wassname/DeepRL
1
0
Fork 0
mirror of https://github.com/wassname/DeepRL.git synced 2026-09-09 11:13:47 +08:00
Code Issues Packages Projects Releases Wiki Activity

Update ACVP

Browse Source
This commit is contained in:
Shangtong Zhang
2018-04-10 11:32:59 -06:00
parent af1cc11669
commit 669c366bc0
3 changed files with 6 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
main.py
+3 -3
View File
@@ -302,8 +302,8 @@ def action_conditional_video_prediction():
# a2c_pixel_atari(game) # a2c_pixel_atari(game)
# Generate a dataset with the trained model # Generate a dataset with the trained model
a2c_model_file = './data/A2CAgent-vanilla-model-%s.bin' % (game) # a2c_model_file = './data/A2CAgent-vanilla-model-%s.bin' % (game)
generate_dataset(game, a2c_model_file, prefix) # generate_dataset(game, a2c_model_file, prefix)
# Train the action conditional video prediction model # Train the action conditional video prediction model
acvp_train(game, prefix) acvp_train(game, prefix)
@@ -334,7 +334,7 @@ if __name__ == '__main__':
# ddpg_continuous() # ddpg_continuous()
# ppo_continuous() # ppo_continuous()
action_conditional_video_prediction() # action_conditional_video_prediction()
# plot() # plot()
model/action_conditional_video_prediction.py
+2
View File
@@ -4,6 +4,8 @@
# declaration at the top # # declaration at the top #
####################################################################### #######################################################################
__all__ = ['acvp_train']
import torch import torch
from torch.autograd import Variable from torch.autograd import Variable
import torch.nn as nn import torch.nn as nn
model/dataset.py
+1
View File
@@ -50,6 +50,7 @@ def generate_dataset(game, a2c_model, prefix):
config.gradient_clip = 0.5 config.gradient_clip = 0.5
config.logger = Logger('./log', logger, skip=True) config.logger = Logger('./log', logger, skip=True)
agent = A2CAgent(config) agent = A2CAgent(config)
agent.close()
agent.load(a2c_model) agent.load(a2c_model)
task = PixelAtari(game, frame_skip=4, history_length=4, log_dir=None, dataset=True) task = PixelAtari(game, frame_skip=4, history_length=4, log_dir=None, dataset=True)