From 3245acdf564e247133fb90cd5e19efc32f5f1352 Mon Sep 17 00:00:00 2001
From: tomrink <rink@ssec.wisc.edu>
Date: Wed, 27 Oct 2021 10:44:25 -0500
Subject: [PATCH] new methods for running predictions
---
modules/deeplearning/icing_cnn.py | 43 +++++++++++++++++++++++++++----
1 file changed, 38 insertions(+), 5 deletions(-)
diff --git a/modules/deeplearning/icing_cnn.py b/modules/deeplearning/icing_cnn.py
index d8cc5523..afb0cf37 100644
--- a/modules/deeplearning/icing_cnn.py
+++ b/modules/deeplearning/icing_cnn.py
@@ -1,6 +1,6 @@
import tensorflow as tf
import tensorflow_addons as tfa
-from util.setup import logdir, modeldir, cachepath, now, ewa_varsdir
+from util.setup import logdir, modeldir, cachepath, now
from util.util import homedir, EarlyStop, normalize, make_for_full_domain_predict
from util.geos_nav import get_navigation
@@ -939,12 +939,14 @@ class IcingIntensityNN:
self.test_preds = preds
- def do_evaluate(self, ckpt_dir, prob_thresh=0.5):
+ def do_evaluate(self, ckpt_dir=None, prob_thresh=0.5):
- ckpt = tf.train.Checkpoint(step=tf.Variable(1), model=self.model)
- ckpt_manager = tf.train.CheckpointManager(ckpt, ckpt_dir, max_to_keep=3)
+ if ckpt_dir is not None: # if is None, this has been done already
+ ckpt = tf.train.Checkpoint(step=tf.Variable(1), model=self.model)
+ ckpt_manager = tf.train.CheckpointManager(ckpt, ckpt_dir, max_to_keep=3)
+ ckpt.restore(ckpt_manager.latest_checkpoint)
- ckpt.restore(ckpt_manager.latest_checkpoint)
+ self.reset_test_metrics()
pred_s = []
for data in self.eval_dataset:
@@ -1079,6 +1081,37 @@ def run_evaluate_static(h5f, ckpt_dir_s_path, flight_level=4, prob_thresh=0.5, s
return ice_lons, ice_lats, preds_2d, lons_2d, lats_2d, x_rad, y_rad
+def run_evaluate_static_new(data_dct, num_lines, num_elems, ckpt_dir_s_path, flight_level=4, prob_thresh=0.5):
+
+ ckpt_dir_s = os.listdir(ckpt_dir_s_path)
+ ckpt_dir = ckpt_dir_s[0]
+
+ nn = IcingIntensityNN()
+ nn.flight_level = flight_level
+ nn.setup_eval_pipeline(data_dct, num_lines * num_elems)
+ nn.build_model()
+ nn.build_training()
+ nn.build_evaluation()
+
+ ckpt = tf.train.Checkpoint(step=tf.Variable(1), model=nn.model)
+ ckpt_manager = tf.train.CheckpointManager(ckpt, ckpt_dir, max_to_keep=3)
+ ckpt.restore(ckpt_manager.latest_checkpoint)
+
+ for flvl in [0, 1, 2, 3, 4]:
+ nn.flight_level = flvl
+ nn.do_evaluate()
+ probs = nn.test_probs
+
+ if NumClasses == 2:
+ preds = np.where(probs > prob_thresh, 1, 0)
+ else:
+ preds = np.argmax(probs, axis=1)
+ preds_2d = preds.reshape((num_lines, num_elems))
+ probs_2d = probs.reshape((num_lines, num_elems))
+
+ return preds_2d, probs_2d
+
+
if __name__ == "__main__":
nn = IcingIntensityNN()
nn.run('matchup_filename')
--
GitLab