Initializing LGAN_A2CAgent with parameters:Initializing LGAN_A2CAgent with parameters:  {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}{'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}

Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
{'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: Initializing LGAN_A2CAgent with parameters: Initializing LGAN_A2CAgent with parameters:{'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}Initializing LGAN_A2CAgent with parameters:{'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192} 

 {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
{'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
Initializing LGAN_A2CAgent with parameters: {'game': 'breakthrough(rows=8,columns=8)', 'agent': 'LGAN_A2C', 'train_group_size': 5000, 'gamma': 0.99, 'epsilon_start': 1, 'epsilon_decay_duration': 1000000, 'epsilon_end': 0.1, 'batch_size': 128, 'd_model': 512, 'num_heads': 4, 'average_func': 'mean', 'buffer_size': 100000, 'lr': 0.001, 'k_embedding_dim': 16, 'num_layers': 1, 'train_interval': 10, 'target_net_update_interval': 1000, 'num_actions': 768, 'state_dim': 192}
