agent.py

import json
import os
import sys
import numpy as np
import random
import math
import time

import torch
import torch.nn as nn
from torch.autograd import Variable
from torch import optim
import torch.nn.functional as F

from env import R2RBatch
from utils import padding_idx, add_idx, Tokenizer
import utils
import model
import param
from param import args
from collections import defaultdict

class BaseAgent(object):
    ''' Base class for an R2R agent to generate and save trajectories. '''

    def __init__(self, env, results_path):
        self.env = env
        self.results_path = results_path
        random.seed(1)
        self.results = {}
        self.losses = [] # For learning agents
    
    def write_results(self):
        output = [{'instr_id':k, 'trajectory': v} for k,v in self.results.items()]
        with open(self.results_path, 'w') as f:
            json.dump(output, f)

    def get_results(self):
        output = [{'instr_id': k, 'trajectory': v} for k, v in self.results.items()]
        return output

    def rollout(self, **args):
        ''' Return a list of dicts containing instr_id:'xx', path:[(viewpointId, heading_rad, elevation_rad)]  '''
        raise NotImplementedError

    @staticmethod
    def get_agent(name):
        return globals()[name+"Agent"]

    def test(self, iters=None, **kwargs):
        self.env.reset_epoch(shuffle=(iters is not None))   # If iters is not none, shuffle the env batch
        self.losses = []
        self.results = {}
        # We rely on env showing the entire batch before repeating anything
        looped = False
        self.loss = 0
        if iters is not None:
            # For each time, it will run the first 'iters' iterations. (It was shuffled before)
            for i in range(iters):
                for traj in self.rollout(**kwargs):
                    self.loss = 0
                    self.results[traj['instr_id']] = traj['path']
        else:   # Do a full round
            while True:
                for traj in self.rollout(**kwargs):
                    if traj['instr_id'] in self.results:
                        looped = True
                    else:
                        self.loss = 0
                        self.results[traj['instr_id']] = traj['path']
                if looped:
                    break

class Seq2SeqAgent(BaseAgent):
    ''' An agent based on an LSTM seq2seq model with attention. '''

    # For now, the agent can't pick which forward move to make - just the one in the middle
    env_actions = {
      'left': (0,-1, 0), # left
      'right': (0, 1, 0), # right
      'up': (0, 0, 1), # up
      'down': (0, 0,-1), # down
      'forward': (1, 0, 0), # forward
      '<end>': (0, 0, 0), # <end>
      '<start>': (0, 0, 0), # <start>
      '<ignore>': (0, 0, 0)  # <ignore>
    }

    def __init__(self, env, results_path, tok, episode_len=20):
        super(Seq2SeqAgent, self).__init__(env, results_path)
        self.tok = tok
        self.episode_len = episode_len
        self.feature_size = self.env.feature_size

        # glove encoder
        with open('img_features/objects/object_vocab.txt', 'r') as f_ov:
            obj_vocab = [k.strip() for k in f_ov.readlines()]
        obj_glove_matrix = utils.get_glove_matrix(obj_vocab, args.glove_dim)
        self.objencoder = model.ObjEncoder(obj_glove_matrix.shape[0], obj_glove_matrix.shape[1], obj_glove_matrix).cuda()
        self.CLIP_language = model.CLIP_language()
        # Models
        enc_hidden_size = args.rnn_dim//2 if args.bidir else args.rnn_dim
        self.encoder = model.EncoderLSTM(tok.vocab_size(), args.wemb, enc_hidden_size, padding_idx,
                                         args.dropout, bidirectional=args.bidir).cuda()
        self.decoder = model.ASODecoderLSTM(args.aemb, args.rnn_dim, args.dropout).cuda()
        self.critic = model.Critic().cuda()
        self.models = (self.encoder, self.decoder, self.critic)
        self.critic_object = model.Critic_object().cuda()
        # Optimizers
        self.encoder_optimizer = args.optimizer(self.encoder.parameters(), lr=args.lr)
        self.decoder_optimizer = args.optimizer(self.decoder.parameters(), lr=args.lr)
        self.critic_optimizer = args.optimizer(self.critic.parameters(), lr=args.lr)
        self.critic_object_optimizer = args.optimizer(self.critic_object.parameters(), lr=args.lr)
        self.optimizers = (self.encoder_optimizer, self.decoder_optimizer, self.critic_optimizer)

        # Evaluations
        self.losses = []
        self.criterion = nn.CrossEntropyLoss(ignore_index=args.ignoreid, size_average=False)
        self.ndtw_criterion = utils.ndtw_initialize()

        # Logs
        sys.stdout.flush()
        self.logs = defaultdict(list)


    def _sort_batch(self, obs):
        ''' Extract instructions from a list of observations and sort by descending
            sequence length (to enable PyTorch packing). '''

        seq_tensor = np.array([ob['instr_encoding'] for ob in obs])
        seq_lengths = np.argmax(seq_tensor == padding_idx, axis=1)
        seq_lengths[seq_lengths == 0] = seq_tensor.shape[1]     # Full length

        seq_tensor = torch.from_numpy(seq_tensor)
        seq_lengths = torch.from_numpy(seq_lengths)

        # Sort sequences by lengths
        seq_lengths, perm_idx = seq_lengths.sort(0, True)       # True -> descending
        sorted_tensor = seq_tensor[perm_idx]
        mask = (sorted_tensor == padding_idx)[:,:seq_lengths[0]]    # seq_lengths[0] is the Maximum length

        return Variable(sorted_tensor, requires_grad=False).long().cuda(), \
               mask.byte().cuda(),  \
               list(seq_lengths), list(perm_idx)

    def _feature_variable(self, obs):
        ''' Extract precomputed features into variable. '''
        features = np.empty((len(obs), args.views, args.visual_feat_size + args.angle_feat_size), dtype=np.float32)
        for i, ob in enumerate(obs):
            features[i, :, :] = ob['feature']   # Image feat
        return Variable(torch.from_numpy(features), requires_grad=False).cuda()

    def _candidate_variable(self, obs):
        cand_leng = [len(ob['candidate']) + 1 for ob in obs]       # +1 is for the end
        max_cand_leng = max(cand_leng)

        cand_visual_feat = np.zeros((len(obs), max_cand_leng, args.visual_feat_size), dtype=np.float32)
        cand_angle_feat = np.zeros((len(obs), max_cand_leng, args.angle_feat_size), dtype=np.float32)
        cand_obj_class = np.zeros((len(obs), max_cand_leng, args.top_N_obj), dtype=np.float32)
        near_visual_mask = np.ones((len(obs), max_cand_leng, 5), dtype=np.uint8)
        near_visual_feat = np.zeros((len(obs), max_cand_leng, 5, args.visual_feat_size), dtype=np.float32)
        near_angle_feat = np.zeros((len(obs), max_cand_leng, 5, args.angle_feat_size), dtype=np.float32)

        near_obj_mask = np.ones((len(obs), max_cand_leng, 4), dtype=np.uint8)
        near_obj_class = np.zeros((len(obs), max_cand_leng, 4, args.top_N_obj), dtype=np.float32)
        near_edge_feat = np.zeros((len(obs), max_cand_leng, 4, args.angle_feat_size), dtype=np.float32)
        near_id_feat = np.zeros((len(obs), max_cand_leng, 5),dtype=np.float32)
        for i, ob in enumerate(obs):
            for j, c in enumerate(ob['candidate']):
                cand_visual_feat[i, j, :] = c['visual_feat']
                cand_angle_feat[i, j, :] = c['angle_feat']
                cand_obj_class[i, j, :] = c['obj_class']
                near_visual_mask[i,j,...] = c['near_mask']
                near_visual_feat[i,j,...] = c['near_visual_feat']
                near_angle_feat[i,j,...] = c['near_angle_feat']
                near_obj_mask[i,j,...] = c['near_mask'][1:,...]
                near_obj_class[i,j,...] = c['near_obj_class'][1:,...]
                near_edge_feat[i,j,...] = c['near_edge_feat'][1:,...]
                near_id_feat[i,j,...] = c['near_view_id']

        cand_visual_feat = torch.from_numpy(cand_visual_feat).cuda()
        cand_angle_feat = torch.from_numpy(cand_angle_feat).cuda()
        cand_obj_feat = self.objencoder(torch.from_numpy(cand_obj_class).cuda().long())

        near_visual_mask = torch.from_numpy(near_visual_mask).cuda()
        near_visual_feat = torch.from_numpy(near_visual_feat).cuda()
        near_angle_feat = torch.from_numpy(near_angle_feat).cuda()

        near_obj_mask = torch.from_numpy(near_obj_mask).cuda()
        near_obj_feat = self.objencoder(torch.from_numpy(near_obj_class).cuda().long())
        near_edge_feat = torch.from_numpy(near_edge_feat).cuda()
        near_id_feat = torch.from_numpy(near_id_feat).cuda()
        return cand_leng, cand_visual_feat, cand_angle_feat, cand_obj_feat, \
               near_visual_mask, near_visual_feat, near_angle_feat, \
               near_obj_mask, near_obj_feat, near_edge_feat, near_id_feat

    def get_input_feat(self, obs):
        input_a_t = None
        f_t = self._feature_variable(obs)      # Image features from obs

        cand_leng, cand_visual_feat, cand_angle_feat, cand_obj_feat, \
        near_visual_mask, near_visual_feat, near_angle_feat, \
        near_obj_mask, near_obj_feat, near_edge_feat, near_id_feat = self._candidate_variable(obs)

        return input_a_t, f_t, \
               cand_leng, cand_visual_feat, cand_angle_feat, cand_obj_feat, \
               near_visual_mask, near_visual_feat, near_angle_feat, \
               near_obj_mask, near_obj_feat, near_edge_feat, near_id_feat
    
    def _teacher_action(self, obs, ended):
        """
        Extract teacher actions into variable.
        :param obs: The observation.
        :param ended: Whether the action seq is ended
        :return:
        """
        a = np.zeros(len(obs), dtype=np.int64)
        for i, ob in enumerate(obs):
            if ended[i]:                                            # Just ignore this index
                a[i] = args.ignoreid
            else:
                for k, candidate in enumerate(ob['candidate']):
                    if candidate['viewpointId'] == ob['teacher']:   # Next view point
                        a[i] = k
                        break
                else:   # Stop here
                    assert ob['teacher'] == ob['viewpoint']         # The teacher action should be "STAY HERE"
                    a[i] = len(ob['candidate'])
        return torch.from_numpy(a).cuda()

    def make_equiv_action(self, a_t, perm_obs, perm_idx=None, traj=None):
        """
        Interface between Panoramic view and Egocentric view 
        It will convert the action panoramic view action a_t to equivalent egocentric view actions for the simulator
        """
        if perm_idx is None:
            perm_idx = range(len(perm_obs))
        actions = [[]] * self.env.batch_size # batch * action_len
        max_len = 0 # for padding stop action
        for i, idx in enumerate(perm_idx):
            action = a_t[i]
            if action != -1:            # -1 is the <stop> action
                select_candidate = perm_obs[i]['candidate'][action]
                src_point = perm_obs[i]['viewIndex']
                trg_point = select_candidate['pointId']
                src_level = (src_point) // 12   # The point idx started from 0
                trg_level = (trg_point) // 12
                src_heading = (src_point) % 12
                trg_heading = (trg_point) % 12
                # adjust elevation
                if trg_level > src_level:
                    actions[idx] = actions[idx] + [self.env_actions['up']] * int(trg_level - src_level)
                elif trg_level < src_level:
                    actions[idx] = actions[idx] + [self.env_actions['down']] * int(src_level - trg_level)
                # adjust heading
                if trg_heading > src_heading:
                    dif = trg_heading - src_heading
                    if dif >= 6: # turn left
                        actions[idx] = actions[idx] + [self.env_actions['left']] * int(12 - dif)
                    else: # turn right
                        actions[idx] = actions[idx] + [self.env_actions['right']] * int(dif)
                elif trg_heading < src_heading:
                    dif = src_heading - trg_heading
                    if dif >=6: # turn right
                        actions[idx] = actions[idx] + [self.env_actions['right']] * int(12 - dif)
                    else: # turn left
                        actions[idx] = actions[idx] + [self.env_actions['left']] * int(dif)

                actions[idx] = actions[idx] + [(select_candidate['idx'], 0, 0)]
                max_len = max(max_len, len(actions[idx]))

        for idx in perm_idx:
            if len(actions[idx]) < max_len:
                actions[idx] = actions[idx] + [self.env_actions['<end>']] * (max_len - len(actions[idx]))
        actions = np.array(actions, dtype = 'float32')

        for i in range(max_len):
            cur_actions = actions[:,i]
            cur_actions = list(cur_actions)
            cur_actions = [tuple(a) for a in cur_actions]
            self.env.env.makeActions(cur_actions)
        
        if traj is not None:
            state = self.env.env.sim.getState()
            for i, idx in enumerate(perm_idx):
                action = a_t[i]
                if action != -1:
                    traj[i]['path'].append((state[idx].location.viewpointId, state[idx].heading, state[idx].elevation))

    def rollout(self, train_ml=None, train_rl=True, reset=True, speaker=None):
        """
        :param train_ml:    The weight to train with maximum likelihood
        :param train_rl:    whether use RL in training
        :param reset:       Reset the environment
        :param speaker:     Speaker used in back translation.
                            If the speaker is not None, use back translation.
                            O.w., normal training
        :return:
        """
        if self.feedback == 'teacher' or self.feedback == 'argmax':
            train_rl = False

        if reset:
            # Reset env
            obs = np.array(self.env.reset())
        else:
            obs = np.array(self.env._get_obs())

        batch_size = len(obs)

        if speaker is not None:         # Trigger the self_train mode!
            noise = self.decoder.drop_env(torch.ones(self.feature_size).cuda())
            batch = self.env.batch.copy()
            speaker.env = self.env
            insts = speaker.infer_batch(featdropmask=noise)     # Use the same drop mask in speaker

            # Create fake environments with the generated instruction
            boss = np.ones((batch_size, 1), np.int64) * self.tok.word_to_index['<BOS>']  # First word is <BOS>
            insts = np.concatenate((boss, insts), 1)
            for i, (datum, inst) in enumerate(zip(batch, insts)):
                if inst[-1] != self.tok.word_to_index['<PAD>']: # The inst is not ended!
                    inst[-1] = self.tok.word_to_index['<EOS>']
                datum.pop('instructions')
                datum.pop('instr_encoding')
                datum['instructions'] = self.tok.decode_sentence(inst)
                datum['instr_encoding'] = inst
            obs = np.array(self.env.reset(batch))

        # Reorder the language input for the encoder (do not ruin the original code)
        seq, seq_mask, seq_lengths, perm_idx = self._sort_batch(obs)
        perm_obs = obs[perm_idx]

        ctx, h_t, c_t = self.encoder(seq, seq_lengths)
        row_text= [x['instructions'] for x in perm_obs]
        clip_language_feature = self.CLIP_language(row_text)
        ctx_mask = seq_mask
        self.decoder.egcn.init_weights(h_t.detach())
        # Record starting point
        traj = [{
            'instr_id': ob['instr_id'],
            'path': [(ob['viewpoint'], ob['heading'], ob['elevation'])]
        } for ob in perm_obs]

        # Init the reward shaping
        last_dist = np.zeros(batch_size, np.float32)
        last_ndtw = np.zeros(batch_size, np.float32)
        for i, ob in enumerate(perm_obs):   # The init distance from the view point to the target
            last_dist[i] = ob['distance']
            path_act = [vp[0] for vp in traj[i]['path']]
            if not args.train == "validlistener":
                last_ndtw[i] = self.ndtw_criterion[ob['scan']](path_act, ob['gt_path'], metric='ndtw')

        # Initialization the tracking state
        ended = np.array([False] * batch_size)  # Indices match permuation of the model, not env

        # Init the logs
        rewards = []
        hidden_states = []
        policy_log_probs = []
        policy_score_probs = []
        masks = []
        entropys = []
        ml_loss = 0.
        scorers = []
        entropys_object = []
        h1 = h_t
        input_a_t = torch.zeros([len(obs), args.angle_feat_size]).cuda()
        for t in range(self.episode_len):
            _, f_t, \
            cand_leng, cand_visual_feat, cand_angle_feat, cand_obj_feat, \
            near_visual_mask, near_visual_feat, near_angle_feat, \
            near_obj_mask, near_obj_feat, near_edge_feat, near_id_feat = self.get_input_feat(perm_obs)
            if speaker is not None:       # Apply the env drop mask to the feat
                f_t[..., :-args.angle_feat_size] *= noise
                cand_visual_feat *= noise
                near_visual_feat *= noise
            h_t, c_t, logit, h1,score_policy,scorer,entropy_object = self.decoder(input_a_t, f_t, 
                                               cand_visual_feat, cand_angle_feat, cand_obj_feat,
                                               near_visual_mask, near_visual_feat, near_angle_feat,
                                               near_obj_mask, near_obj_feat, near_edge_feat, near_id_feat,
                                               h_t, h1, c_t,
                                               ctx, ctx_mask, clip_language_feature,
                                               already_dropfeat=(speaker is not None))
            policy_score_probs.append(score_policy)
            hidden_states.append(h_t)
            scorers.append(scorer)
            entropys_object.append(entropy_object)
            # Mask outputs where agent can't move forward
            # Here the logit is [b, max_candidate]
            candidate_mask = utils.length2mask(cand_leng)
            logit.masked_fill_(candidate_mask, -float('inf'))

            # Supervised training
            target = self._teacher_action(perm_obs, ended)
            ml_loss += self.criterion(logit, target)

            # Determine next model inputs
            if self.feedback == 'teacher':
                a_t = target                # teacher forcing
            elif self.feedback == 'argmax': 
                _, a_t = logit.max(1)        # student forcing - argmax
                a_t = a_t.detach()
                log_probs = F.log_softmax(logit, 1)                              # Calculate the log_prob here
                policy_log_probs.append(log_probs.gather(1, a_t.unsqueeze(1)))   # Gather the log_prob for each batch
            elif self.feedback == 'sample':
                probs = F.softmax(logit, 1)    # sampling an action from model
                c = torch.distributions.Categorical(probs)
                self.logs['entropy'].append(c.entropy().sum().item())      # For log
                entropys.append(c.entropy())                                # For optimization
                a_t = c.sample().detach()
                policy_log_probs.append(c.log_prob(a_t))
            else:
                print(self.feedback)
                sys.exit('Invalid feedback option')

            # Prepare environment action
            # NOTE: Env action is in the perm_obs space
            cpu_a_t = a_t.cpu().numpy()
            for i, next_id in enumerate(cpu_a_t):
                if next_id == (cand_leng[i]-1) or next_id == args.ignoreid or ended[i]:    # The last action is <end>
                    cpu_a_t[i] = -1             # Change the <end> and ignore action to -1

            # Make action and get the new state
            input_a_t = []
            for i,a in enumerate(cpu_a_t):
                if a!=-1:
                    input_a_t.append(cand_angle_feat[i,a,:])
                else:
                    input_a_t.append(torch.zeros(args.angle_feat_size).cuda())
            input_a_t = torch.stack(input_a_t)
            self.make_equiv_action(cpu_a_t, perm_obs, perm_idx, traj)
            obs = np.array(self.env._get_obs())
            perm_obs = obs[perm_idx]                    # Perm the obs for the resu

            if train_rl:
                # Calculate the mask and reward
                dist = np.zeros(batch_size, np.float32)
                ndtw_score = np.zeros(batch_size, np.float32)
                reward = np.zeros(batch_size, np.float32)
                mask = np.ones(batch_size, np.float32)
                for i, ob in enumerate(perm_obs):
                    dist[i] = ob['distance']
                    path_act = [vp[0] for vp in traj[i]['path']]
                    ndtw_score[i] = self.ndtw_criterion[ob['scan']](path_act, ob['gt_path'], metric='ndtw')

                    if ended[i]:            # If the action is already finished BEFORE THIS ACTION.
                        reward[i] = 0.
                        mask[i] = 0.
                    else:       # Calculate the reward
                        action_idx = cpu_a_t[i]
                        if action_idx == -1:        # If the action now is end
                            if dist[i] < 3:         # Correct
                                #reward[i] = 2.
                                reward[i] = 2.0 + ndtw_score[i] * 2.0
                            else:                   # Incorrect
                                reward[i] = -2.
                        else:                       # The action is not end
                            reward[i] = - (dist[i] - last_dist[i])      # Change of distance
                            ndtw_reward = ndtw_score[i] - last_ndtw[i]
                            if reward[i] > 0:                           # Quantification
                                #reward[i] = 1
                                reward[i] = 1 + ndtw_reward
                            elif reward[i] < 0:
                                #reward[i] = -1
                                reward[i] = -1 + ndtw_reward
                            else:
                                raise NameError("The action doesn't change the move")
                            if (last_dist[i] <= 1.0) and (dist[i]-last_dist[i] > 0.0):
                                reward[i] -= (1.0 - last_dist[i]) * 2.0
                rewards.append(reward)
                masks.append(mask)
                last_dist[:] = dist
                last_ndtw[:] = ndtw_score

            # Update the finished actions
            # -1 means ended or ignored (already ended)
            ended[:] = np.logical_or(ended, (cpu_a_t == -1))

            # Early exit if all ended
            if ended.all(): 
                break

        if train_rl:
            # Last action in A2C
            _, f_t, \
            cand_leng, cand_visual_feat, cand_angle_feat, cand_obj_feat, \
            near_visual_mask, near_visual_feat, near_angle_feat, \
            near_obj_mask, near_obj_feat, near_edge_feat, near_id_feat  = self.get_input_feat(perm_obs)
            if speaker is not None:       # Apply the env drop mask to the feat
                f_t[..., :-args.angle_feat_size] *= noise
                cand_visual_feat *= noise
                near_visual_feat *= noise

            last_h_, _, _, _, _,scorer,_ = self.decoder(input_a_t, f_t,
                                            cand_visual_feat, cand_angle_feat, cand_obj_feat,
                                            near_visual_mask, near_visual_feat, near_angle_feat,
                                            near_obj_mask, near_obj_feat, near_edge_feat,near_id_feat,  
                                            h_t, h1, c_t,
                                            ctx, ctx_mask,clip_language_feature,
                                            already_dropfeat=(speaker is not None))
            rl_loss = 0.
            rl_loss_object = 0.
            # NOW, A2C!!!
            # Calculate the final discounted reward
            #RLM init
            last_value_object = self.critic_object(scorer).detach()
            discount_reward_object = np.zeros(batch_size, np.float32)
            #agent RL init
            last_value__ = self.critic(last_h_).detach()    # The value esti of the last state, remove the grad for safety
            discount_reward = np.zeros(batch_size, np.float32)  # The inital reward is zero
            
            for i in range(batch_size):
                if not ended[i]:        # If the action is not ended, use the value function as the last reward
                    discount_reward[i] = last_value__[i]
                    discount_reward_object[i] = last_value_object[i]
            
            length = len(rewards)
            total = 0
            for t in range(length-1, -1, -1):
                #RLM RL training
                mask_ = Variable(torch.from_numpy(masks[t]), requires_grad=False).cuda()
                discount_reward_object = discount_reward_object *args.gamma + rewards[t]
                clip_reward_object = discount_reward_object.copy()
                r_object = Variable(torch.from_numpy(clip_reward_object),requires_grad=False).cuda()
                v_object = self.critic_object(scorers[t])
                a_object = (r_object -v_object).detach()
                rl_loss_object +=(-policy_score_probs[t]*a_object*mask_).sum()
                rl_loss_object +=(((r_object - v_object)**2)*mask_).sum()*0.5
                
                discount_reward = discount_reward * args.gamma + rewards[t]   # If it ended, the reward will be 0
                clip_reward = discount_reward.copy()
                r_ = Variable(torch.from_numpy(clip_reward), requires_grad=False).cuda()
                v_ = self.critic(hidden_states[t])
                a_ = (r_ - v_).detach()
                rl_loss +=(-policy_log_probs[t] * a_ * mask_).sum()
                rl_loss +=(((r_ - v_) ** 2) * mask_).sum() * 0.5     # 1/2 L2 loss
                if self.feedback == 'sample':
                    rl_loss += (-0.01 * entropys[t] * mask_).sum()
                    rl_loss += (-0.01 *entropys_object[t] *mask_).sum()
                self.logs['critic_loss'].append((((r_ - v_) ** 2) * mask_).sum().item())

                total = total + np.sum(masks[t])
            rl_loss += rl_loss_object*0.2
            self.logs['total'].append(total)
            # Normalize the loss function
            if args.normalize_loss == 'total':
                rl_loss /= total
            elif args.normalize_loss == 'batch':
                rl_loss /= batch_size
            else:
                assert args.normalize_loss == 'none'

            self.loss += rl_loss
            self.logs['RL_loss'].append(rl_loss.item())

        if train_ml is not None:
            self.loss += ml_loss * train_ml / batch_size
            self.logs['IL_loss'].append((ml_loss * train_ml / batch_size).item())

        if type(self.loss) is int:  # For safety, it will be activated if no losses are added
            self.losses.append(0.)
        else:
            self.losses.append(self.loss.item() / self.episode_len)    # This argument is useless.

        return traj

    def test(self, use_dropout=False, feedback='argmax', allow_cheat=False, iters=None):
        ''' Evaluate once on each instruction in the current environment '''
        self.feedback = feedback
        if use_dropout:
            self.encoder.train()
            self.decoder.train()
            self.critic.train()
        else:
            self.encoder.eval()
            self.decoder.eval()
            self.critic.eval()
        super(Seq2SeqAgent, self).test(iters)

    def zero_grad(self):
        self.loss = 0.
        self.losses = []
        for model, optimizer in zip(self.models, self.optimizers):
            model.train()
            optimizer.zero_grad()

    def accumulate_gradient(self, feedback='teacher', **kwargs):
        if feedback == 'teacher':
            self.feedback = 'teacher'
            self.rollout(train_ml=args.teacher_weight, train_rl=False, **kwargs)
        elif feedback == 'sample':
            self.feedback = 'teacher'
            self.rollout(train_ml=args.ml_weight, train_rl=False, **kwargs)
            self.feedback = 'sample'
            self.rollout(train_ml=None, train_rl=True, **kwargs)
        else:
            assert False

    def optim_step(self):
        self.loss.backward()

        torch.nn.utils.clip_grad_norm(self.encoder.parameters(), 40.)
        torch.nn.utils.clip_grad_norm(self.decoder.parameters(), 40.)

        self.encoder_optimizer.step()
        self.decoder_optimizer.step()
        self.critic_optimizer.step()

    def train(self, n_iters, feedback='teacher', **kwargs):
        ''' Train for a given number of iterations '''
        self.feedback = feedback

        self.encoder.train()
        self.decoder.train()
        self.critic.train()

        self.losses = []
        for iter in range(1, n_iters + 1):

            self.encoder_optimizer.zero_grad()
            self.decoder_optimizer.zero_grad()
            self.critic_optimizer.zero_grad()

            self.loss = 0
            if feedback == 'teacher':
                self.feedback = 'teacher'
                self.rollout(train_ml=args.teacher_weight, train_rl=False, **kwargs)
            elif feedback == 'sample':
                if args.ml_weight != 0:
                    self.feedback = 'teacher'
                    self.rollout(train_ml=args.ml_weight, train_rl=False, **kwargs)
                self.feedback = 'sample'
                self.rollout(train_ml=None, train_rl=True, **kwargs)
            else:
                assert False

            self.loss.backward()

            torch.nn.utils.clip_grad_norm(self.encoder.parameters(), 40.)
            torch.nn.utils.clip_grad_norm(self.decoder.parameters(), 40.)

            self.encoder_optimizer.step()
            self.decoder_optimizer.step()
            self.critic_optimizer.step()

            if args.aug is None:
                utils.print_progress(iter, n_iters+1, prefix='Progress:', suffix='Complete', bar_length=50)
                
    def save(self, epoch, path):
        ''' Snapshot models '''
        the_dir, _ = os.path.split(path)
        os.makedirs(the_dir, exist_ok=True)
        states = {}
        def create_state(name, model, optimizer):
            states[name] = {
                'epoch': epoch + 1,
                'state_dict': model.state_dict(),
                'optimizer': optimizer.state_dict(),
            }
        all_tuple = [("encoder", self.encoder, self.encoder_optimizer),
                     ("decoder", self.decoder, self.decoder_optimizer),
                     ("critic", self.critic, self.critic_optimizer)]
        for param in all_tuple:
            create_state(*param)
        torch.save(states, path)

    def load(self, path):
        ''' Loads parameters (but not training state) '''
        states = torch.load(path)
        def recover_state(name, model, optimizer):
            state = model.state_dict()
            model_keys = set(state.keys())
            load_keys = set(states[name]['state_dict'].keys())
            if model_keys != load_keys:
                print("NOTICE: DIFFERENT KEYS IN THE LISTEREN")
            state.update(states[name]['state_dict'])
            model.load_state_dict(state)
            if args.loadOptim:
                optimizer.load_state_dict(states[name]['optimizer'])
        all_tuple = [("encoder", self.encoder, self.encoder_optimizer),
                     ("decoder", self.decoder, self.decoder_optimizer),
                     ("critic", self.critic, self.critic_optimizer)]
        for param in all_tuple:
            recover_state(*param)
        return states['encoder']['epoch'] - 1