From e49b74d614a795061892653397e127844334da87 Mon Sep 17 00:00:00 2001
From: Alex Lee <alexleegk@gmail.com>
Date: Wed, 11 Apr 2018 19:38:57 -0700
Subject: [PATCH] Change default context_frames and sequence_lengths defaults
to -1. Set default batch_size=32 for mocogan.
---
video_prediction/models/base_model.py | 12 ++++++------
video_prediction/models/mocogan_model.py | 1 +
2 files changed, 7 insertions(+), 6 deletions(-)
diff --git a/video_prediction/models/base_model.py b/video_prediction/models/base_model.py
index c5da777f..3facdbcb 100644
--- a/video_prediction/models/base_model.py
+++ b/video_prediction/models/base_model.py
@@ -47,10 +47,10 @@ class BaseVideoPredictionModel(object):
self.eval_num_samples = eval_num_samples
self.eval_parallel_iterations = eval_parallel_iterations
self.hparams = self.parse_hparams(hparams_dict, hparams)
- if not self.hparams.context_frames:
+ if self.hparams.context_frames == -1:
raise ValueError('Invalid context_frames %r. It might have to be '
'specified.' % self.hparams.context_frames)
- if not self.hparams.sequence_length:
+ if self.hparams.sequence_length == -1:
raise ValueError('Invalid sequence_length %r. It might have to be '
'specified.' % self.hparams.sequence_length)
@@ -84,8 +84,8 @@ class BaseVideoPredictionModel(object):
repeat: the number of repeat actions (if applicable).
"""
hparams = dict(
- context_frames=0,
- sequence_length=0,
+ context_frames=-1,
+ sequence_length=-1,
repeat=1,
)
return hparams
@@ -348,8 +348,8 @@ class VideoPredictionModel(BaseVideoPredictionModel):
max_steps=300000,
beta1=0.9,
beta2=0.999,
- context_frames=0,
- sequence_length=0,
+ context_frames=-1,
+ sequence_length=-1,
clip_length=10,
l1_weight=0.0,
l2_weight=1.0,
diff --git a/video_prediction/models/mocogan_model.py b/video_prediction/models/mocogan_model.py
index ac9c470a..22490c4d 100644
--- a/video_prediction/models/mocogan_model.py
+++ b/video_prediction/models/mocogan_model.py
@@ -291,6 +291,7 @@ class MoCoGANVideoPredictionModel(VideoPredictionModel):
def get_default_hparams_dict(self):
default_hparams = super(MoCoGANVideoPredictionModel, self).get_default_hparams_dict()
hparams = dict(
+ batch_size=32,
l1_weight=10.0,
l2_weight=0.0,
image_gan_weight=1.0,
--
GitLab