From 070fd9fa2b7dd83485b9bafe5dbde6ad8c56ac3b Mon Sep 17 00:00:00 2001
From: tomrink <rink@ssec.wisc.edu>
Date: Thu, 12 Oct 2023 10:31:24 -0500
Subject: [PATCH] snapshot...
---
modules/GSOC/E2_ESRGAN/lib/train.py | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/modules/GSOC/E2_ESRGAN/lib/train.py b/modules/GSOC/E2_ESRGAN/lib/train.py
index 7d234fcd..f8ff4150 100644
--- a/modules/GSOC/E2_ESRGAN/lib/train.py
+++ b/modules/GSOC/E2_ESRGAN/lib/train.py
@@ -92,14 +92,15 @@ class Trainer(object):
logging.debug("Starting Distributed Step")
with tf.GradientTape() as tape:
fake = generator.unsigned_call(image_lr)
- loss = utils.pixel_loss(image_hr, fake) * (1.0 / self.batch_size)
- # loss = utils.pixel_loss_mse(image_hr, fake) * (1.0 / self.batch_size)
+ loss_mae = utils.pixel_loss(image_hr, fake) * (1.0 / self.batch_size)
+ # loss_mse = utils.pixel_loss_mse(image_hr, fake) * (1.0 / self.batch_size)
+ loss = loss_mae
+ mean_loss = metric(loss_mae)
psnr_metric(tf.reduce_mean(tf.image.psnr(fake, image_hr, max_val=PSNR_MAX)))
# gen_vars = list(set(generator.trainable_variables))
gen_vars = generator.trainable_variables
gradient = tape.gradient(loss, gen_vars)
G_optimizer.apply_gradients(zip(gradient, gen_vars))
- mean_loss = metric(loss)
logging.debug("Ending Distributed Step")
return tf.cast(G_optimizer.iterations, tf.float32)
--
GitLab