models/babyai11.py

import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.autograd import Variable
from torch.distributions.categorical import Categorical
from torch.nn.utils.rnn import pack_padded_sequence, pad_packed_sequence
from utils.babyai_utils.supervised_losses import required_heads
import torch_ac



# From https://github.com/ikostrikov/pytorch-a2c-ppo-acktr/blob/master/model.py
def initialize_parameters(m):
    classname = m.__class__.__name__
    if classname.find('Linear') != -1:
        m.weight.data.normal_(0, 1)
        m.weight.data *= 1 / torch.sqrt(m.weight.data.pow(2).sum(1, keepdim=True))
        if m.bias is not None:
            m.bias.data.fill_(0)


# Inspired by FiLMedBlock from https://arxiv.org/abs/1709.07871
class FiLM(nn.Module):
    def __init__(self, in_features, out_features, in_channels, imm_channels):
        super().__init__()
        self.conv1 = nn.Conv2d(
            in_channels=in_channels, out_channels=imm_channels,
            kernel_size=(3, 3), padding=1)
        self.bn1 = nn.BatchNorm2d(imm_channels)
        self.conv2 = nn.Conv2d(
            in_channels=imm_channels, out_channels=out_features,
            kernel_size=(3, 3), padding=1)
        self.bn2 = nn.BatchNorm2d(out_features)

        self.weight = nn.Linear(in_features, out_features)
        self.bias = nn.Linear(in_features, out_features)

        self.apply(initialize_parameters)

    def forward(self, x, y):
        x = F.relu(self.bn1(self.conv1(x)))
        x = self.conv2(x)
        weight = self.weight(y).unsqueeze(2).unsqueeze(3)
        bias = self.bias(y).unsqueeze(2).unsqueeze(3)
        out = x * weight + bias
        return F.relu(self.bn2(out))


class ImageBOWEmbedding(nn.Module):
    def __init__(self, space, embedding_dim):
        super().__init__()
        self.max_value = max(space)
        self.space = space
        self.embedding_dim = embedding_dim
        self.embedding = nn.Embedding(len(self.space) * self.max_value, embedding_dim)
        self.apply(initialize_parameters)

    def forward(self, inputs):
        offsets = torch.Tensor([x * self.max_value for x in range(self.space[-1])]).to(inputs.device)
        inputs = (inputs + offsets[None, :, None, None]).long()
        return self.embedding(inputs).sum(1).permute(0, 3, 1, 2)

#notes: what they call instr is what we call text

#class ACModel(nn.Module, babyai.rl.RecurrentACModel):
class Baby11ACModel(nn.Module, torch_ac.RecurrentACModel):
    def __init__(self, obs_space, action_space,
                 image_dim=128, memory_dim=128, instr_dim=128,
                 use_instr=False, lang_model="gru", use_memory=False,
                 arch="bow_endpool_res", aux_info=None):
        super().__init__()

        # store config
        self.config = locals()

        endpool = 'endpool' in arch
        use_bow = 'bow' in arch
        pixel = 'pixel' in arch
        self.res = 'res' in arch

        # Decide which components are enabled
        self.use_instr = use_instr
        self.use_memory = use_memory
        self.arch = arch
        self.lang_model = lang_model
        self.aux_info = aux_info
        self.env_action_space = action_space
        self.model_raw_action_space = action_space
        if self.res and image_dim != 128:
            raise ValueError(f"image_dim is {image_dim}, expected 128")
        self.image_dim = image_dim
        self.memory_dim = memory_dim
        self.instr_dim = instr_dim

        self.obs_space = obs_space
        # transform given 3d obs_space into what babyai11 baseline uses, i.e. 1d embedding size
        n = obs_space["image"][0]
        m = obs_space["image"][1]
        nb_img_channels = self.obs_space['image'][2]
        self.obs_space = ((n-1)//2-2)*((m-1)//2-2)*64

        for part in self.arch.split('_'):
            if part not in ['original', 'bow', 'pixels', 'endpool', 'res']:
                raise ValueError("Incorrect architecture name: {}".format(self.arch))

        # if not self.use_instr:
        #     raise ValueError("FiLM architecture can be used when instructions are enabled")
        self.image_conv = nn.Sequential(*[
            *([ImageBOWEmbedding(obs_space['image'], 128)] if use_bow else []),
            *([nn.Conv2d(
                in_channels=nb_img_channels, out_channels=128, kernel_size=(8, 8),
                stride=8, padding=0)] if pixel else []),
            nn.Conv2d(
                in_channels=128 if use_bow or pixel else nb_img_channels, out_channels=128,
                kernel_size=(3, 3) if endpool else (2, 2), stride=1, padding=1),
            nn.BatchNorm2d(128),
            nn.ReLU(),
            *([] if endpool else [nn.MaxPool2d(kernel_size=(2, 2), stride=2)]),
            nn.Conv2d(in_channels=128, out_channels=128, kernel_size=(3, 3), padding=1),
            nn.BatchNorm2d(128),
            nn.ReLU(),
            *([] if endpool else [nn.MaxPool2d(kernel_size=(2, 2), stride=2)])
        ])
        self.film_pool = nn.MaxPool2d(kernel_size=(7, 7) if endpool else (2, 2), stride=2)

        # Define instruction embedding
        if self.use_instr:
            if self.lang_model in ['gru', 'bigru', 'attgru']:
                #self.word_embedding = nn.Embedding(obs_space["instr"], self.instr_dim)
                self.word_embedding = nn.Embedding(obs_space["text"], self.instr_dim)
                if self.lang_model in ['gru', 'bigru', 'attgru']:
                    gru_dim = self.instr_dim
                    if self.lang_model in ['bigru', 'attgru']:
                        gru_dim //= 2
                    self.instr_rnn = nn.GRU(
                        self.instr_dim, gru_dim, batch_first=True,
                        bidirectional=(self.lang_model in ['bigru', 'attgru']))
                    self.final_instr_dim = self.instr_dim
                else:
                    kernel_dim = 64
                    kernel_sizes = [3, 4]
                    self.instr_convs = nn.ModuleList([
                        nn.Conv2d(1, kernel_dim, (K, self.instr_dim)) for K in kernel_sizes])
                    self.final_instr_dim = kernel_dim * len(kernel_sizes)

            if self.lang_model == 'attgru':
                self.memory2key = nn.Linear(self.memory_size, self.final_instr_dim)

            num_module = 2
            self.controllers = []
            for ni in range(num_module):
                mod = FiLM(
                    in_features=self.final_instr_dim,
                    out_features=128 if ni < num_module-1 else self.image_dim,
                    in_channels=128, imm_channels=128)
                self.controllers.append(mod)
                self.add_module('FiLM_' + str(ni), mod)

        # Define memory and resize image embedding
        self.embedding_size = self.image_dim
        if self.use_memory:
            self.memory_rnn = nn.LSTMCell(self.image_dim, self.memory_dim)
            self.embedding_size = self.semi_memory_size

        # Define actor's model
        self.actor = nn.Sequential(
            nn.Linear(self.embedding_size, 64),
            nn.Tanh(),
            nn.Linear(64, action_space.nvec[0])
        )

        # Define critic's model
        self.critic = nn.Sequential(
            nn.Linear(self.embedding_size, 64),
            nn.Tanh(),
            nn.Linear(64, 1)
        )

        # Initialize parameters correctly
        self.apply(initialize_parameters)

        # Define head for extra info
        if self.aux_info:
            self.extra_heads = None
            self.add_heads()

    def add_heads(self):
        '''
        When using auxiliary tasks, the environment yields at each step some binary, continous, or multiclass
        information. The agent needs to predict those information. This function add extra heads to the model
        that output the predictions. There is a head per extra information (the head type depends on the extra
        information type).
        '''
        self.extra_heads = nn.ModuleDict()
        for info in self.aux_info:
            if required_heads[info] == 'binary':
                self.extra_heads[info] = nn.Linear(self.embedding_size, 1)
            elif required_heads[info].startswith('multiclass'):
                n_classes = int(required_heads[info].split('multiclass')[-1])
                self.extra_heads[info] = nn.Linear(self.embedding_size, n_classes)
            elif required_heads[info].startswith('continuous'):
                if required_heads[info].endswith('01'):
                    self.extra_heads[info] = nn.Sequential(nn.Linear(self.embedding_size, 1), nn.Sigmoid())
                else:
                    raise ValueError('Only continous01 is implemented')
            else:
                raise ValueError('Type not supported')
            # initializing these parameters independently is done in order to have consistency of results when using
            # supervised-loss-coef = 0 and when not using any extra binary information
            self.extra_heads[info].apply(initialize_parameters)

    def add_extra_heads_if_necessary(self, aux_info):
        '''
        This function allows using a pre-trained model without aux_info and add aux_info to it and still make
        it possible to finetune.
        '''
        try:
            if not hasattr(self, 'aux_info') or not set(self.aux_info) == set(aux_info):
                self.aux_info = aux_info
                self.add_heads()
        except Exception:
            raise ValueError('Could not add extra heads')

    @property
    def memory_size(self):
        return 2 * self.semi_memory_size

    @property
    def semi_memory_size(self):
        return self.memory_dim

    def forward(self, obs, memory, instr_embedding=None):
        if self.use_instr and instr_embedding is None:
            #instr_embedding = self._get_instr_embedding(obs.instr)
            instr_embedding = self._get_instr_embedding(obs.text)
        if self.use_instr and self.lang_model == "attgru":
            # outputs: B x L x D
            # memory: B x M
            #mask = (obs.instr != 0).float()
            mask = (obs.text != 0).float()
            # The mask tensor has the same length as obs.instr, and
            # thus can be both shorter and longer than instr_embedding.
            # It can be longer if instr_embedding is computed
            # for a subbatch of obs.instr.
            # It can be shorter if obs.instr is a subbatch of
            # the batch that instr_embeddings was computed for.
            # Here, we make sure that mask and instr_embeddings
            # have equal length along dimension 1.
            mask = mask[:, :instr_embedding.shape[1]]
            instr_embedding = instr_embedding[:, :mask.shape[1]]

            keys = self.memory2key(memory)
            pre_softmax = (keys[:, None, :] * instr_embedding).sum(2) + 1000 * mask
            attention = F.softmax(pre_softmax, dim=1)
            instr_embedding = (instr_embedding * attention[:, :, None]).sum(1)

        x = torch.transpose(torch.transpose(obs.image, 1, 3), 2, 3)

        if 'pixel' in self.arch:
            x /= 256.0
        x = self.image_conv(x)
        if self.use_instr:
            for controller in self.controllers:
                out = controller(x, instr_embedding)
                if self.res:
                    out += x
                x = out
        x = F.relu(self.film_pool(x))
        x = x.reshape(x.shape[0], -1)

        if self.use_memory:
            hidden = (memory[:, :self.semi_memory_size], memory[:, self.semi_memory_size:])
            hidden = self.memory_rnn(x, hidden)
            embedding = hidden[0]
            memory = torch.cat(hidden, dim=1)
        else:
            embedding = x

        if hasattr(self, 'aux_info') and self.aux_info:
            extra_predictions = {info: self.extra_heads[info](embedding) for info in self.extra_heads}
        else:
            extra_predictions = dict()

        x = self.actor(embedding)
        dist = Categorical(logits=F.log_softmax(x, dim=1))

        x = self.critic(embedding)
        value = x.squeeze(1)

        #return {'dist': dist, 'value': value, 'memory': memory, 'extra_predictions': extra_predictions}
        return [dist], value, memory

    def _get_instr_embedding(self, instr):
        lengths = (instr != 0).sum(1).long()
        if self.lang_model == 'gru':
            out, _ = self.instr_rnn(self.word_embedding(instr))
            hidden = out[range(len(lengths)), lengths-1, :]
            return hidden

        elif self.lang_model in ['bigru', 'attgru']:
            masks = (instr != 0).float()

            if lengths.shape[0] > 1:
                seq_lengths, perm_idx = lengths.sort(0, descending=True)
                iperm_idx = torch.LongTensor(perm_idx.shape).fill_(0)
                if instr.is_cuda: iperm_idx = iperm_idx.cuda()
                for i, v in enumerate(perm_idx):
                    iperm_idx[v.data] = i

                inputs = self.word_embedding(instr)
                inputs = inputs[perm_idx]

                inputs = pack_padded_sequence(inputs, seq_lengths.data.cpu().numpy(), batch_first=True)

                outputs, final_states = self.instr_rnn(inputs)
            else:
                instr = instr[:, 0:lengths[0]]
                outputs, final_states = self.instr_rnn(self.word_embedding(instr))
                iperm_idx = None
            final_states = final_states.transpose(0, 1).contiguous()
            final_states = final_states.view(final_states.shape[0], -1)
            if iperm_idx is not None:
                outputs, _ = pad_packed_sequence(outputs, batch_first=True)
                outputs = outputs[iperm_idx]
                final_states = final_states[iperm_idx]

            return outputs if self.lang_model == 'attgru' else final_states

        else:
            ValueError("Undefined instruction architecture: {}".format(self.use_instr))

    # add action sampling to fit our interaction pipeline
    def sample_action(self, dist):
        return torch.stack([d.sample() for d in dist], dim=1)

    # add construct final action to fit our interaction pipeline
    def construct_final_action(self, action):
        return action

    # add calculate log probs to fit our interaction pipeline
    def calculate_log_probs(self, dist, action):
        return torch.stack([d.log_prob(action[:, i]) for i, d in enumerate(dist)], dim=1)

    # add calculate action masks to fit our interaction pipeline
    def calculate_action_masks(self, action):
        mask = torch.ones_like(action)
        assert action.shape == mask.shape
        return mask

    def get_config_dict(self):
        del self.config['__class__']
        self.config['self'] = str(self.config['self'])
        self.config['action_space'] = self.config['action_space'].nvec.tolist()
        return self.config