Rename folder with training script so they are not used for enhancing by users accidentally.

2016-11-11 10:00:23 +01:00
parent fabe04f1d0
commit a22d7ced3c
2 changed files with 0 additions and 0 deletions
@@ -0,0 +1,18 @@
+#!/bin/sh
+
+python3.4 enhance.py \
+    --train "$OPEN_IMAGES_PATH/*/*.jpg" --type photo --model repair \
+    --epochs=50 --batch-shape=256 --device=gpu1 \
+    --generator-downscale=2 --generator-upscale=2 \
+    --generator-blocks=8 --generator-filters=128 --generator-residual=0 \
+    --perceptual-layer=conv2_2 --smoothness-weight=1e7 --adversary-weight=0.0 \
+    --train-noise=2.0 --train-jpeg=30
+
+python3.4 enhance.py \
+    --train "$OPEN_IMAGES_PATH/*/*.jpg" --type photo --model repair \
+    --epochs=500 --batch-shape=240 --device=gpu1 \
+    --generator-downscale=2 --generator-upscale=2 \
+    --perceptual-layer=conv5_2 --smoothness-weight=5e3 --adversary-weight=5e1 \
+    --generator-start=10 --discriminator-start=0 --adversarial-start=10 \
+    --discriminator-size=48 \
+    --train-noise=2.0 --train-jpeg=30
@@ -0,0 +1,18 @@
+#!/bin/sh
+
+python3.4 enhance.py \
+    --train "$OPEN_IMAGES_PATH/*/*.jpg" --type photo --model default \
+    --epochs=50 --batch-shape=256 --device=gpu0 \
+    --generator-downscale=0 --generator-upscale=1 \
+    --generator-blocks=8 --generator-filters=128 --generator-residual=0 \
+    --perceptual-layer=conv2_2 --smoothness-weight=1e7 --adversary-weight=0.0 \
+    --train-noise=1.0
+
+python3.4 enhance.py \
+    --train "$OPEN_IMAGES_PATH/*/*.jpg" --type photo --model default \
+    --epochs=500 --batch-shape=240 --device=gpu0 \
+    --generator-downscale=0 --generator-upscale=1 \
+    --perceptual-layer=conv5_2 --smoothness-weight=5e3 --adversary-weight=5e1 \
+    --generator-start=10 --discriminator-start=0 --adversarial-start=10 \
+    --discriminator-size=64 \
+    --train-noise=1.0