('cal_metric_during_train',
                                                    False)
+    calc_epoch_interval = config['Global'].get('calc_epoch_interval', 1)
     log_smooth_window = config['Global']['log_smooth_window']
     epoch_num = config['Global']['epoch_num']
     print_batch_step = config['Global']['print_batch_step']
@@ -244,6 +245,16 @@ def train(config,
                 optimizer.step()
             optimizer.clear_grad()
 
+            if cal_metric_during_train and epoch % calc_epoch_interval == 0:  # only rec and cls need
+                batch = [item.numpy() for item in batch]
+                if model_type in ['table', 'kie']:
+                    eval_class(preds, batch)
+                else:
+                    post_result = post_process_class(preds, batch[1])
+                    eval_class(post_result, batch)
+                metric = eval_class.get_metric()
+                train_stats.update(metric)
+
             train_batch_time = time.time() - reader_start
             train_batch_cost += train_batch_time
             eta_meter.update(train_batch_time)
@@ -258,16 +269,6 @@ def train(config,
             stats['lr'] = lr
             train_stats.update(stats)
 
-            if cal_metric_during_train:  # only rec and cls need
-                batch = [item.numpy() for item in batch]
-                if model_type in ['table', 'kie']:
-                    eval_class(preds, batch)
-                else:
-                    post_result = post_process_class(preds, batch[1])
-                    eval_class(post_result, batch)
-                metric = eval_class.get_metric()
-                train_stats.update(metric)
-
             if vdl_writer is not None and dist.get_rank() == 0:
                 for k, v in train_stats.get().items():
                     vdl_writer.add_scalar('TRAIN/{}'.format(k), v, global_step)
-- 
GitLab