# Copyright 2018 The Google AI Language Team Authors and The HuggingFace Inc. team. # Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """PyTorch MPLUG model. """ import math import os from typing import Tuple import torch import torch.nn.functional as F import torch.utils.checkpoint import transformers from torch import Tensor, device, nn from torch.nn import CrossEntropyLoss from transformers import BertConfig, BertTokenizer from transformers.activations import ACT2FN from transformers.file_utils import (add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_model_forward, replace_return_docstrings) from transformers.modeling_outputs import ( BaseModelOutputWithPastAndCrossAttentions, BaseModelOutputWithPoolingAndCrossAttentions, CausalLMOutputWithCrossAttentions) from transformers.modeling_utils import PreTrainedModel from transformers.utils import logging from modelscope.models.multi_modal.mplug.configuration_mplug import ( HiTeAConfig, MPlugConfig) from modelscope.models.multi_modal.mplug.mvit import MViTv2, MViTv2_Base_config from modelscope.models.multi_modal.mplug.predictor import TextGenerator from modelscope.utils.constant import ModelFile from modelscope.utils.torch_utils import (apply_chunking_to_forward, find_pruneable_heads_and_indices, prune_linear_layer) transformers.logging.set_verbosity_error() logger = logging.get_logger() CONFIG_NAME = 'config.yaml' _CONFIG_FOR_DOC = 'BertConfig' _TOKENIZER_FOR_DOC = 'BertTokenizer' def load_tf_weights_in_bert(model, config, tf_checkpoint_path): """Load tf checkpoints in a pytorch model.""" try: import re import numpy as np import tensorflow as tf except ImportError: logger.error( 'Loading a TensorFlow model in PyTorch, requires TensorFlow to be installed. Please see ' 'https://www.tensorflow.org/install/ for installation instructions.' ) raise tf_path = os.path.abspath(tf_checkpoint_path) logger.info('Converting TensorFlow checkpoint from {}'.format(tf_path)) # Load weights from TF model init_vars = tf.train.list_variables(tf_path) names = [] arrays = [] for name, shape in init_vars: logger.info('Loading TF weight {} with shape {}'.format(name, shape)) array = tf.train.load_variable(tf_path, name) names.append(name) arrays.append(array) for name, array in zip(names, arrays): name = name.split('/') # adam_v and adam_m are variables used in AdamWeightDecayOptimizer to calculated m and v # which are not required for using pretrained model if any(n in [ 'adam_v', 'adam_m', 'AdamWeightDecayOptimizer', 'AdamWeightDecayOptimizer_1', 'global_step' ] for n in name): logger.info('Skipping {}'.format('/'.join(name))) continue pointer = model for m_name in name: if re.fullmatch(r'[A-Za-z]+_\d+', m_name): scope_names = re.split(r'_(\d+)', m_name) else: scope_names = [m_name] if scope_names[0] == 'kernel' or scope_names[0] == 'gamma': pointer = getattr(pointer, 'weight') elif scope_names[0] == 'output_bias' or scope_names[0] == 'beta': pointer = getattr(pointer, 'bias') elif scope_names[0] == 'output_weights': pointer = getattr(pointer, 'weight') elif scope_names[0] == 'squad': pointer = getattr(pointer, 'classifier') else: try: pointer = getattr(pointer, scope_names[0]) except AttributeError: logger.info('Skipping {}'.format('/'.join(name))) continue if len(scope_names) >= 2: num = int(scope_names[1]) pointer = pointer[num] if m_name[-11:] == '_embeddings': pointer = getattr(pointer, 'weight') elif m_name == 'kernel': array = np.transpose(array) try: assert ( pointer.shape == array.shape ), f'Pointer shape {pointer.shape} and array shape {array.shape} mismatched' except AssertionError as e: e.args += (pointer.shape, array.shape) raise logger.info('Initialize PyTorch weight {}'.format(name)) pointer.data = torch.from_numpy(array) return model def clamp_inf(tensor): if tensor.dtype == torch.float16 and torch.isinf(tensor).any(): clamp_value = torch.finfo(tensor.dtype).max - 1000 tensor = torch.clamp(tensor, min=-clamp_value, max=clamp_value) return tensor class BertEmbeddings(nn.Module): """Construct the embeddings from word, position and token_type embeddings.""" def __init__(self, config): super().__init__() self.word_embeddings = nn.Embedding( config.vocab_size, config.hidden_size, padding_idx=config.pad_token_id) self.position_embeddings = nn.Embedding(config.max_position_embeddings, config.hidden_size) self.token_type_embeddings = nn.Embedding(config.type_vocab_size, config.hidden_size) # self.LayerNorm is not snake-cased to stick with TensorFlow model variable name and be able to load # any TensorFlow checkpoint file self.LayerNorm = nn.LayerNorm( config.hidden_size, eps=config.layer_norm_eps) self.dropout = nn.Dropout(config.hidden_dropout_prob) # position_ids (1, len position emb) is contiguous in memory and exported when serialized self.register_buffer( 'position_ids', torch.arange(config.max_position_embeddings).expand((1, -1))) self.position_embedding_type = getattr(config, 'position_embedding_type', 'absolute') self.config = config def forward(self, input_ids=None, token_type_ids=None, position_ids=None, inputs_embeds=None, past_key_values_length=0): if input_ids is not None: input_shape = input_ids.size() else: input_shape = inputs_embeds.size()[:-1] seq_length = input_shape[1] if position_ids is None: position_ids = self.position_ids[:, past_key_values_length:seq_length + past_key_values_length] if token_type_ids is None: token_type_ids = torch.zeros( input_shape, dtype=torch.long, device=self.position_ids.device) if inputs_embeds is None: inputs_embeds = self.word_embeddings(input_ids) token_type_embeddings = self.token_type_embeddings(token_type_ids) embeddings = inputs_embeds + token_type_embeddings if self.position_embedding_type == 'absolute': position_embeddings = self.position_embeddings(position_ids) embeddings += position_embeddings embeddings = self.LayerNorm(embeddings) embeddings = self.dropout(embeddings) return embeddings class BertSelfAttention(nn.Module): def __init__(self, config, is_cross_attention): super().__init__() self.config = config if config.hidden_size % config.num_attention_heads != 0 and not hasattr( config, 'embedding_size'): raise ValueError( 'The hidden size (%d) is not a multiple of the number of attention ' 'heads (%d)' % (config.hidden_size, config.num_attention_heads)) self.num_attention_heads = config.num_attention_heads self.attention_head_size = int(config.hidden_size / config.num_attention_heads) self.all_head_size = self.num_attention_heads * self.attention_head_size self.query = nn.Linear(config.hidden_size, self.all_head_size) if is_cross_attention: self.key = nn.Linear(config.encoder_width, self.all_head_size) self.value = nn.Linear(config.encoder_width, self.all_head_size) else: self.key = nn.Linear(config.hidden_size, self.all_head_size) self.value = nn.Linear(config.hidden_size, self.all_head_size) self.dropout = nn.Dropout(config.attention_probs_dropout_prob) self.position_embedding_type = getattr(config, 'position_embedding_type', 'absolute') if self.position_embedding_type == 'relative_key' or self.position_embedding_type == 'relative_key_query': self.max_position_embeddings = config.max_position_embeddings self.distance_embedding = nn.Embedding( 2 * config.max_position_embeddings - 1, self.attention_head_size) self.save_attention = False def save_attn_gradients(self, attn_gradients): self.attn_gradients = attn_gradients def get_attn_gradients(self): return self.attn_gradients def save_attention_map(self, attention_map): self.attention_map = attention_map def get_attention_map(self): return self.attention_map def transpose_for_scores(self, x): new_x_shape = x.size()[:-1] + (self.num_attention_heads, self.attention_head_size) x = x.view(*new_x_shape) return x.permute(0, 2, 1, 3) def forward( self, hidden_states, attention_mask=None, head_mask=None, encoder_hidden_states=None, encoder_attention_mask=None, past_key_value=None, output_attentions=False, ): mixed_query_layer = self.query(hidden_states) # If this is instantiated as a cross-attention module, the keys # and values come from an encoder; the attention mask needs to be # such that the encoder's padding tokens are not attended to. is_cross_attention = encoder_hidden_states is not None if is_cross_attention: key_layer = self.transpose_for_scores( self.key(encoder_hidden_states)) value_layer = self.transpose_for_scores( self.value(encoder_hidden_states)) attention_mask = encoder_attention_mask elif past_key_value is not None: key_layer = self.transpose_for_scores(self.key(hidden_states)) value_layer = self.transpose_for_scores(self.value(hidden_states)) key_layer = torch.cat([past_key_value[0], key_layer], dim=2) value_layer = torch.cat([past_key_value[1], value_layer], dim=2) else: key_layer = self.transpose_for_scores(self.key(hidden_states)) value_layer = self.transpose_for_scores(self.value(hidden_states)) query_layer = self.transpose_for_scores(mixed_query_layer) past_key_value = (key_layer, value_layer) # Take the dot product between "query" and "key" to get the raw attention scores. attention_scores = torch.matmul(query_layer, key_layer.transpose(-1, -2)) attention_scores = clamp_inf(attention_scores) if self.position_embedding_type == 'relative_key' or self.position_embedding_type == 'relative_key_query': seq_length = hidden_states.size()[1] position_ids_l = torch.arange( seq_length, dtype=torch.long, device=hidden_states.device).view(-1, 1) position_ids_r = torch.arange( seq_length, dtype=torch.long, device=hidden_states.device).view(1, -1) distance = position_ids_l - position_ids_r positional_embedding = self.distance_embedding( distance + self.max_position_embeddings - 1) positional_embedding = positional_embedding.to( dtype=query_layer.dtype) # fp16 compatibility if self.position_embedding_type == 'relative_key': relative_position_scores = torch.einsum( 'bhld,lrd->bhlr', query_layer, positional_embedding) attention_scores = attention_scores + relative_position_scores elif self.position_embedding_type == 'relative_key_query': relative_position_scores_query = torch.einsum( 'bhld,lrd->bhlr', query_layer, positional_embedding) relative_position_scores_key = torch.einsum( 'bhrd,lrd->bhlr', key_layer, positional_embedding) attention_scores = attention_scores + relative_position_scores_query + relative_position_scores_key attention_scores = attention_scores / math.sqrt( self.attention_head_size) if attention_mask is not None: # Apply the attention mask is (precomputed for all layers in BertModel forward() function) attention_scores = attention_scores + attention_mask # Normalize the attention scores to probabilities. attention_probs = nn.Softmax(dim=-1)(attention_scores) if is_cross_attention and self.save_attention: self.save_attention_map(attention_probs) attention_probs.register_hook(self.save_attn_gradients) # This is actually dropping out entire tokens to attend to, which might # seem a bit unusual, but is taken from the original Transformer paper. attention_probs_dropped = self.dropout(attention_probs) # Mask heads if we want to if head_mask is not None: attention_probs_dropped = attention_probs_dropped * head_mask context_layer = torch.matmul(attention_probs_dropped, value_layer) context_layer = context_layer.permute(0, 2, 1, 3).contiguous() new_context_layer_shape = context_layer.size()[:-2] + ( self.all_head_size, ) context_layer = context_layer.view(*new_context_layer_shape) outputs = (context_layer, attention_probs) if output_attentions else (context_layer, ) outputs = outputs + (past_key_value, ) return outputs class BertSelfOutput(nn.Module): def __init__(self, config): super().__init__() self.dense = nn.Linear(config.hidden_size, config.hidden_size) self.LayerNorm = nn.LayerNorm( config.hidden_size, eps=config.layer_norm_eps) self.dropout = nn.Dropout(config.hidden_dropout_prob) def forward(self, hidden_states, input_tensor): hidden_states = self.dense(hidden_states) hidden_states = self.dropout(hidden_states) hidden_states = self.LayerNorm(hidden_states + input_tensor) return hidden_states class BertAttention(nn.Module): def __init__(self, config, is_cross_attention=False): super().__init__() self.self = BertSelfAttention(config, is_cross_attention) self.output = BertSelfOutput(config) self.pruned_heads = set() def prune_heads(self, heads): if len(heads) == 0: return heads, index = find_pruneable_heads_and_indices( heads, self.self.num_attention_heads, self.self.attention_head_size, self.pruned_heads) # Prune linear layers self.self.query = prune_linear_layer(self.self.query, index) self.self.key = prune_linear_layer(self.self.key, index) self.self.value = prune_linear_layer(self.self.value, index) self.output.dense = prune_linear_layer(self.output.dense, index, dim=1) # Update hyper params and store pruned heads self.self.num_attention_heads = self.self.num_attention_heads - len( heads) self.self.all_head_size = self.self.attention_head_size * self.self.num_attention_heads self.pruned_heads = self.pruned_heads.union(heads) def forward( self, hidden_states, attention_mask=None, head_mask=None, encoder_hidden_states=None, encoder_attention_mask=None, past_key_value=None, output_attentions=False, ): self_outputs = self.self( hidden_states, attention_mask, head_mask, encoder_hidden_states, encoder_attention_mask, past_key_value, output_attentions, ) attention_output = self.output(self_outputs[0], hidden_states) outputs = (attention_output, ) + self_outputs[1:] # add attentions if we output them return outputs class BertIntermediate(nn.Module): def __init__(self, config): super().__init__() self.dense = nn.Linear(config.hidden_size, config.intermediate_size) if isinstance(config.hidden_act, str): self.intermediate_act_fn = ACT2FN[config.hidden_act] else: self.intermediate_act_fn = config.hidden_act def forward(self, hidden_states): hidden_states = self.dense(hidden_states) hidden_states = self.intermediate_act_fn(hidden_states) return hidden_states class BertOutput(nn.Module): def __init__(self, config): super().__init__() self.dense = nn.Linear(config.intermediate_size, config.hidden_size) self.LayerNorm = nn.LayerNorm( config.hidden_size, eps=config.layer_norm_eps) self.dropout = nn.Dropout(config.hidden_dropout_prob) def forward(self, hidden_states, input_tensor): hidden_states = self.dense(hidden_states) hidden_states = clamp_inf(hidden_states) hidden_states = self.dropout(hidden_states) hidden_states = clamp_inf(hidden_states) hidden_states = self.LayerNorm(hidden_states + input_tensor) return hidden_states class FusionLayer(nn.Module): def __init__(self, config, layer_num): super().__init__() self.config = config self.stride_layer = getattr(self.config, 'stride_layer', 100) self.chunk_size_feed_forward = config.chunk_size_feed_forward self.seq_len_dim = 1 self.attention = BertAttention(config) self.crossattention = BertAttention(config, is_cross_attention=True) self.intermediate = BertIntermediate(config) self.output = BertOutput(config) def forward( self, hidden_states, attention_mask=None, head_mask=None, encoder_hidden_states=None, encoder_attention_mask=None, layer_nums=None, past_key_value=None, output_attentions=False, ): # decoder uni-directional self-attention cached key/values tuple is at positions 1,2 self_attn_past_key_value = past_key_value[: 2] if past_key_value is not None else None if layer_nums == 0 or layer_nums % self.stride_layer != 0: self_attention_outputs = self.attention( hidden_states, attention_mask, head_mask, output_attentions=output_attentions, past_key_value=self_attn_past_key_value, ) attention_output = self_attention_outputs[0] outputs = self_attention_outputs[1:-1] present_key_value = self_attention_outputs[-1] assert encoder_hidden_states is not None, 'encoder_hidden_states must be given for cross-attention layers' cross_attention_outputs = self.crossattention( attention_output, attention_mask, head_mask, encoder_hidden_states, encoder_attention_mask, output_attentions=output_attentions, ) attention_output = cross_attention_outputs[0] outputs = outputs + cross_attention_outputs[ 1:-1] # add cross attentions if we output attention weights elif layer_nums != 0 and layer_nums % self.stride_layer == 0: self_attention_outputs = self.attention( torch.cat([encoder_hidden_states, hidden_states], 1), torch.cat([encoder_attention_mask, attention_mask], 3), head_mask, output_attentions=output_attentions, past_key_value=self_attn_past_key_value, ) attention_output = self_attention_outputs[0] outputs = self_attention_outputs[1:-1] present_key_value = self_attention_outputs[-1] layer_output = apply_chunking_to_forward(self.feed_forward_chunk, self.chunk_size_feed_forward, self.seq_len_dim, attention_output) outputs = (layer_output, ) + outputs outputs = outputs + (present_key_value[0], present_key_value[1]) return outputs def feed_forward_chunk(self, attention_output): intermediate_output = self.intermediate(attention_output) layer_output = self.output(intermediate_output, attention_output) return layer_output class BertLayer(nn.Module): def __init__(self, config, layer_num): super().__init__() self.config = config self.chunk_size_feed_forward = config.chunk_size_feed_forward self.seq_len_dim = 1 self.attention = BertAttention(config) self.has_cross_attention = getattr(self.config, 'add_cross_attention', False) if self.has_cross_attention: self.crossattention = BertAttention( config, is_cross_attention=True) self.intermediate = BertIntermediate(config) self.output = BertOutput(config) def forward( self, hidden_states, attention_mask=None, head_mask=None, encoder_hidden_states=None, encoder_attention_mask=None, past_key_value=None, output_attentions=False, ): # decoder uni-directional self-attention cached key/values tuple is at positions 1,2 self_attn_past_key_value = past_key_value[: 2] if past_key_value is not None else None self_attention_outputs = self.attention( hidden_states, attention_mask, head_mask, output_attentions=output_attentions, past_key_value=self_attn_past_key_value, ) attention_output = self_attention_outputs[0] outputs = self_attention_outputs[1:-1] present_key_value = self_attention_outputs[-1] if self.has_cross_attention: assert encoder_hidden_states is not None, 'encoder_hidden_states must be given for cross-attention layers' if type(encoder_hidden_states) == list: cross_attention_outputs = self.crossattention( attention_output, attention_mask, head_mask, encoder_hidden_states[(self.layer_num - self.config.fusion_layer) % len(encoder_hidden_states)], encoder_attention_mask[(self.layer_num - self.config.fusion_layer) % len(encoder_hidden_states)], output_attentions=output_attentions, ) attention_output = cross_attention_outputs[0] outputs = outputs + cross_attention_outputs[1:-1] else: cross_attention_outputs = self.crossattention( attention_output, attention_mask, head_mask, encoder_hidden_states, encoder_attention_mask, output_attentions=output_attentions, ) attention_output = cross_attention_outputs[0] outputs = outputs + cross_attention_outputs[ 1: -1] # add cross attentions if we output attention weights layer_output = apply_chunking_to_forward(self.feed_forward_chunk, self.chunk_size_feed_forward, self.seq_len_dim, attention_output) outputs = (layer_output, ) + outputs outputs = outputs + (present_key_value[0], present_key_value[1]) return outputs def feed_forward_chunk(self, attention_output): intermediate_output = self.intermediate(attention_output) layer_output = self.output(intermediate_output, attention_output) return layer_output class FusionEncoder(nn.Module): def __init__(self, config): super().__init__() self.config = config self.layer = nn.ModuleList( [FusionLayer(config, i) for i in range(config.num_hidden_layers)]) self.start_layer = max(0, config.num_hidden_layers - config.fusion_layers) def forward(self, hidden_states, attention_mask=None, head_mask=None, encoder_hidden_states=None, encoder_attention_mask=None, past_key_values=None, use_cache=None, output_attentions=False, output_hidden_states=False, return_dict=True): all_hidden_states = () if output_hidden_states else None all_self_attentions = () if output_attentions else None next_decoder_cache = () if use_cache else None self.stride_layer = getattr(self.config, 'stride_layer', 100) image_length = encoder_hidden_states.shape[1] text_length = hidden_states.shape[1] for i in range(self.start_layer, len(self.layer)): layer_module = self.layer[i] if output_hidden_states: all_hidden_states = all_hidden_states + (hidden_states, ) layer_head_mask = head_mask[i] if head_mask is not None else None past_key_value = past_key_values[ i] if past_key_values is not None else None if getattr(self.config, 'gradient_checkpointing', False) and self.training: if use_cache: logger.warning( '`use_cache=True` is incompatible with `config.gradient_checkpointing=True`. Setting ' '`use_cache=False`...') use_cache = False def create_custom_forward(module): def custom_forward(*inputs): return tuple( module(*inputs, past_key_value, output_attentions)) return custom_forward layer_outputs = torch.utils.checkpoint.checkpoint( create_custom_forward(layer_module), hidden_states, attention_mask, layer_head_mask, encoder_hidden_states, encoder_attention_mask, i - self.start_layer, ) else: layer_outputs = layer_module( hidden_states, attention_mask, layer_head_mask, encoder_hidden_states, encoder_attention_mask, i - self.start_layer, past_key_value, output_attentions, ) hidden_states = layer_outputs[0] if use_cache: next_decoder_cache += (layer_outputs[-1], ) if output_attentions: all_self_attentions = all_self_attentions + ( layer_outputs[1], ) if hidden_states.shape[1] == (image_length + text_length): encoder_hidden_states_new, hidden_states = torch.split( hidden_states, (image_length, text_length), 1) encoder_hidden_states += encoder_hidden_states_new if output_hidden_states: all_hidden_states = all_hidden_states + (hidden_states, ) return [encoder_hidden_states, hidden_states] class BertEncoder(nn.Module): def __init__(self, config): super().__init__() self.config = config self.layer = nn.ModuleList( [BertLayer(config, i) for i in range(config.num_hidden_layers)]) def forward(self, hidden_states, attention_mask=None, head_mask=None, encoder_hidden_states=None, encoder_attention_mask=None, past_key_values=None, use_cache=None, output_attentions=False, output_hidden_states=False, return_dict=True): all_hidden_states = () if output_hidden_states else None all_self_attentions = () if output_attentions else None all_cross_attentions = ( ) if output_attentions and self.config.add_cross_attention else None next_decoder_cache = () if use_cache else None for i in range(len(self.layer)): layer_module = self.layer[i] if output_hidden_states: all_hidden_states = all_hidden_states + (hidden_states, ) layer_head_mask = head_mask[i] if head_mask is not None else None past_key_value = past_key_values[ i] if past_key_values is not None else None if getattr(self.config, 'gradient_checkpointing', False) and self.training: if use_cache: logger.warning( '`use_cache=True` is incompatible with `config.gradient_checkpointing=True`. Setting ' '`use_cache=False`...') use_cache = False def create_custom_forward(module): def custom_forward(*inputs): return tuple( module(*inputs, past_key_value, output_attentions)) return custom_forward layer_outputs = torch.utils.checkpoint.checkpoint( create_custom_forward(layer_module), hidden_states, attention_mask, layer_head_mask, encoder_hidden_states, encoder_attention_mask) else: layer_outputs = layer_module( hidden_states, attention_mask, layer_head_mask, encoder_hidden_states, encoder_attention_mask, past_key_value, output_attentions, ) hidden_states = layer_outputs[0] if use_cache: next_decoder_cache += (layer_outputs[-1], ) if output_attentions: all_self_attentions = all_self_attentions + ( layer_outputs[1], ) if output_hidden_states: all_hidden_states = all_hidden_states + (hidden_states, ) if not return_dict: return tuple(v for v in [ hidden_states, next_decoder_cache, all_hidden_states, all_self_attentions, all_cross_attentions, ] if v is not None) return BaseModelOutputWithPastAndCrossAttentions( last_hidden_state=hidden_states, past_key_values=next_decoder_cache, hidden_states=all_hidden_states, attentions=all_self_attentions, cross_attentions=all_cross_attentions, ) class BertPooler(nn.Module): def __init__(self, config): super().__init__() self.dense = nn.Linear(config.hidden_size, config.hidden_size) self.activation = nn.Tanh() def forward(self, hidden_states): # We "pool" the model by simply taking the hidden state corresponding # to the first token. first_token_tensor = hidden_states[:, 0] pooled_output = self.dense(first_token_tensor) pooled_output = self.activation(pooled_output) return pooled_output class BertPredictionHeadTransform(nn.Module): def __init__(self, config): super().__init__() self.dense = nn.Linear(config.hidden_size, config.hidden_size) if isinstance(config.hidden_act, str): self.transform_act_fn = ACT2FN[config.hidden_act] else: self.transform_act_fn = config.hidden_act self.LayerNorm = nn.LayerNorm( config.hidden_size, eps=config.layer_norm_eps) def forward(self, hidden_states): hidden_states = self.dense(hidden_states) hidden_states = self.transform_act_fn(hidden_states) hidden_states = self.LayerNorm(hidden_states) return hidden_states class BertLMPredictionHead(nn.Module): def __init__(self, config): super().__init__() self.transform = BertPredictionHeadTransform(config) # The output weights are the same as the input embeddings, but there is # an output-only bias for each token. self.decoder = nn.Linear( config.hidden_size, config.vocab_size, bias=False) self.bias = nn.Parameter(torch.zeros(config.vocab_size)) # Need a link between the two variables so that the bias is correctly resized with `resize_token_embeddings` self.decoder.bias = self.bias def forward(self, hidden_states): hidden_states = self.transform(hidden_states) hidden_states = self.decoder(hidden_states) return hidden_states class BertOnlyMLMHead(nn.Module): def __init__(self, config): super().__init__() self.predictions = BertLMPredictionHead(config) def forward(self, sequence_output): prediction_scores = self.predictions(sequence_output) return prediction_scores class BertOnlyNSPHead(nn.Module): def __init__(self, config): super().__init__() self.seq_relationship = nn.Linear(config.hidden_size, 2) def forward(self, pooled_output): seq_relationship_score = self.seq_relationship(pooled_output) return seq_relationship_score class BertPreTrainingHeads(nn.Module): def __init__(self, config): super().__init__() self.predictions = BertLMPredictionHead(config) self.seq_relationship = nn.Linear(config.hidden_size, 2) def forward(self, sequence_output, pooled_output): prediction_scores = self.predictions(sequence_output) seq_relationship_score = self.seq_relationship(pooled_output) return prediction_scores, seq_relationship_score class BertPreTrainedModel(PreTrainedModel): """ An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained models. """ config_class = BertConfig load_tf_weights = load_tf_weights_in_bert base_model_prefix = 'bert' _keys_to_ignore_on_load_missing = [r'position_ids'] def _init_weights(self, module): """ Initialize the weights """ if isinstance(module, (nn.Linear, nn.Embedding)): # Slightly different from the TF version which uses truncated_normal for initialization # cf https://github.com/pytorch/pytorch/pull/5617 module.weight.data.normal_( mean=0.0, std=self.config.initializer_range) elif isinstance(module, nn.LayerNorm): module.bias.data.zero_() module.weight.data.fill_(1.0) if isinstance(module, nn.Linear) and module.bias is not None: module.bias.data.zero_() BERT_START_DOCSTRING = r""" This model inherits from :class:`~transformers.PreTrainedModel`. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.) This model is also a PyTorch `torch.nn.Module `__ subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior. Parameters: config (:class:`~transformers.BertConfig`): Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the :meth:`~transformers.PreTrainedModel.from_pretrained` method to load the model weights. """ BERT_INPUTS_DOCSTRING = r""" Args: input_ids (:obj:`torch.LongTensor` of shape :obj:`({0})`): Indices of input sequence tokens in the vocabulary. Indices can be obtained using :class:`~transformers.BertTokenizer`. See :meth:`transformers.PreTrainedTokenizer.encode` and :meth:`transformers.PreTrainedTokenizer.__call__` for details. `What are input IDs? <../glossary.html#input-ids>`__ attention_mask (:obj:`torch.FloatTensor` of shape :obj:`({0})`, `optional`): Mask to avoid performing attention on padding token indices. Mask values selected in ``[0, 1]``: - 1 for tokens that are **not masked**, - 0 for tokens that are **masked**. `What are attention masks? <../glossary.html#attention-mask>`__ token_type_ids (:obj:`torch.LongTensor` of shape :obj:`({0})`, `optional`): Segment token indices to indicate first and second portions of the inputs. Indices are selected in ``[0, 1]``: - 0 corresponds to a `sentence A` token, - 1 corresponds to a `sentence B` token. `What are token type IDs? <../glossary.html#token-type-ids>`_ position_ids (:obj:`torch.LongTensor` of shape :obj:`({0})`, `optional`): Indices of positions of each input sequence tokens in the position embeddings. Selected in the range ``[0, config.max_position_embeddings - 1]``. `What are position IDs? <../glossary.html#position-ids>`_ head_mask (:obj:`torch.FloatTensor` of shape :obj:`(num_heads,)` or :obj:`(num_layers, num_heads)`, `optional`): Mask to nullify selected heads of the self-attention modules. Mask values selected in ``[0, 1]``: - 1 indicates the head is **not masked**, - 0 indicates the head is **masked**. inputs_embeds (:obj:`torch.FloatTensor` of shape :obj:`({0}, hidden_size)`, `optional`): Optionally, instead of passing :obj:`input_ids` you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert :obj:`input_ids` indices into associated vectors than the model's internal embedding lookup matrix. output_attentions (:obj:`bool`, `optional`): Whether or not to return the attentions tensors of all attention layers. See ``attentions`` under returned tensors for more detail. output_hidden_states (:obj:`bool`, `optional`): Whether or not to return the hidden states of all layers. See ``hidden_states`` under returned tensors for more detail. return_dict (:obj:`bool`, `optional`): Whether or not to return a :class:`~transformers.file_utils.ModelOutput` instead of a plain tuple. """ @add_start_docstrings( 'The bare Bert Model transformer outputting raw hidden-states without any specific head on top.', BERT_START_DOCSTRING, ) class BertModel(BertPreTrainedModel): """ The model can behave as an encoder (with only self-attention) as well as a decoder, in which case a layer of cross-attention is added between the self-attention layers, following the architecture described in `Attention is all you need `__ by Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser and Illia Polosukhin. argument and :obj:`add_cross_attention` set to :obj:`True`; an :obj:`encoder_hidden_states` is then expected as an input to the forward pass. """ def __init__(self, config, add_pooling_layer=True): super().__init__(config) self.config = config self.embeddings = BertEmbeddings(config) self.encoder = BertEncoder(config) self.pooler = BertPooler(config) if add_pooling_layer else None self.init_weights() def get_input_embeddings(self): return self.embeddings.word_embeddings def set_input_embeddings(self, value): self.embeddings.word_embeddings = value def _prune_heads(self, heads_to_prune): """ Prunes heads of the model. heads_to_prune: dict of {layer_num: list of heads to prune in this layer} See base class PreTrainedModel """ for layer, heads in heads_to_prune.items(): self.encoder.layer[layer].attention.prune_heads(heads) @add_start_docstrings_to_model_forward( BERT_INPUTS_DOCSTRING.format('batch_size, sequence_length')) @add_code_sample_docstrings( processor_class=_TOKENIZER_FOR_DOC, checkpoint='bert-base-uncased', output_type=BaseModelOutputWithPoolingAndCrossAttentions, config_class=_CONFIG_FOR_DOC, ) def get_extended_attention_mask(self, attention_mask: Tensor, input_shape: Tuple[int], device: device, is_decoder: bool) -> Tensor: """ Makes broadcastable attention and causal masks so that future and masked tokens are ignored. Arguments: attention_mask (:obj:`torch.Tensor`): Mask with ones indicating tokens to attend to, zeros for tokens to ignore. input_shape (:obj:`Tuple[int]`): The shape of the input to the model. device: (:obj:`torch.device`): The device of the input to the model. Returns: :obj:`torch.Tensor` The extended attention mask, with a the same dtype as :obj:`attention_mask.dtype`. """ # We can provide a self-attention mask of dimensions [batch_size, from_seq_length, to_seq_length] # ourselves in which case we just need to make it broadcastable to all heads. if attention_mask.dim() == 3: extended_attention_mask = attention_mask[:, None, :, :] elif attention_mask.dim() == 2: # Provided a padding mask of dimensions [batch_size, seq_length] # - if the model is a decoder, apply a causal mask in addition to the padding mask # - if the model is an encoder, make the mask broadcastable to # [batch_size, num_heads, seq_length, seq_length] if is_decoder: batch_size, seq_length = input_shape seq_ids = torch.arange(seq_length, device=device) causal_mask = seq_ids[None, None, :].repeat( batch_size, seq_length, 1) <= seq_ids[None, :, None] # in case past_key_values are used we need to add a prefix ones mask to the causal mask # causal and attention masks must have same type with pytorch version < 1.3 causal_mask = causal_mask.to(attention_mask.dtype) if causal_mask.shape[1] < attention_mask.shape[1]: prefix_seq_len = attention_mask.shape[ 1] - causal_mask.shape[1] causal_mask = torch.cat( [ torch.ones( (batch_size, seq_length, prefix_seq_len), device=device, dtype=causal_mask.dtype), causal_mask, ], axis=-1, ) extended_attention_mask = causal_mask[:, None, :, :] * attention_mask[:, None, None, :] else: extended_attention_mask = attention_mask[:, None, None, :] else: raise ValueError( 'Wrong shape for input_ids (shape {}) or attention_mask (shape {})' .format(input_shape, attention_mask.shape)) # Since attention_mask is 1.0 for positions we want to attend and 0.0 for # masked positions, this operation will create a tensor which is 0.0 for # positions we want to attend and -10000.0 for masked positions. # Since we are adding it to the raw scores before the softmax, this is # effectively the same as removing these entirely. extended_attention_mask = extended_attention_mask.to( dtype=self.dtype) # fp16 compatibility extended_attention_mask = (1.0 - extended_attention_mask) * -10000.0 return extended_attention_mask def forward( self, input_ids=None, attention_mask=None, token_type_ids=None, position_ids=None, head_mask=None, inputs_embeds=None, encoder_embeds=None, encoder_hidden_states=None, encoder_attention_mask=None, past_key_values=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None, is_decoder=False, ): r""" encoder_hidden_states (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, sequence_length, hidden_size)`, `optional`): Sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention if the model is configured as a decoder. encoder_attention_mask (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, sequence_length)`, `optional`): Mask to avoid performing attention on the padding token indices of the encoder input. This mask is used in the cross-attention if the model is configured as a decoder. Mask values selected in ``[0, 1]``: - 1 for tokens that are **not masked**, - 0 for tokens that are **masked**. past_key_values (:obj:`tuple(tuple(torch.FloatTensor))` of length :obj:`config.n_layers` with each tuple having 4 tensors of shape :obj:`(batch_size, num_heads, sequence_length - 1, embed_size_per_head)`): Contains precomputed key and value hidden states of the attention blocks. Can be used to speed up decoding. If :obj:`past_key_values` are used, the user can optionally input only the last :obj:`decoder_input_ids` (those that don't have their past key value states given to this model) of shape :obj:`(batch_size, 1)` instead of all :obj:`decoder_input_ids` of shape :obj:`(batch_size, sequence_length)`. use_cache (:obj:`bool`, `optional`): If set to :obj:`True`, :obj:`past_key_values` key value states are returned and can be used to speed up decoding (see :obj:`past_key_values`). """ output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions output_hidden_states = ( output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states) return_dict = return_dict if return_dict is not None else self.config.use_return_dict if is_decoder: use_cache = use_cache if use_cache is not None else self.config.use_cache else: use_cache = False if input_ids is not None and inputs_embeds is not None: raise ValueError( 'You cannot specify both input_ids and inputs_embeds at the same time' ) elif input_ids is not None: input_shape = input_ids.size() batch_size, seq_length = input_shape device = input_ids.device elif inputs_embeds is not None: input_shape = inputs_embeds.size()[:-1] batch_size, seq_length = input_shape device = inputs_embeds.device elif encoder_embeds is not None: input_shape = encoder_embeds.size()[:-1] batch_size, seq_length = input_shape device = encoder_embeds.device else: raise ValueError( 'You have to specify either input_ids or inputs_embeds or encoder_embeds' ) # past_key_values_length past_key_values_length = past_key_values[0][0].shape[ 2] if past_key_values is not None else 0 if attention_mask is None: attention_mask = torch.ones( ((batch_size, seq_length + past_key_values_length)), device=device) if token_type_ids is None: token_type_ids = torch.zeros( input_shape, dtype=torch.long, device=device) # We can provide a self-attention mask of dimensions [batch_size, from_seq_length, to_seq_length] # ourselves in which case we just need to make it broadcastable to all heads. extended_attention_mask: torch.Tensor = self.get_extended_attention_mask( attention_mask, input_shape, device, is_decoder) # If a 2D or 3D attention mask is provided for the cross-attention # we need to make broadcastable to [batch_size, num_heads, seq_length, seq_length] if encoder_hidden_states is not None: if type(encoder_hidden_states) == list: encoder_batch_size, encoder_sequence_length, _ = encoder_hidden_states[ 0].size() else: encoder_batch_size, encoder_sequence_length, _ = encoder_hidden_states.size( ) encoder_hidden_shape = (encoder_batch_size, encoder_sequence_length) if type(encoder_attention_mask) == list: encoder_extended_attention_mask = [ self.invert_attention_mask(mask) for mask in encoder_attention_mask ] elif encoder_attention_mask is None: encoder_attention_mask = torch.ones( encoder_hidden_shape, device=device) encoder_extended_attention_mask = self.invert_attention_mask( encoder_attention_mask) else: encoder_extended_attention_mask = self.invert_attention_mask( encoder_attention_mask) else: encoder_extended_attention_mask = None # Prepare head mask if needed # 1.0 in head_mask indicate we keep the head # attention_probs has shape bsz x n_heads x N x N # input head_mask has shape [num_heads] or [num_hidden_layers x num_heads] # and head_mask is converted to shape [num_hidden_layers x batch x num_heads x seq_length x seq_length] head_mask = self.get_head_mask(head_mask, self.config.num_hidden_layers) if encoder_embeds is None: embedding_output = self.embeddings( input_ids=input_ids, position_ids=position_ids, token_type_ids=token_type_ids, inputs_embeds=inputs_embeds, past_key_values_length=past_key_values_length, ) else: embedding_output = encoder_embeds encoder_outputs = self.encoder( embedding_output, attention_mask=extended_attention_mask, head_mask=head_mask, encoder_hidden_states=encoder_hidden_states, encoder_attention_mask=encoder_extended_attention_mask, past_key_values=past_key_values, use_cache=use_cache, output_attentions=output_attentions, output_hidden_states=output_hidden_states, return_dict=return_dict) sequence_output = encoder_outputs[0] pooled_output = self.pooler( sequence_output) if self.pooler is not None else None if not return_dict: return (sequence_output, pooled_output) + encoder_outputs[1:] return BaseModelOutputWithPoolingAndCrossAttentions( last_hidden_state=sequence_output, pooler_output=pooled_output, past_key_values=encoder_outputs.past_key_values, hidden_states=encoder_outputs.hidden_states, attentions=encoder_outputs.attentions, cross_attentions=encoder_outputs.cross_attentions, ) class FusionModel(BertPreTrainedModel): """ The model can behave as an encoder (with only self-attention) as well as a decoder, in which case a layer of cross-attention is added between the self-attention layers, following the architecture described in `Attention is all you need `__ by Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser and Illia Polosukhin. argument and :obj:`add_cross_attention` set to :obj:`True`; an :obj:`encoder_hidden_states` is then expected as an input to the forward pass. """ def __init__(self, config, add_pooling_layer=True): super().__init__(config) self.config = config self.encoder = FusionEncoder(config) self.pooler = BertPooler(config) if add_pooling_layer else None self.init_weights() def get_input_embeddings(self): return self.embeddings.word_embeddings def set_input_embeddings(self, value): self.embeddings.word_embeddings = value def _prune_heads(self, heads_to_prune): """ Prunes heads of the model. heads_to_prune: dict of {layer_num: list of heads to prune in this layer} See base class PreTrainedModel """ for layer, heads in heads_to_prune.items(): self.encoder.layer[layer].attention.prune_heads(heads) @add_start_docstrings_to_model_forward( BERT_INPUTS_DOCSTRING.format('batch_size, sequence_length')) @add_code_sample_docstrings( # tokenizer_class=_TOKENIZER_FOR_DOC, processor_class=_TOKENIZER_FOR_DOC, checkpoint='bert-base-uncased', output_type=BaseModelOutputWithPoolingAndCrossAttentions, config_class=_CONFIG_FOR_DOC, ) def get_extended_attention_mask(self, attention_mask: Tensor, input_shape: Tuple[int], device: device, is_decoder: bool) -> Tensor: """ Makes broadcastable attention and causal masks so that future and masked tokens are ignored. Arguments: attention_mask (:obj:`torch.Tensor`): Mask with ones indicating tokens to attend to, zeros for tokens to ignore. input_shape (:obj:`Tuple[int]`): The shape of the input to the model. device: (:obj:`torch.device`): The device of the input to the model. Returns: :obj:`torch.Tensor` The extended attention mask, with a the same dtype as :obj:`attention_mask.dtype`. """ # We can provide a self-attention mask of dimensions [batch_size, from_seq_length, to_seq_length] # ourselves in which case we just need to make it broadcastable to all heads. if attention_mask.dim() == 3: extended_attention_mask = attention_mask[:, None, :, :] elif attention_mask.dim() == 2: # Provided a padding mask of dimensions [batch_size, seq_length] # - if the model is a decoder, apply a causal mask in addition to the padding mask # - if the model is an encoder, make the mask broadcastable to # [batch_size, num_heads, seq_length, seq_length] if is_decoder: batch_size, seq_length = input_shape seq_ids = torch.arange(seq_length, device=device) causal_mask = seq_ids[None, None, :].repeat( batch_size, seq_length, 1) <= seq_ids[None, :, None] # in case past_key_values are used we need to add a prefix ones mask to the causal mask # causal and attention masks must have same type with pytorch version < 1.3 causal_mask = causal_mask.to(attention_mask.dtype) if causal_mask.shape[1] < attention_mask.shape[1]: prefix_seq_len = attention_mask.shape[ 1] - causal_mask.shape[1] causal_mask = torch.cat( [ torch.ones( (batch_size, seq_length, prefix_seq_len), device=device, dtype=causal_mask.dtype), causal_mask, ], axis=-1, ) extended_attention_mask = causal_mask[:, None, :, :] * attention_mask[:, None, None, :] else: extended_attention_mask = attention_mask[:, None, None, :] else: raise ValueError( 'Wrong shape for input_ids (shape {}) or attention_mask (shape {})' .format(input_shape, attention_mask.shape)) # Since attention_mask is 1.0 for positions we want to attend and 0.0 for # masked positions, this operation will create a tensor which is 0.0 for # positions we want to attend and -10000.0 for masked positions. # Since we are adding it to the raw scores before the softmax, this is # effectively the same as removing these entirely. extended_attention_mask = extended_attention_mask.to( dtype=self.dtype) # fp16 compatibility extended_attention_mask = (1.0 - extended_attention_mask) * -10000.0 return extended_attention_mask def forward(self, input_ids=None, attention_mask=None, token_type_ids=None, position_ids=None, head_mask=None, inputs_embeds=None, encoder_embeds=None, encoder_hidden_states=None, encoder_attention_mask=None, past_key_values=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None, is_decoder=False): r""" encoder_hidden_states (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, sequence_length, hidden_size)`, `optional`): Sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention if the model is configured as a decoder. encoder_attention_mask (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, sequence_length)`, `optional`): Mask to avoid performing attention on the padding token indices of the encoder input. This mask is used in the cross-attention if the model is configured as a decoder. Mask values selected in ``[0, 1]``: - 1 for tokens that are **not masked**, - 0 for tokens that are **masked**. past_key_values (:obj:`tuple(tuple(torch.FloatTensor))` of length :obj:`config.n_layers` with each tuple having 4 tensors of shape :obj:`(batch_size, num_heads, sequence_length - 1, embed_size_per_head)`): Contains precomputed key and value hidden states of the attention blocks. Can be used to speed up decoding. If :obj:`past_key_values` are used, the user can optionally input only the last :obj:`decoder_input_ids` (those that don't have their past key value states given to this model) of shape :obj:`(batch_size, 1)` instead of all :obj:`decoder_input_ids` of shape :obj:`(batch_size, sequence_length)`. use_cache (:obj:`bool`, `optional`): If set to :obj:`True`, :obj:`past_key_values` key value states are returned and can be used to speed up decoding (see :obj:`past_key_values`). """ output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions output_hidden_states = ( output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states) return_dict = return_dict if return_dict is not None else self.config.use_return_dict if is_decoder: use_cache = use_cache if use_cache is not None else self.config.use_cache else: use_cache = False if input_ids is not None and inputs_embeds is not None: raise ValueError( 'You cannot specify both input_ids and inputs_embeds at the same time' ) elif input_ids is not None: input_shape = input_ids.size() batch_size, seq_length = input_shape device = input_ids.device elif inputs_embeds is not None: input_shape = inputs_embeds.size()[:-1] batch_size, seq_length = input_shape device = inputs_embeds.device elif encoder_embeds is not None: input_shape = encoder_embeds.size()[:-1] batch_size, seq_length = input_shape device = encoder_embeds.device else: raise ValueError( 'You have to specify either input_ids or inputs_embeds or encoder_embeds' ) # past_key_values_length past_key_values_length = past_key_values[0][0].shape[ 2] if past_key_values is not None else 0 if attention_mask is None: attention_mask = torch.ones( ((batch_size, seq_length + past_key_values_length)), device=device) if token_type_ids is None: token_type_ids = torch.zeros( input_shape, dtype=torch.long, device=device) # We can provide a self-attention mask of dimensions [batch_size, from_seq_length, to_seq_length] # ourselves in which case we just need to make it broadcastable to all heads. extended_attention_mask: torch.Tensor = self.get_extended_attention_mask( attention_mask, input_shape, device, is_decoder) # If a 2D or 3D attention mask is provided for the cross-attention # we need to make broadcastable to [batch_size, num_heads, seq_length, seq_length] if encoder_hidden_states is not None: if type(encoder_hidden_states) == list: encoder_batch_size, encoder_sequence_length, _ = encoder_hidden_states[ 0].size() else: encoder_batch_size, encoder_sequence_length, _ = encoder_hidden_states.size( ) encoder_hidden_shape = (encoder_batch_size, encoder_sequence_length) if type(encoder_attention_mask) == list: encoder_extended_attention_mask = [ self.invert_attention_mask(mask) for mask in encoder_attention_mask ] elif encoder_attention_mask is None: encoder_attention_mask = torch.ones( encoder_hidden_shape, device=device) encoder_extended_attention_mask = self.invert_attention_mask( encoder_attention_mask) else: encoder_extended_attention_mask = self.invert_attention_mask( encoder_attention_mask) else: encoder_extended_attention_mask = None # Prepare head mask if needed # 1.0 in head_mask indicate we keep the head # attention_probs has shape bsz x n_heads x N x N # input head_mask has shape [num_heads] or [num_hidden_layers x num_heads] # and head_mask is converted to shape [num_hidden_layers x batch x num_heads x seq_length x seq_length] head_mask = self.get_head_mask(head_mask, self.config.num_hidden_layers) if encoder_embeds is None: embedding_output = self.embeddings( input_ids=input_ids, position_ids=position_ids, token_type_ids=token_type_ids, inputs_embeds=inputs_embeds, past_key_values_length=past_key_values_length, ) else: embedding_output = encoder_embeds encoder_outputs = self.encoder( embedding_output, attention_mask=extended_attention_mask, head_mask=head_mask, encoder_hidden_states=encoder_hidden_states, encoder_attention_mask=encoder_extended_attention_mask, past_key_values=past_key_values, use_cache=use_cache, output_attentions=output_attentions, output_hidden_states=output_hidden_states, return_dict=return_dict, ) encoder_hidden_states, sequence_output = encoder_outputs pooled_output = self.pooler( sequence_output) if self.pooler is not None else None if not return_dict: return [encoder_hidden_states, sequence_output] return BaseModelOutputWithPoolingAndCrossAttentions( last_hidden_state=sequence_output, pooler_output=pooled_output, past_key_values=encoder_outputs.past_key_values, hidden_states=encoder_outputs.hidden_states, attentions=encoder_outputs.attentions, cross_attentions=encoder_outputs.cross_attentions, ) @add_start_docstrings( """Bert Model with a `language modeling` head on top for CLM fine-tuning. """, BERT_START_DOCSTRING) class BertLMHeadModel(BertPreTrainedModel): _keys_to_ignore_on_load_unexpected = [r'pooler'] _keys_to_ignore_on_load_missing = [ r'position_ids', r'predictions.decoder.bias' ] def __init__(self, config): super().__init__(config) self.bert = BertModel(config, add_pooling_layer=False) self.cls = BertOnlyMLMHead(config) self.init_weights() def get_output_embeddings(self): return self.cls.predictions.decoder def set_output_embeddings(self, new_embeddings): self.cls.predictions.decoder = new_embeddings @add_start_docstrings_to_model_forward( BERT_INPUTS_DOCSTRING.format('batch_size, sequence_length')) @replace_return_docstrings( output_type=CausalLMOutputWithCrossAttentions, config_class=_CONFIG_FOR_DOC) def forward( self, input_ids=None, attention_mask=None, token_type_ids=None, position_ids=None, head_mask=None, inputs_embeds=None, encoder_hidden_states=None, encoder_attention_mask=None, labels=None, past_key_values=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None, is_decoder=True, reduction='mean', soft_labels=None, alpha=0, return_logits=False, ): r""" encoder_hidden_states (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, sequence_length, hidden_size)`, `optional`): Sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention if the model is configured as a decoder. encoder_attention_mask (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, sequence_length)`, `optional`): Mask to avoid performing attention on the padding token indices of the encoder input. This mask is used in the cross-attention if the model is configured as a decoder. Mask values selected in ``[0, 1]``: - 1 for tokens that are **not masked**, - 0 for tokens that are **masked**. labels (:obj:`torch.LongTensor` of shape :obj:`(batch_size, sequence_length)`, `optional`): Labels for computing the left-to-right language modeling loss (next word prediction). Indices should be in ``[-100, 0, ..., config.vocab_size]`` (see ``input_ids`` docstring) Tokens with indices set to ``-100`` are ignored (masked), the loss is only computed for the tokens with labels n ``[0, ..., config.vocab_size]`` past_key_values (:obj:`tuple(tuple(torch.FloatTensor))` of length :obj:`config.n_layers` with each tuple having 4 tensors of shape :obj:`(batch_size, num_heads, sequence_length - 1, embed_size_per_head)`): Contains precomputed key and value hidden states of the attention blocks. Can be used to speed up decoding. If :obj:`past_key_values` are used, the user can optionally input only the last :obj:`decoder_input_ids` (those that don't have their past key value states given to this model) of shape :obj:`(batch_size, 1)` instead of all :obj:`decoder_input_ids` of shape :obj:`(batch_size, sequence_length)`. use_cache (:obj:`bool`, `optional`): If set to :obj:`True`, :obj:`past_key_values` key value states are returned and can be used to speed up decoding (see :obj:`past_key_values`). Returns: Example: >>> from transformers import BertTokenizer, BertLMHeadModel, BertConfig >>> import torch >>> tokenizer = BertTokenizer.from_pretrained('bert-base-cased') >>> config = BertConfig.from_pretrained("bert-base-cased") >>> model = BertLMHeadModel.from_pretrained('bert-base-cased', config=config) >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") >>> outputs = model(**inputs) >>> prediction_logits = outputs.logits """ return_dict = return_dict if return_dict is not None else self.config.use_return_dict if labels is not None: use_cache = False outputs = self.bert( input_ids, attention_mask=attention_mask, token_type_ids=token_type_ids, position_ids=position_ids, head_mask=head_mask, inputs_embeds=inputs_embeds, encoder_hidden_states=encoder_hidden_states, encoder_attention_mask=encoder_attention_mask, past_key_values=past_key_values, use_cache=use_cache, output_attentions=output_attentions, output_hidden_states=output_hidden_states, return_dict=return_dict, is_decoder=is_decoder, ) sequence_output = outputs[0] prediction_scores = self.cls(sequence_output) if return_logits: return prediction_scores[:, :-1, :].contiguous() lm_loss = None if labels is not None: # we are doing next-token prediction; shift prediction scores and input ids by one shifted_prediction_scores = prediction_scores[:, : -1, :].contiguous() labels = labels[:, 1:].contiguous() loss_fct = CrossEntropyLoss(reduction=reduction) lm_loss = loss_fct( shifted_prediction_scores.view(-1, self.config.vocab_size), labels.view(-1)) lm_loss = lm_loss.view(prediction_scores.size(0), -1).sum(1) if soft_labels is not None: loss_distill = -torch.sum( F.log_softmax(shifted_prediction_scores, dim=1) * soft_labels, dim=-1) loss_distill = (loss_distill * (labels != -100)).sum(1) lm_loss = (1 - alpha) * lm_loss + alpha * loss_distill if not return_dict: output = (prediction_scores, ) + outputs[2:] return ((lm_loss, ) + output) if lm_loss is not None else output return CausalLMOutputWithCrossAttentions( loss=lm_loss, logits=prediction_scores, past_key_values=outputs.past_key_values, hidden_states=outputs.hidden_states, attentions=outputs.attentions, cross_attentions=outputs.cross_attentions, ) def prepare_inputs_for_generation(self, input_ids, past=None, attention_mask=None, **model_kwargs): input_shape = input_ids.shape # if model is used as a decoder in encoder-decoder model, the decoder attention mask is created on the fly if attention_mask is None: attention_mask = input_ids.new_ones(input_shape) # cut decoder_input_ids if past is used if past is not None: input_ids = input_ids[:, -1:] return { 'input_ids': input_ids, 'attention_mask': attention_mask, 'past_key_values': past, 'encoder_hidden_states': model_kwargs.get('encoder_hidden_states', None), 'encoder_attention_mask': model_kwargs.get('encoder_attention_mask', None), 'is_decoder': True, } def _reorder_cache(self, past, beam_idx): reordered_past = () for layer_past in past: reordered_past += (tuple( past_state.index_select(0, beam_idx) for past_state in layer_past), ) return reordered_past class BertPrefixModel(BertPreTrainedModel): _keys_to_ignore_on_load_unexpected = [r'pooler'] _keys_to_ignore_on_load_missing = [ r'position_ids', r'predictions.decoder.bias' ] def __init__(self, config): super().__init__(config) self.bert = BertModel(config, add_pooling_layer=False) self.cls = BertOnlyMLMHead(config) self.init_weights() def get_output_embeddings(self): return self.cls.predictions.decoder def set_output_embeddings(self, new_embeddings): self.cls.predictions.decoder = new_embeddings @add_start_docstrings_to_model_forward( BERT_INPUTS_DOCSTRING.format('batch_size, sequence_length')) @add_code_sample_docstrings( processor_class=_TOKENIZER_FOR_DOC, checkpoint='bert-base-uncased', output_type=CausalLMOutputWithCrossAttentions, config_class=_CONFIG_FOR_DOC, ) def forward( self, input_ids=None, attention_mask=None, token_type_ids=None, position_ids=None, head_mask=None, inputs_embeds=None, encoder_hidden_states=None, encoder_attention_mask=None, labels=None, past_key_values=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None, is_decoder=True, reduction='mean', soft_labels=None, alpha=0, return_logits=False, ): return_dict = return_dict if return_dict is not None else self.config.use_return_dict if labels is not None: use_cache = False outputs = self.bert( input_ids, attention_mask=attention_mask, token_type_ids=token_type_ids, position_ids=position_ids, head_mask=head_mask, inputs_embeds=inputs_embeds, encoder_hidden_states=encoder_hidden_states, encoder_attention_mask=encoder_attention_mask, past_key_values=past_key_values, use_cache=use_cache, output_attentions=output_attentions, output_hidden_states=output_hidden_states, return_dict=return_dict, is_decoder=is_decoder, ) sequence_output = outputs[0] prediction_scores = self.cls(sequence_output) if return_logits: return prediction_scores[:, :-1, :].contiguous() lm_loss = None if labels is not None: # we are doing next-token prediction; shift prediction scores and input ids by one shifted_prediction_scores = prediction_scores[:, : -1, :].contiguous() labels = labels[:, 1:].contiguous() loss_fct = CrossEntropyLoss() lm_loss = loss_fct( shifted_prediction_scores.view(-1, self.config.vocab_size), labels.view(-1)) if soft_labels is not None: loss_distill = -torch.sum( F.log_softmax(shifted_prediction_scores, dim=1) * soft_labels, dim=-1) loss_distill = loss_distill[labels != -100].mean() lm_loss = (1 - alpha) * lm_loss + alpha * loss_distill if not return_dict: output = (prediction_scores, ) + outputs[2:] return ((lm_loss, ) + output) if lm_loss is not None else output return CausalLMOutputWithCrossAttentions( loss=lm_loss, logits=prediction_scores, past_key_values=outputs.past_key_values, hidden_states=outputs.hidden_states, attentions=outputs.attentions, cross_attentions=outputs.cross_attentions, ) class MPlug(PreTrainedModel): config_class = MPlugConfig def __init__(self, config): super().__init__(config) self.config = config self.tokenizer = BertTokenizer.from_pretrained( os.path.join(config.model_dir, ModelFile.VOCAB_FILE)) self.module_setting(config) self.visual_encoder = self._initialize_clip(config) self.text_encoder = BertModel( self.config_encoder, add_pooling_layer=False) self.fusion_encoder = FusionModel( self.config_fusion, add_pooling_layer=False) @classmethod def from_pretrained(cls, model_dir, task=None, load_checkpoint=True): from modelscope.utils.constant import Tasks task_mapping = { Tasks.visual_question_answering: MPlugForVisualQuestionAnswering, Tasks.image_captioning: MPlugForImageCaption, Tasks.image_text_retrieval: MPlugForImageTextRetrieval, } config = cls.config_class.from_yaml_file( os.path.join(model_dir, CONFIG_NAME)) config.model_dir = model_dir if task is None: task = config.task model = task_mapping[task](config) if load_checkpoint: checkpoint_path = os.path.join(model_dir, ModelFile.TORCH_MODEL_BIN_FILE) checkpoint = torch.load(checkpoint_path, map_location='cpu') if 'model' in checkpoint: checkpoint = checkpoint['model'] if 'module' in checkpoint: checkpoint = checkpoint['module'] checkpoint = { k.replace('model.', ''): v for k, v in checkpoint.items() } msg = model.load_state_dict(checkpoint, strict=False) print('load checkpoint from %s' % checkpoint_path) print(msg) return model @staticmethod def _initialize_clip(config, num_patches=240): def resize_pos_embed(posemb, posemb_new): # Rescale the grid of position embeddings when loading from state_dict. Adapted from # https://github.com/google-research/vision_transformer/blob/00883dd691c63a6830751563748663526e811cee/vit_jax/checkpoint.py#L224 ntok_new = posemb_new.shape[1] posemb_tok, posemb_grid = posemb[:, :1], posemb[0, 1:] ntok_new -= 1 gs_old = int(math.sqrt(len(posemb_grid))) gs_new = int(math.sqrt(ntok_new)) # _logger.info('Position embedding grid-size from %s to %s', gs_old, gs_new) posemb_grid = posemb_grid.reshape(1, gs_old, gs_old, -1).permute(0, 3, 1, 2) orig = posemb_grid.dtype posemb_grid = F.interpolate( posemb_grid.float(), size=(gs_new, gs_new), mode='bilinear') posemb_grid = posemb_grid.to(orig) posemb_grid = posemb_grid.permute(0, 2, 3, 1).reshape( 1, gs_new * gs_new, -1) posemb = torch.cat([posemb_tok, posemb_grid], dim=1) return posemb from .clip import clip clip_model = clip.load_from_config(config) if 'ViT-B-16' in config.clip_name: num_patches = int(config.image_res * config.image_res / (16 * 16)) pos_embed = nn.Parameter(torch.zeros(num_patches + 1, 768).float()) else: num_patches = int(config.image_res * config.image_res / (14 * 14)) pos_embed = nn.Parameter( torch.zeros(num_patches + 1, 1024).float()) pos_embed.weight = resize_pos_embed( clip_model.visual.positional_embedding.unsqueeze(0), pos_embed.unsqueeze(0)) clip_model.visual.positional_embedding = pos_embed return clip_model def init_distill(self, config): self.distill = config.distill if self.distill: self.visual_encoder_m = self._initialize_clip(config) self.text_encoder_m = BertModel( self.config_encoder, add_pooling_layer=False) self.fusion_encoder_m = FusionModel( self.config_fusion, add_pooling_layer=False) self.text_decoder_m = BertLMHeadModel(self.config_decoder) self.model_pairs = [ [self.visual_encoder, self.visual_encoder_m], [self.text_encoder, self.text_encoder_m], [self.text_decoder, self.text_decoder_m], ] if self.config_encoder.hidden_size != config.vision_width: self.visn_fc_m = nn.Linear(config.vision_width, self.config_encoder.hidden_size) self.visn_layer_norm_m = nn.LayerNorm( self.config_encoder.hidden_size, eps=1e-12) self.dropout_m = nn.Dropout( self.config_encoder.hidden_dropout_prob) self.model_pairs.extend( [[self.visn_fc, self.visn_fc_m], [self.visn_layer_norm, self.visn_layer_norm_m]]) self.copy_params() self.momentum = 0.995 def forward(self, *args, **kwargs): raise NotImplementedError def module_setting(self, config): bert_config_path = os.path.join(config.model_dir, config.bert_config) self.config_encoder = BertConfig.from_json_file(bert_config_path) self.config_encoder.num_hidden_layers = self.config_encoder.text_encoder_layers self.config_fusion = BertConfig.from_json_file(bert_config_path) self.config_decoder = BertConfig.from_json_file(bert_config_path) self.config_decoder.add_cross_attention = True self.config_decoder.num_hidden_layers = self.config_decoder.text_decode_layers self.large = False if self.config_encoder.hidden_size != config.vision_width: self.visn_fc = nn.Linear(config.vision_width, self.config_encoder.hidden_size) self.visn_layer_norm = nn.LayerNorm( self.config_encoder.hidden_size, eps=1e-12) self.dropout = nn.Dropout(self.config_encoder.hidden_dropout_prob) self.large = True @torch.no_grad() def copy_params(self): for model_pair in self.model_pairs: for param, param_m in zip(model_pair[0].parameters(), model_pair[1].parameters()): param_m.data.copy_(param.data) # initialize param_m.requires_grad = False # not update by gradient @torch.no_grad() def _momentum_update(self): for model_pair in self.model_pairs: for param, param_m in zip(model_pair[0].parameters(), model_pair[1].parameters()): param_m.data = param_m.data * self.momentum + param.data * ( 1. - self.momentum) def generation(self, question_states, question_atts, out_size=1): encoder_inputs = [question_states, question_atts] topk_ids, topk_scores = self.beam_generator.translate_batch( encoder_inputs, out_size=out_size) return topk_ids, topk_scores @staticmethod def _tile(x, dim, n_tile): import numpy as np init_dim = x.size(dim) repeat_idx = [1] * x.dim() repeat_idx[dim] = n_tile x = x.repeat(*(repeat_idx)) order_index = torch.LongTensor( np.concatenate( [init_dim * np.arange(n_tile) + i for i in range(init_dim)])) return torch.index_select(x, dim, order_index.to(x.device)) class MPlugForVisualQuestionAnswering(MPlug): def __init__(self, config): super().__init__(config) self.text_decoder = BertLMHeadModel(self.config_decoder) self.beam_generator = TextGenerator(config, self.text_decoder) self.init_distill(config) def forward(self, image, question, answer=None, alpha=0, k=None, weights=None, train=True): image = image.to(dtype=next(self.parameters()).dtype) image_embeds = self.visual_encoder.visual(image, skip_last_layer=True) if self.large: image_embeds = self.dropout( self.visn_layer_norm(self.visn_fc(image_embeds))) image_atts = torch.ones( image_embeds.size()[:-1], dtype=torch.long).to(image.device) if train: ''' k: number of answers for each question weights: weight for each answer ''' answer_targets = answer.input_ids.masked_fill( answer.input_ids == self.tokenizer.pad_token_id, -100) text_output = self.text_encoder( question.input_ids, attention_mask=question.attention_mask, return_dict=True) text_embeds = text_output.last_hidden_state fusion_output = self.fusion_encoder( encoder_embeds=text_embeds, attention_mask=question.attention_mask, encoder_hidden_states=image_embeds, encoder_attention_mask=image_atts, return_dict=False) image_output, question_output = fusion_output question_output = torch.cat([image_output, question_output], 1) merge_text_attention = torch.cat( [image_atts, question.attention_mask], 1) if k is None: k = [1] * question_output.shape[0] question_states = [] question_atts = [] for b, n in enumerate(k): question_states += [question_output[b]] * n question_atts += [merge_text_attention[b]] * n question_states = torch.stack(question_states, 0) question_atts = torch.stack(question_atts, 0) if self.distill: with torch.no_grad(): self._momentum_update() image_embeds_m = self.visual_encoder_m.visual( image, skip_last_layer=True) if self.large: image_embeds_m = self.dropout_m( self.visn_layer_norm_m( self.visn_fc_m(image_embeds_m))) text_output_m = self.text_encoder_m( question.input_ids, attention_mask=question.attention_mask, return_dict=True) text_embeds_m = text_output_m.last_hidden_state fusion_output_m = self.fusion_encoder_m( encoder_embeds=text_embeds_m, attention_mask=question.attention_mask, encoder_hidden_states=image_embeds_m, encoder_attention_mask=image_atts, return_dict=False) image_output_m, question_output_m = fusion_output_m question_output_m = torch.cat( [image_output_m, question_output_m], 1) question_states_m = [] for b, n in enumerate(k): question_states_m += [question_output_m[b]] * n question_states_m = torch.stack(question_states_m, 0) logits_m = self.text_decoder_m( answer.input_ids, attention_mask=answer.attention_mask, encoder_hidden_states=question_states_m, encoder_attention_mask=question_atts, return_logits=True, ) answer_output = self.text_decoder( answer.input_ids, attention_mask=answer.attention_mask, encoder_hidden_states=question_states, encoder_attention_mask=question_atts, labels=answer_targets, return_dict=True, soft_labels=F.softmax(logits_m, dim=-1), reduction='none', ) else: answer_output = self.text_decoder( answer.input_ids, attention_mask=answer.attention_mask, encoder_hidden_states=question_states, encoder_attention_mask=question_atts, labels=answer_targets, return_dict=True, reduction='none', ) if weights is None: weights = 1 loss = weights * answer_output.loss loss = loss.sum() / image.size(0) return loss else: text_output = self.text_encoder( question.input_ids, attention_mask=question.attention_mask, return_dict=True) text_embeds = text_output.last_hidden_state fusion_output = self.fusion_encoder( encoder_embeds=text_embeds, attention_mask=question.attention_mask, encoder_hidden_states=image_embeds, encoder_attention_mask=image_atts, return_dict=False) image_output, question_output = fusion_output question_output = torch.cat([image_output, question_output], 1) merge_text_attention = torch.cat( [image_atts, question.attention_mask], 1) topk_ids, topk_probs = self.generation(question_output, merge_text_attention) return topk_ids, topk_probs class MPlugForImageCaption(MPlug): def __init__(self, config): super().__init__(config) self.text_decoder = BertPrefixModel(self.config_decoder) self.beam_generator = TextGenerator(config, self.text_decoder) def beam_search(self, image, question, answer=None, train=True, out_size=5): image_embeds = self.visual_encoder.visual(image, skip_last_layer=True) if self.large: image_embeds = self.dropout( self.visn_layer_norm(self.visn_fc(image_embeds))) image_atts = torch.ones( image_embeds.size()[:-1], dtype=torch.long).to(image.device) text_output = self.text_encoder( question.input_ids, attention_mask=question.attention_mask, return_dict=True) text_embeds = text_output.last_hidden_state fusion_output = self.fusion_encoder( encoder_embeds=text_embeds, attention_mask=question.attention_mask, encoder_hidden_states=image_embeds, encoder_attention_mask=image_atts, return_dict=False) image_output, question_output = fusion_output question_output = torch.cat([image_output, question_output], 1) merge_text_attention = torch.cat([image_atts, question.attention_mask], 1) topk_ids, topk_probs = self.generation( question_output, merge_text_attention, out_size=out_size) return topk_ids, topk_probs def forward(self, image, question, answer=None, train=True, out_size=5, scst=False): if (scst): return self.beam_search( image, question, answer, train=True, out_size=out_size) image = image.to(dtype=next(self.parameters()).dtype) image_embeds = self.visual_encoder.visual(image, skip_last_layer=True) if self.large: image_embeds = self.dropout( self.visn_layer_norm(self.visn_fc(image_embeds))) image_atts = torch.ones( image_embeds.size()[:-1], dtype=torch.long).to(image.device) if train: answer_targets = answer.input_ids.masked_fill( answer.input_ids == self.tokenizer.pad_token_id, -100) answer_output = self.text_decoder( answer.input_ids, attention_mask=answer.attention_mask, encoder_hidden_states=image_embeds, encoder_attention_mask=image_atts, labels=answer_targets, return_dict=True, reduction='none') loss = answer_output.loss return loss else: topk_ids, topk_probs = self.generation(image_embeds, image_atts) return topk_ids, topk_probs class MPlugForImageTextRetrieval(MPlug): def __init__(self, config): super().__init__(config) self.embed_dim = config.embed_dim self.temp = nn.Parameter(torch.ones([]) * config.temp) self.queue_size = config.queue_size self.momentum = config.momentum self.alpha = config.alpha self.queue_size = config.queue_size self.text_width = self.config_encoder.hidden_size self.embed_dim = config.embed_dim self.vision_proj = nn.Linear(self.text_width, self.embed_dim) self.text_proj = nn.Linear(self.text_width, self.embed_dim) self.itm_head = nn.Linear(self.text_width, 2) self.register_buffer('image_queue', torch.randn(self.embed_dim, self.queue_size)) self.register_buffer('text_queue', torch.randn(self.embed_dim, self.queue_size)) self.register_buffer('idx_queue', torch.full((1, self.queue_size), -100)) self.register_buffer('queue_ptr', torch.zeros(1, dtype=torch.long)) self.image_queue = F.normalize(self.image_queue, dim=0) self.text_queue = F.normalize(self.text_queue, dim=0) self.init_distill(config) def init_distill(self, config): self.distill = config.distill if self.distill: self.visual_encoder_m = self._initialize_clip(config) self.text_encoder_m = BertModel( self.config_encoder, add_pooling_layer=False) self.fusion_encoder_m = FusionModel( self.config_fusion, add_pooling_layer=False) self.vision_proj_m = nn.Linear(self.text_width, self.embed_dim) self.text_proj_m = nn.Linear(self.text_width, self.embed_dim) self.model_pairs = [ [self.visual_encoder, self.visual_encoder_m], [self.text_encoder, self.text_encoder_m], [self.text_proj, self.text_proj_m], [self.vision_proj, self.vision_proj_m], ] if self.config_encoder.hidden_size != config.vision_width: self.visn_fc_m = nn.Linear(config.vision_width, self.config_encoder.hidden_size) self.visn_layer_norm_m = nn.LayerNorm( self.config_encoder.hidden_size, eps=1e-12) self.dropout_m = nn.Dropout( self.config_encoder.hidden_dropout_prob) self.model_pairs.extend( [[self.visn_fc, self.visn_fc_m], [self.visn_layer_norm, self.visn_layer_norm_m]]) self.copy_params() self.momentum = 0.995 @torch.no_grad() def _dequeue_and_enqueue(self, image_feat, text_feat, idx): def concat_all_gather(tensor): """ Performs all_gather operation on the provided tensors. *** Warning ***: torch.distributed.all_gather has no gradient. """ if not torch.distributed.is_initialized(): return tensor tensors_gather = [ torch.ones_like(tensor) for _ in range(torch.distributed.get_world_size()) ] torch.distributed.all_gather( tensors_gather, tensor, async_op=False) output = torch.cat(tensors_gather, dim=0) return output # gather keys before updating queue image_feats = concat_all_gather(image_feat) text_feats = concat_all_gather(text_feat) idxs = concat_all_gather(idx) batch_size = image_feats.shape[0] ptr = int(self.queue_ptr) # assert self.queue_size % batch_size == 0 # for simplicity # replace the keys at ptr (dequeue and enqueue) self.image_queue[:, ptr:ptr + batch_size] = image_feats.T self.text_queue[:, ptr:ptr + batch_size] = text_feats.T self.idx_queue[:, ptr:ptr + batch_size] = idxs.T ptr = (ptr + batch_size) % self.queue_size # move pointer self.queue_ptr[0] = ptr def forward(self, image, text, idx=None, train=True): if train: image_embeds = self.visual_encoder.visual( image, skip_last_layer=True) if self.large: image_embeds = self.dropout( self.visn_layer_norm(self.visn_fc(image_embeds))) image_atts = torch.ones( image_embeds.size()[:-1], dtype=torch.long).to(image.device) image_feat = F.normalize( self.vision_proj(image_embeds[:, 0, :]), dim=-1) text_output = self.text_encoder( text.input_ids, attention_mask=text.attention_mask, return_dict=True) text_embeds = text_output.last_hidden_state text_feat = F.normalize( self.text_proj(text_embeds[:, 0, :]), dim=-1) idx = idx.view(-1, 1) idx_all = torch.cat( [idx.t(), self.idx_queue.clone().detach()], dim=1) pos_idx = torch.eq(idx, idx_all).float() sim_targets = pos_idx / pos_idx.sum(1, keepdim=True) with torch.no_grad(): self._momentum_update() image_embeds_m = self.visual_encoder_m.visual( image, skip_last_layer=True) if self.large: image_embeds_m = self.dropout_m( self.visn_layer_norm_m(self.visn_fc_m(image_embeds_m))) image_feat_m = F.normalize( self.vision_proj_m(image_embeds_m[:, 0, :]), dim=-1) image_feat_all = torch.cat( [image_feat_m.t(), self.image_queue.clone().detach()], dim=1) text_output_m = self.text_encoder_m( text.input_ids, attention_mask=text.attention_mask, return_dict=True) text_feat_m = F.normalize( self.text_proj_m(text_output_m.last_hidden_state[:, 0, :]), dim=-1) text_feat_all = torch.cat( [text_feat_m.t(), self.text_queue.clone().detach()], dim=1) if self.distill: sim_i2t_m = image_feat_m @ text_feat_all / self.temp sim_t2i_m = text_feat_m @ image_feat_all / self.temp sim_i2t_targets = self.alpha * F.softmax( sim_i2t_m, dim=1) + (1 - self.alpha) * sim_targets sim_t2i_targets = self.alpha * F.softmax( sim_t2i_m, dim=1) + (1 - self.alpha) * sim_targets sim_i2t = image_feat @ text_feat_all / self.temp sim_t2i = text_feat @ image_feat_all / self.temp if self.distill: loss_i2t = -torch.sum( F.log_softmax(sim_i2t, dim=1) * sim_i2t_targets, dim=1).mean() loss_t2i = -torch.sum( F.log_softmax(sim_t2i, dim=1) * sim_t2i_targets, dim=1).mean() else: loss_i2t = -torch.sum( F.log_softmax(sim_i2t, dim=1) * sim_targets, dim=1).mean() loss_t2i = -torch.sum( F.log_softmax(sim_t2i, dim=1) * sim_targets, dim=1).mean() loss_ita = (loss_i2t + loss_t2i) / 2 self._dequeue_and_enqueue(image_feat_m, text_feat_m, idx) # forward the positive image-text pair _, output_pos = self.fusion_encoder( encoder_embeds=text_embeds, attention_mask=text.attention_mask, encoder_hidden_states=image_embeds, encoder_attention_mask=image_atts, return_dict=False, ) with torch.no_grad(): bs = image.size(0) weights_i2t = F.softmax(sim_i2t[:, :bs], dim=1) weights_t2i = F.softmax(sim_t2i[:, :bs], dim=1) mask = torch.eq(idx, idx.T) weights_i2t.masked_fill_(mask, 0) weights_t2i.masked_fill_(mask, 0) # select a negative image for each text image_embeds_neg = [] for b in range(bs): neg_idx = torch.multinomial(weights_t2i[b], 1).item() image_embeds_neg.append(image_embeds[neg_idx]) image_embeds_neg = torch.stack(image_embeds_neg, dim=0) # select a negative text for each image text_embeds_neg = [] text_atts_neg = [] for b in range(bs): neg_idx = torch.multinomial(weights_i2t[b], 1).item() text_embeds_neg.append(text_embeds[neg_idx]) text_atts_neg.append(text.attention_mask[neg_idx]) text_embeds_neg = torch.stack(text_embeds_neg, dim=0) text_atts_neg = torch.stack(text_atts_neg, dim=0) text_embeds_all = torch.cat([text_embeds, text_embeds_neg], dim=0) text_atts_all = torch.cat([text.attention_mask, text_atts_neg], dim=0) image_embeds_all = torch.cat([image_embeds_neg, image_embeds], dim=0) image_atts_all = torch.cat([image_atts, image_atts], dim=0) _, output_neg = self.fusion_encoder( encoder_embeds=text_embeds_all, attention_mask=text_atts_all, encoder_hidden_states=image_embeds_all, encoder_attention_mask=image_atts_all, return_dict=False, ) vl_embeddings = torch.cat( [output_pos[:, 0, :], output_neg[:, 0, :]], dim=0) vl_output = self.itm_head(vl_embeddings) ones_tmp = torch.ones(bs, dtype=torch.long) zeros_tmp = torch.zeros(2 * bs, dtype=torch.long) itm_labels = torch.cat([ones_tmp, zeros_tmp], dim=0).to(image.device) loss_itm = F.cross_entropy(vl_output, itm_labels) return loss_ita + loss_itm else: text_output = self.text_encoder( text.input_ids, attention_mask=text.attention_mask) text_feat = text_output.last_hidden_state image_feat = self.visual_encoder.visual( image, skip_last_layer=True) image_feat = self.visn_layer_norm(self.visn_fc(image_feat)) image_att = torch.ones( image_feat.size()[:-1], dtype=torch.long, device=image_feat.device) _, output = self.fusion_encoder( encoder_embeds=text_feat, attention_mask=text.attention_mask, encoder_hidden_states=image_feat, encoder_attention_mask=image_att, return_dict=False, ) scores = self.itm_head(output[:, 0, :]) scores = F.softmax(scores, dim=-1) return scores class HiTeA(PreTrainedModel): config_class = HiTeAConfig def __init__(self, config): super().__init__(config) self.config = config self.tokenizer = BertTokenizer.from_pretrained( os.path.join(config.model_dir, ModelFile.VOCAB_FILE)) self.module_setting(config) self.visual_encoder = MViTv2( img_size=config.image_res, config=MViTv2_Base_config, num_frames=config.num_frames) self.text_encoder = BertModel( self.config_encoder, add_pooling_layer=False) self.fusion_encoder = FusionModel( self.config_fusion, add_pooling_layer=False) @classmethod def from_pretrained(cls, model_dir, load_checkpoint=True): from modelscope.utils.constant import Tasks task_mapping = { Tasks.video_question_answering: HiTeAForVideoQuestionAnswering, Tasks.video_captioning: HiTeAForVideoCaption, } config = cls.config_class.from_yaml_file( os.path.join(model_dir, CONFIG_NAME)) config.model_dir = model_dir model = task_mapping[config.task](config) if load_checkpoint: checkpoint_path = os.path.join(model_dir, ModelFile.TORCH_MODEL_BIN_FILE) checkpoint = torch.load(checkpoint_path, map_location='cpu') if 'model' in checkpoint: checkpoint = checkpoint['model'] if 'module' in checkpoint: checkpoint = checkpoint['module'] checkpoint = { k.replace('model.', ''): v for k, v in checkpoint.items() } model.load_state_dict(checkpoint, strict=False) return model def init_distill(self, config): self.distill = config.distill if self.distill: self.visual_encoder_m = MViTv2( img_size=config.image_res, config=MViTv2_Base_config, num_frames=config.num_frames) self.text_encoder_m = BertModel( self.config_encoder, add_pooling_layer=False) self.fusion_encoder_m = FusionModel( self.config_fusion, add_pooling_layer=False) self.text_decoder_m = BertLMHeadModel(self.config_decoder) self.model_pairs = [ [self.visual_encoder, self.visual_encoder_m], [self.text_encoder, self.text_encoder_m], [self.text_decoder, self.text_decoder_m], ] self.copy_params() self.momentum = 0.995 def forward(self, *args, **kwargs): raise NotImplementedError def module_setting(self, config): bert_config_path = os.path.join(config.model_dir, config.bert_config) self.config_encoder = BertConfig.from_json_file(bert_config_path) self.config_encoder.num_hidden_layers = self.config_encoder.text_encoder_layers self.config_fusion = BertConfig.from_json_file(bert_config_path) self.config_decoder = BertConfig.from_json_file(bert_config_path) self.config_decoder.add_cross_attention = True self.config_decoder.num_hidden_layers = self.config_decoder.text_decode_layers @torch.no_grad() def copy_params(self): for model_pair in self.model_pairs: for param, param_m in zip(model_pair[0].parameters(), model_pair[1].parameters()): param_m.data.copy_(param.data) # initialize param_m.requires_grad = False # not update by gradient @torch.no_grad() def _momentum_update(self): for model_pair in self.model_pairs: for param, param_m in zip(model_pair[0].parameters(), model_pair[1].parameters()): param_m.data = param_m.data * self.momentum + param.data * ( 1. - self.momentum) def generation(self, question_states, question_atts, out_size=1): encoder_inputs = [question_states, question_atts] topk_ids, topk_scores = self.beam_generator.translate_batch( encoder_inputs, out_size=out_size) return topk_ids, topk_scores @staticmethod def _tile(x, dim, n_tile): import numpy as np init_dim = x.size(dim) repeat_idx = [1] * x.dim() repeat_idx[dim] = n_tile x = x.repeat(*(repeat_idx)) order_index = torch.LongTensor( np.concatenate( [init_dim * np.arange(n_tile) + i for i in range(init_dim)])) return torch.index_select(x, dim, order_index.to(x.device)) class HiTeAForVideoQuestionAnswering(HiTeA): def __init__(self, config): super().__init__(config) self.text_decoder = BertLMHeadModel(self.config_decoder) self.beam_generator = TextGenerator(config, self.text_decoder) self.init_distill(config) def forward(self, video, question, answer=None, alpha=0, k=None, weights=None, train=True): video = video.to(dtype=next(self.parameters()).dtype) video_embeds = self.visual_encoder(video) video_atts = torch.ones( video_embeds.size()[:-1], dtype=torch.long).to(video.device) if train: ''' k: number of answers for each question weights: weight for each answer ''' answer_targets = answer.input_ids.masked_fill( answer.input_ids == self.tokenizer.pad_token_id, -100) text_output = self.text_encoder( question.input_ids, attention_mask=question.attention_mask, return_dict=True) text_embeds = text_output.last_hidden_state fusion_output = self.fusion_encoder( encoder_embeds=text_embeds, attention_mask=question.attention_mask, encoder_hidden_states=video_embeds, encoder_attention_mask=video_atts, return_dict=False) video_output, question_output = fusion_output question_output = torch.cat([video_output, question_output], 1) merge_text_attention = torch.cat( [video_atts, question.attention_mask], 1) if k is None: k = [1] * question_output.shape[0] question_states = [] question_atts = [] for b, n in enumerate(k): question_states += [question_output[b]] * n question_atts += [merge_text_attention[b]] * n question_states = torch.stack(question_states, 0) question_atts = torch.stack(question_atts, 0) if self.distill: with torch.no_grad(): self._momentum_update() video_embeds_m = self.visual_encoder_m(video) text_output_m = self.text_encoder_m( question.input_ids, attention_mask=question.attention_mask, return_dict=True) text_embeds_m = text_output_m.last_hidden_state fusion_output_m = self.fusion_encoder_m( encoder_embeds=text_embeds_m, attention_mask=question.attention_mask, encoder_hidden_states=video_embeds_m, encoder_attention_mask=video_atts, return_dict=False) image_output_m, question_output_m = fusion_output_m question_output_m = torch.cat( [image_output_m, question_output_m], 1) question_states_m = [] for b, n in enumerate(k): question_states_m += [question_output_m[b]] * n question_states_m = torch.stack(question_states_m, 0) logits_m = self.text_decoder_m( answer.input_ids, attention_mask=answer.attention_mask, encoder_hidden_states=question_states_m, encoder_attention_mask=question_atts, return_logits=True, ) answer_output = self.text_decoder( answer.input_ids, attention_mask=answer.attention_mask, encoder_hidden_states=question_states, encoder_attention_mask=question_atts, labels=answer_targets, return_dict=True, soft_labels=F.softmax(logits_m, dim=-1), reduction='none', ) else: answer_output = self.text_decoder( answer.input_ids, attention_mask=answer.attention_mask, encoder_hidden_states=question_states, encoder_attention_mask=question_atts, labels=answer_targets, return_dict=True, reduction='none', ) if weights is None: weights = 1 loss = weights * answer_output.loss loss = loss.sum() / video.size(0) return loss else: text_output = self.text_encoder( question.input_ids, attention_mask=question.attention_mask, return_dict=True) text_embeds = text_output.last_hidden_state fusion_output = self.fusion_encoder( encoder_embeds=text_embeds, attention_mask=question.attention_mask, encoder_hidden_states=video_embeds, encoder_attention_mask=video_atts, return_dict=False) video_output, question_output = fusion_output question_output = torch.cat([video_output, question_output], 1) merge_text_attention = torch.cat( [video_atts, question.attention_mask], 1) topk_ids, topk_probs = self.generation(question_output, merge_text_attention) return topk_ids, topk_probs class HiTeAForVideoCaption(HiTeA): def __init__(self, config): super().__init__(config) self.text_decoder = BertPrefixModel(self.config_decoder) self.beam_generator = TextGenerator(config, self.text_decoder) def beam_search(self, video, question, answer=None, train=True, out_size=5): video_embeds = self.visual_encoder(video) video_atts = torch.ones( video_embeds.size()[:-1], dtype=torch.long).to(video.device) text_output = self.text_encoder( question.input_ids, attention_mask=question.attention_mask, return_dict=True) text_embeds = text_output.last_hidden_state fusion_output = self.fusion_encoder( encoder_embeds=text_embeds, attention_mask=question.attention_mask, encoder_hidden_states=video_embeds, encoder_attention_mask=video_atts, return_dict=False) video_output, question_output = fusion_output question_output = torch.cat([video_output, question_output], 1) merge_text_attention = torch.cat([video_atts, question.attention_mask], 1) topk_ids, topk_probs = self.generation( question_output, merge_text_attention, out_size=out_size) return topk_ids, topk_probs def forward(self, video, question, answer=None, train=True, out_size=5, scst=False): if (scst): return self.beam_search( video, question, answer, train=True, out_size=out_size) video = video.to(dtype=next(self.parameters()).dtype) video_embeds = self.visual_encoder(video) video_atts = torch.ones( video_embeds.size()[:-1], dtype=torch.long).to(video.device) if train: answer_targets = answer.input_ids.masked_fill( answer.input_ids == self.tokenizer.pad_token_id, -100) answer_output = self.text_decoder( answer.input_ids, attention_mask=answer.attention_mask, encoder_hidden_states=video_embeds, encoder_attention_mask=video_atts, labels=answer_targets, return_dict=True, reduction='none') loss = answer_output.loss return loss else: topk_ids, topk_probs = self.generation(video_embeds, video_atts) return topk_ids, topk_probs