def get_generator(checkpoint, evalArgs):
args = AttrDict(checkpoint['args'])
generator = TrajectoryGenerator(
obs_len=args.obs_len,
pred_len=args.pred_len,
embedding_dim=args.embedding_dim,
encoder_h_dim=args.encoder_h_dim_g,
decoder_h_dim=args.decoder_h_dim_g,
mlp_dim=args.mlp_dim,
num_layers=args.num_layers,
noise_dim=args.noise_dim,
noise_type=args.noise_type,
noise_mix_type=args.noise_mix_type,
pooling_type=args.pooling_type,
pool_every_timestep=args.pool_every_timestep,
dropout=args.dropout,
bottleneck_dim=args.bottleneck_dim,
neighborhood_size=args.neighborhood_size,
grid_size=args.grid_size,
batch_norm=args.batch_norm,
use_gpu=evalArgs.use_gpu)
generator.load_state_dict(checkpoint['g_state'])
if evalArgs.use_gpu:
generator.cuda()
else:
generator.cpu()
generator.train()
return generator
def get_generator(self, checkpoint):
self.import_modules()
from sgan.models import TrajectoryGenerator
args = attrdict.AttrDict(checkpoint['args'])
generator = TrajectoryGenerator(
obs_len=args.obs_len,
pred_len=args.pred_len,
embedding_dim=args.embedding_dim,
encoder_h_dim=args.encoder_h_dim_g,
decoder_h_dim=args.decoder_h_dim_g,
mlp_dim=args.mlp_dim,
num_layers=args.num_layers,
noise_dim=args.noise_dim,
noise_type=args.noise_type,
noise_mix_type=args.noise_mix_type,
pooling_type=args.pooling_type,
pool_every_timestep=args.pool_every_timestep,
dropout=args.dropout,
bottleneck_dim=args.bottleneck_dim,
neighborhood_size=args.neighborhood_size,
grid_size=args.grid_size,
batch_norm=args.batch_norm)
generator.load_state_dict(checkpoint['g_state'])
generator.cpu()
generator.train()
return generator