|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
""" PyTorch RoFormer model.""" |
|
|
|
|
|
import math |
|
import os |
|
from typing import Optional, Tuple, Union |
|
|
|
import numpy as np |
|
import torch |
|
import torch.utils.checkpoint |
|
from torch import nn |
|
from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss |
|
|
|
from transformers.activations import ACT2FN |
|
from transformers.modeling_outputs import ( |
|
BaseModelOutputWithPastAndCrossAttentions, |
|
CausalLMOutputWithCrossAttentions, |
|
MaskedLMOutput, |
|
MultipleChoiceModelOutput, |
|
QuestionAnsweringModelOutput, |
|
SequenceClassifierOutput, |
|
TokenClassifierOutput, |
|
) |
|
from transformers.modeling_utils import PreTrainedModel, SequenceSummary |
|
from transformers.pytorch_utils import apply_chunking_to_forward, find_pruneable_heads_and_indices, prune_linear_layer |
|
from transformers.utils import ( |
|
add_code_sample_docstrings, |
|
add_start_docstrings, |
|
add_start_docstrings_to_model_forward, |
|
logging, |
|
replace_return_docstrings, |
|
) |
|
from transformers import RoFormerConfig |
|
|
|
|
|
logger = logging.get_logger(__name__) |
|
|
|
_CHECKPOINT_FOR_DOC = "junnyu/roformer_chinese_base" |
|
_CONFIG_FOR_DOC = "RoFormerConfig" |
|
_TOKENIZER_FOR_DOC = "RoFormerTokenizer" |
|
|
|
ROFORMER_PRETRAINED_MODEL_ARCHIVE_LIST = [ |
|
"junnyu/roformer_chinese_small", |
|
"junnyu/roformer_chinese_base", |
|
"junnyu/roformer_chinese_char_small", |
|
"junnyu/roformer_chinese_char_base", |
|
"junnyu/roformer_small_discriminator", |
|
"junnyu/roformer_small_generator" |
|
|
|
] |
|
|
|
|
|
|
|
class RoFormerSinusoidalPositionalEmbedding(nn.Embedding): |
|
"""This module produces sinusoidal positional embeddings of any length.""" |
|
|
|
def __init__(self, num_positions: int, embedding_dim: int, padding_idx: Optional[int] = None) -> None: |
|
super().__init__(num_positions, embedding_dim) |
|
self.weight = self._init_weight(self.weight) |
|
|
|
@staticmethod |
|
def _init_weight(out: nn.Parameter) -> nn.Parameter: |
|
""" |
|
Identical to the XLM create_sinusoidal_embeddings except features are not interleaved. The cos features are in |
|
the 2nd half of the vector. [dim // 2:] |
|
""" |
|
n_pos, dim = out.shape |
|
position_enc = np.array( |
|
[[pos / np.power(10000, 2 * (j // 2) / dim) for j in range(dim)] for pos in range(n_pos)] |
|
) |
|
out.requires_grad = False |
|
sentinel = dim // 2 if dim % 2 == 0 else (dim // 2) + 1 |
|
out[:, 0:sentinel] = torch.FloatTensor(np.sin(position_enc[:, 0::2])) |
|
out[:, sentinel:] = torch.FloatTensor(np.cos(position_enc[:, 1::2])) |
|
out.detach_() |
|
return out |
|
|
|
@torch.no_grad() |
|
def forward(self, input_ids_shape: torch.Size, past_key_values_length: int = 0) -> torch.Tensor: |
|
"""`input_ids_shape` is expected to be [bsz x seqlen].""" |
|
bsz, seq_len = input_ids_shape[:2] |
|
positions = torch.arange( |
|
past_key_values_length, past_key_values_length + seq_len, dtype=torch.long, device=self.weight.device |
|
) |
|
return super().forward(positions) |
|
|
|
|
|
def load_tf_weights_in_roformer(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(f"Converting TensorFlow checkpoint from {tf_path}") |
|
|
|
init_vars = tf.train.list_variables(tf_path) |
|
names = [] |
|
arrays = [] |
|
for name, shape in init_vars: |
|
logger.info(f"Loading TF weight {name} with shape {shape}") |
|
array = tf.train.load_variable(tf_path, name) |
|
names.append(name.replace("bert", "roformer")) |
|
arrays.append(array) |
|
|
|
for name, array in zip(names, arrays): |
|
name = name.split("/") |
|
|
|
|
|
if any( |
|
n in ["adam_v", "adam_m", "AdamWeightDecayOptimizer", "AdamWeightDecayOptimizer_1", "global_step"] |
|
for n in name |
|
): |
|
logger.info(f"Skipping {'/'.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(f"Skipping {'/'.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: |
|
if not pointer.shape == array.shape: |
|
raise ValueError(f"Pointer shape {pointer.shape} and array shape {array.shape} mismatched") |
|
except AssertionError as e: |
|
e.args += (pointer.shape, array.shape) |
|
raise |
|
logger.info(f"Initialize PyTorch weight {name}") |
|
pointer.data = torch.from_numpy(array) |
|
return model |
|
|
|
|
|
class RoFormerEmbeddings(nn.Module): |
|
"""Construct the embeddings from word and token_type embeddings.""" |
|
|
|
def __init__(self, config): |
|
super().__init__() |
|
self.word_embeddings = nn.Embedding(config.vocab_size, config.embedding_size, padding_idx=config.pad_token_id) |
|
self.token_type_embeddings = nn.Embedding(config.type_vocab_size, config.embedding_size) |
|
|
|
|
|
|
|
self.LayerNorm = nn.LayerNorm(config.embedding_size, eps=config.layer_norm_eps) |
|
self.dropout = nn.Dropout(config.hidden_dropout_prob) |
|
|
|
def forward(self, input_ids=None, token_type_ids=None, inputs_embeds=None): |
|
if input_ids is not None: |
|
input_shape = input_ids.size() |
|
else: |
|
input_shape = inputs_embeds.size()[:-1] |
|
|
|
if inputs_embeds is None: |
|
inputs_embeds = self.word_embeddings(input_ids) |
|
|
|
if token_type_ids is None: |
|
token_type_ids = torch.zeros(input_shape, dtype=torch.long, device=inputs_embeds.device) |
|
|
|
token_type_embeddings = self.token_type_embeddings(token_type_ids) |
|
|
|
embeddings = inputs_embeds + token_type_embeddings |
|
|
|
embeddings = self.LayerNorm(embeddings) |
|
embeddings = self.dropout(embeddings) |
|
return embeddings |
|
|
|
|
|
class RoFormerSelfAttention(nn.Module): |
|
def __init__(self, config): |
|
super().__init__() |
|
if config.hidden_size % config.num_attention_heads != 0 and not hasattr(config, "embedding_size"): |
|
raise ValueError( |
|
f"The hidden size ({config.hidden_size}) is not a multiple of the number of attention " |
|
f"heads ({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) |
|
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.is_decoder = config.is_decoder |
|
self.rotary_value = config.rotary_value |
|
|
|
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, |
|
sinusoidal_pos=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) |
|
query_layer = self.transpose_for_scores(mixed_query_layer) |
|
|
|
|
|
|
|
is_cross_attention = encoder_hidden_states is not None |
|
|
|
if is_cross_attention and past_key_value is not None: |
|
|
|
key_layer = past_key_value[0] |
|
value_layer = past_key_value[1] |
|
attention_mask = encoder_attention_mask |
|
elif 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)) |
|
if sinusoidal_pos is not None: |
|
if self.rotary_value: |
|
query_layer, key_layer, value_layer = self.apply_rotary_position_embeddings( |
|
sinusoidal_pos, query_layer, key_layer, value_layer |
|
) |
|
else: |
|
query_layer, key_layer = self.apply_rotary_position_embeddings( |
|
sinusoidal_pos, query_layer, key_layer |
|
) |
|
if self.is_decoder: |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
past_key_value = (key_layer, value_layer) |
|
|
|
|
|
attention_scores = torch.matmul(query_layer, key_layer.transpose(-1, -2)) |
|
|
|
attention_scores = attention_scores / math.sqrt(self.attention_head_size) |
|
if attention_mask is not None: |
|
|
|
attention_scores = attention_scores + attention_mask |
|
|
|
|
|
attention_probs = nn.functional.softmax(attention_scores, dim=-1) |
|
|
|
|
|
|
|
attention_probs = self.dropout(attention_probs) |
|
|
|
|
|
if head_mask is not None: |
|
attention_probs = attention_probs * head_mask |
|
|
|
context_layer = torch.matmul(attention_probs, 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,) |
|
|
|
if self.is_decoder: |
|
outputs = outputs + (past_key_value,) |
|
return outputs |
|
|
|
@staticmethod |
|
def apply_rotary_position_embeddings(sinusoidal_pos, query_layer, key_layer, value_layer=None): |
|
|
|
|
|
|
|
sin, cos = sinusoidal_pos.chunk(2, dim=-1) |
|
|
|
sin_pos = torch.stack([sin, sin], dim=-1).reshape_as(sinusoidal_pos) |
|
|
|
cos_pos = torch.stack([cos, cos], dim=-1).reshape_as(sinusoidal_pos) |
|
|
|
rotate_half_query_layer = torch.stack([-query_layer[..., 1::2], query_layer[..., ::2]], dim=-1).reshape_as( |
|
query_layer |
|
) |
|
query_layer = query_layer * cos_pos + rotate_half_query_layer * sin_pos |
|
|
|
rotate_half_key_layer = torch.stack([-key_layer[..., 1::2], key_layer[..., ::2]], dim=-1).reshape_as(key_layer) |
|
key_layer = key_layer * cos_pos + rotate_half_key_layer * sin_pos |
|
if value_layer is not None: |
|
|
|
rotate_half_value_layer = torch.stack([-value_layer[..., 1::2], value_layer[..., ::2]], dim=-1).reshape_as( |
|
value_layer |
|
) |
|
value_layer = value_layer * cos_pos + rotate_half_value_layer * sin_pos |
|
return query_layer, key_layer, value_layer |
|
return query_layer, key_layer |
|
|
|
|
|
|
|
class RoFormerSelfOutput(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: torch.Tensor, input_tensor: torch.Tensor) -> torch.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 RoFormerAttention(nn.Module): |
|
def __init__(self, config): |
|
super().__init__() |
|
self.self = RoFormerSelfAttention(config) |
|
self.output = RoFormerSelfOutput(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 |
|
) |
|
|
|
|
|
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) |
|
|
|
|
|
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, |
|
sinusoidal_pos=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, |
|
sinusoidal_pos, |
|
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:] |
|
return outputs |
|
|
|
|
|
|
|
class RoFormerIntermediate(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: torch.Tensor) -> torch.Tensor: |
|
hidden_states = self.dense(hidden_states) |
|
hidden_states = self.intermediate_act_fn(hidden_states) |
|
return hidden_states |
|
|
|
|
|
|
|
class RoFormerOutput(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: torch.Tensor, input_tensor: torch.Tensor) -> torch.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 RoFormerLayer(nn.Module): |
|
def __init__(self, config): |
|
super().__init__() |
|
self.chunk_size_feed_forward = config.chunk_size_feed_forward |
|
self.seq_len_dim = 1 |
|
self.attention = RoFormerAttention(config) |
|
self.is_decoder = config.is_decoder |
|
self.add_cross_attention = config.add_cross_attention |
|
if self.add_cross_attention: |
|
if not self.is_decoder: |
|
raise ValueError(f"{self} should be used as a decoder model if cross attention is added") |
|
self.crossattention = RoFormerAttention(config) |
|
self.intermediate = RoFormerIntermediate(config) |
|
self.output = RoFormerOutput(config) |
|
|
|
def forward( |
|
self, |
|
hidden_states, |
|
attention_mask=None, |
|
sinusoidal_pos=None, |
|
head_mask=None, |
|
encoder_hidden_states=None, |
|
encoder_attention_mask=None, |
|
past_key_value=None, |
|
output_attentions=False, |
|
): |
|
|
|
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, |
|
sinusoidal_pos, |
|
head_mask, |
|
output_attentions=output_attentions, |
|
past_key_value=self_attn_past_key_value, |
|
) |
|
attention_output = self_attention_outputs[0] |
|
|
|
|
|
if self.is_decoder: |
|
outputs = self_attention_outputs[1:-1] |
|
present_key_value = self_attention_outputs[-1] |
|
else: |
|
outputs = self_attention_outputs[1:] |
|
|
|
cross_attn_present_key_value = None |
|
if self.is_decoder and encoder_hidden_states is not None: |
|
if not hasattr(self, "crossattention"): |
|
raise ValueError( |
|
f"If `encoder_hidden_states` are passed, {self} has to be instantiated with cross-attention " |
|
"layers by setting `config.add_cross_attention=True`" |
|
) |
|
|
|
|
|
cross_attn_past_key_value = past_key_value[-2:] if past_key_value is not None else None |
|
cross_attention_outputs = self.crossattention( |
|
attention_output, |
|
attention_mask, |
|
sinusoidal_pos, |
|
head_mask, |
|
encoder_hidden_states, |
|
encoder_attention_mask, |
|
cross_attn_past_key_value, |
|
output_attentions, |
|
) |
|
attention_output = cross_attention_outputs[0] |
|
outputs = outputs + cross_attention_outputs[1:-1] |
|
|
|
|
|
cross_attn_present_key_value = cross_attention_outputs[-1] |
|
present_key_value = present_key_value + cross_attn_present_key_value |
|
|
|
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 |
|
|
|
|
|
if self.is_decoder: |
|
outputs = outputs + (present_key_value,) |
|
|
|
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 RoFormerEncoder(nn.Module): |
|
def __init__(self, config): |
|
super().__init__() |
|
self.config = config |
|
self.embed_positions = RoFormerSinusoidalPositionalEmbedding( |
|
config.max_position_embeddings, config.hidden_size // config.num_attention_heads |
|
) |
|
self.layer = nn.ModuleList([RoFormerLayer(config) for _ in range(config.num_hidden_layers)]) |
|
self.gradient_checkpointing = False |
|
|
|
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 |
|
|
|
|
|
sinusoidal_pos = self.embed_positions(hidden_states.shape[:-1])[None, None, :, :] |
|
|
|
next_decoder_cache = () if use_cache else None |
|
for i, layer_module in enumerate(self.layer): |
|
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 self.gradient_checkpointing and self.training: |
|
|
|
if use_cache: |
|
logger.warning( |
|
"`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`..." |
|
) |
|
use_cache = False |
|
|
|
def create_custom_forward(module): |
|
def custom_forward(*inputs): |
|
return 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, |
|
sinusoidal_pos, |
|
layer_head_mask, |
|
encoder_hidden_states, |
|
encoder_attention_mask, |
|
) |
|
else: |
|
layer_outputs = layer_module( |
|
hidden_states, |
|
attention_mask, |
|
sinusoidal_pos, |
|
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 self.config.add_cross_attention: |
|
all_cross_attentions = all_cross_attentions + (layer_outputs[2],) |
|
|
|
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 RoFormerPredictionHeadTransform(nn.Module): |
|
def __init__(self, config): |
|
super().__init__() |
|
self.dense = nn.Linear(config.hidden_size, config.embedding_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.embedding_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 RoFormerLMPredictionHead(nn.Module): |
|
def __init__(self, config): |
|
super().__init__() |
|
self.transform = RoFormerPredictionHeadTransform(config) |
|
|
|
|
|
|
|
self.decoder = nn.Linear(config.embedding_size, config.vocab_size, bias=False) |
|
|
|
self.bias = nn.Parameter(torch.zeros(config.vocab_size)) |
|
|
|
|
|
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 RoFormerOnlyMLMHead(nn.Module): |
|
def __init__(self, config): |
|
super().__init__() |
|
self.predictions = RoFormerLMPredictionHead(config) |
|
|
|
def forward(self, sequence_output: torch.Tensor) -> torch.Tensor: |
|
prediction_scores = self.predictions(sequence_output) |
|
return prediction_scores |
|
|
|
|
|
class RoFormerPreTrainedModel(PreTrainedModel): |
|
""" |
|
An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained |
|
models. |
|
""" |
|
|
|
config_class = RoFormerConfig |
|
load_tf_weights = load_tf_weights_in_roformer |
|
base_model_prefix = "roformer" |
|
supports_gradient_checkpointing = True |
|
_keys_to_ignore_on_load_missing = [] |
|
_keys_to_ignore_on_load_unexpected = [ |
|
r"roformer.embeddings_project.weight", |
|
r"roformer.embeddings_project.bias", |
|
] |
|
|
|
def _init_weights(self, module): |
|
"""Initialize the weights""" |
|
if isinstance(module, nn.Linear): |
|
|
|
|
|
module.weight.data.normal_(mean=0.0, std=self.config.initializer_range) |
|
if module.bias is not None: |
|
module.bias.data.zero_() |
|
elif isinstance(module, RoFormerSinusoidalPositionalEmbedding): |
|
pass |
|
elif isinstance(module, nn.Embedding): |
|
module.weight.data.normal_(mean=0.0, std=self.config.initializer_range) |
|
if module.padding_idx is not None: |
|
module.weight.data[module.padding_idx].zero_() |
|
elif isinstance(module, nn.LayerNorm): |
|
module.bias.data.zero_() |
|
module.weight.data.fill_(1.0) |
|
|
|
def _set_gradient_checkpointing(self, module, value=False): |
|
if isinstance(module, RoFormerEncoder): |
|
module.gradient_checkpointing = value |
|
|
|
|
|
ROFORMER_START_DOCSTRING = r""" |
|
This model is a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) sub-class. Use |
|
it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and |
|
behavior. |
|
|
|
Parameters: |
|
config ([`RoFormerConfig`]): 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 [`~PreTrainedModel.from_pretrained`] method to load the model weights. |
|
""" |
|
|
|
ROFORMER_INPUTS_DOCSTRING = r""" |
|
Args: |
|
input_ids (`torch.LongTensor` of shape `({0})`): |
|
Indices of input sequence tokens in the vocabulary. |
|
|
|
Indices can be obtained using [`RoFormerTokenizer`]. See [`PreTrainedTokenizer.encode`] and |
|
[`PreTrainedTokenizer.__call__`] for details. |
|
|
|
[What are input IDs?](../glossary#input-ids) |
|
attention_mask (`torch.FloatTensor` of shape `({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#attention-mask) |
|
token_type_ids (`torch.LongTensor` of shape `({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#token-type-ids) |
|
head_mask (`torch.FloatTensor` of shape `(num_heads,)` or `(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 (`torch.FloatTensor` of shape `({0}, hidden_size)`, *optional*): |
|
Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This |
|
is useful if you want more control over how to convert *input_ids* indices into associated vectors than the |
|
model's internal embedding lookup matrix. |
|
output_attentions (`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 (`bool`, *optional*): |
|
Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for |
|
more detail. |
|
return_dict (`bool`, *optional*): |
|
Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. |
|
""" |
|
|
|
|
|
@add_start_docstrings( |
|
"The bare RoFormer Model transformer outputting raw hidden-states without any specific head on top.", |
|
ROFORMER_START_DOCSTRING, |
|
) |
|
class RoFormerModel(RoFormerPreTrainedModel): |
|
""" |
|
|
|
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](https://arxiv.org/abs/1706.03762) by Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, |
|
Llion Jones, Aidan N. Gomez, Lukasz Kaiser and Illia Polosukhin. |
|
|
|
To behave as an decoder the model needs to be initialized with the `is_decoder` argument of the configuration set |
|
to `True`. To be used in a Seq2Seq model, the model needs to initialized with both `is_decoder` argument and |
|
`add_cross_attention` set to `True`; an `encoder_hidden_states` is then expected as an input to the forward pass. |
|
""" |
|
|
|
def __init__(self, config): |
|
super().__init__(config) |
|
self.config = config |
|
self.embeddings = RoFormerEmbeddings(config) |
|
|
|
if config.embedding_size != config.hidden_size: |
|
self.embeddings_project = nn.Linear(config.embedding_size, config.hidden_size) |
|
|
|
self.encoder = RoFormerEncoder(config) |
|
|
|
|
|
self.post_init() |
|
|
|
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(ROFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length")) |
|
@add_code_sample_docstrings( |
|
processor_class=_TOKENIZER_FOR_DOC, |
|
checkpoint=_CHECKPOINT_FOR_DOC, |
|
output_type=BaseModelOutputWithPastAndCrossAttentions, |
|
config_class=_CONFIG_FOR_DOC, |
|
) |
|
def forward( |
|
self, |
|
input_ids: Optional[torch.LongTensor] = None, |
|
attention_mask: Optional[torch.FloatTensor] = None, |
|
token_type_ids: Optional[torch.LongTensor] = None, |
|
head_mask: Optional[torch.FloatTensor] = None, |
|
inputs_embeds: Optional[torch.FloatTensor] = None, |
|
encoder_hidden_states: Optional[torch.FloatTensor] = None, |
|
encoder_attention_mask: Optional[torch.FloatTensor] = None, |
|
past_key_values: Optional[Tuple[Tuple[torch.FloatTensor]]] = None, |
|
use_cache: Optional[bool] = None, |
|
output_attentions: Optional[bool] = None, |
|
output_hidden_states: Optional[bool] = None, |
|
return_dict: Optional[bool] = None, |
|
) -> Union[BaseModelOutputWithPastAndCrossAttentions, Tuple[torch.Tensor]]: |
|
r""" |
|
encoder_hidden_states (`torch.FloatTensor` of shape `(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 (`torch.FloatTensor` of shape `(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 (`tuple(tuple(torch.FloatTensor))` of length `config.n_layers` with each tuple having 4 tensors of shape `(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 `past_key_values` are used, the user can optionally input only the last `decoder_input_ids` (those that |
|
don't have their past key value states given to this model) of shape `(batch_size, 1)` instead of all |
|
`decoder_input_ids` of shape `(batch_size, sequence_length)`. |
|
use_cache (`bool`, *optional*): |
|
If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see |
|
`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 self.config.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() |
|
elif inputs_embeds is not None: |
|
input_shape = inputs_embeds.size()[:-1] |
|
else: |
|
raise ValueError("You have to specify either input_ids or inputs_embeds") |
|
|
|
batch_size, seq_length = input_shape |
|
device = input_ids.device if input_ids is not None else inputs_embeds.device |
|
|
|
|
|
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) |
|
|
|
|
|
|
|
extended_attention_mask: torch.Tensor = self.get_extended_attention_mask(attention_mask, input_shape) |
|
|
|
|
|
|
|
if self.config.is_decoder and encoder_hidden_states is not None: |
|
encoder_batch_size, encoder_sequence_length, _ = encoder_hidden_states.size() |
|
encoder_hidden_shape = (encoder_batch_size, encoder_sequence_length) |
|
if 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 = None |
|
|
|
|
|
|
|
|
|
|
|
|
|
head_mask = self.get_head_mask(head_mask, self.config.num_hidden_layers) |
|
|
|
embedding_output = self.embeddings( |
|
input_ids=input_ids, token_type_ids=token_type_ids, inputs_embeds=inputs_embeds |
|
) |
|
if hasattr(self, "embeddings_project"): |
|
embedding_output = self.embeddings_project(embedding_output) |
|
|
|
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] |
|
|
|
if not return_dict: |
|
return (sequence_output,) + encoder_outputs[1:] |
|
|
|
return BaseModelOutputWithPastAndCrossAttentions( |
|
last_hidden_state=sequence_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("""RoFormer Model with a `language modeling` head on top.""", ROFORMER_START_DOCSTRING) |
|
class RoFormerForMaskedLM(RoFormerPreTrainedModel): |
|
_keys_to_ignore_on_load_missing = ["cls.predictions.decoder.bias", "cls.predictions.decoder.weight"] |
|
|
|
def __init__(self, config): |
|
super().__init__(config) |
|
|
|
if config.is_decoder: |
|
logger.warning( |
|
"If you want to use `RoFormerForMaskedLM` make sure `config.is_decoder=False` for " |
|
"bi-directional self-attention." |
|
) |
|
|
|
self.roformer = RoFormerModel(config) |
|
self.cls = RoFormerOnlyMLMHead(config) |
|
|
|
|
|
self.post_init() |
|
|
|
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(ROFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length")) |
|
@add_code_sample_docstrings( |
|
processor_class=_TOKENIZER_FOR_DOC, |
|
checkpoint=_CHECKPOINT_FOR_DOC, |
|
output_type=MaskedLMOutput, |
|
config_class=_CONFIG_FOR_DOC, |
|
) |
|
def forward( |
|
self, |
|
input_ids: Optional[torch.LongTensor] = None, |
|
attention_mask: Optional[torch.FloatTensor] = None, |
|
token_type_ids: Optional[torch.LongTensor] = None, |
|
head_mask: Optional[torch.FloatTensor] = None, |
|
inputs_embeds: Optional[torch.FloatTensor] = None, |
|
encoder_hidden_states: Optional[torch.FloatTensor] = None, |
|
encoder_attention_mask: Optional[torch.FloatTensor] = None, |
|
labels: Optional[torch.LongTensor] = None, |
|
output_attentions: Optional[bool] = None, |
|
output_hidden_states: Optional[bool] = None, |
|
return_dict: Optional[bool] = None, |
|
loss_weight: Optional[torch.FloatTensor] = None, |
|
) -> Union[MaskedLMOutput, Tuple[torch.Tensor]]: |
|
r""" |
|
labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): |
|
Labels for computing the masked language modeling loss. 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 in `[0, ..., config.vocab_size]`. |
|
""" |
|
return_dict = return_dict if return_dict is not None else self.config.use_return_dict |
|
|
|
outputs = self.roformer( |
|
input_ids, |
|
attention_mask=attention_mask, |
|
token_type_ids=token_type_ids, |
|
head_mask=head_mask, |
|
inputs_embeds=inputs_embeds, |
|
encoder_hidden_states=encoder_hidden_states, |
|
encoder_attention_mask=encoder_attention_mask, |
|
output_attentions=output_attentions, |
|
output_hidden_states=output_hidden_states, |
|
return_dict=return_dict, |
|
) |
|
|
|
sequence_output = outputs[0] |
|
prediction_scores = self.cls(sequence_output) |
|
|
|
masked_lm_loss = None |
|
if labels is not None: |
|
loss_fct = CrossEntropyLoss(reduction="none") |
|
labels = labels.view(-1) |
|
masked_lm_loss = loss_fct(prediction_scores.view(-1, self.config.vocab_size), labels) |
|
loss_weight = loss_weight.view(-1) |
|
loss_weight[labels==-100] = 0.0 |
|
masked_lm_loss = (masked_lm_loss * loss_weight / loss_weight.sum()).sum() |
|
|
|
if not return_dict: |
|
output = (prediction_scores,) + outputs[1:] |
|
return ((masked_lm_loss,) + output) if masked_lm_loss is not None else output |
|
|
|
return MaskedLMOutput( |
|
loss=masked_lm_loss, |
|
logits=prediction_scores, |
|
hidden_states=outputs.hidden_states, |
|
attentions=outputs.attentions, |
|
) |
|
|
|
def prepare_inputs_for_generation(self, input_ids, attention_mask=None, **model_kwargs): |
|
input_shape = input_ids.shape |
|
effective_batch_size = input_shape[0] |
|
|
|
|
|
assert self.config.pad_token_id is not None, "The PAD token should be defined for generation" |
|
attention_mask = torch.cat([attention_mask, attention_mask.new_zeros((attention_mask.shape[0], 1))], dim=-1) |
|
dummy_token = torch.full( |
|
(effective_batch_size, 1), self.config.pad_token_id, dtype=torch.long, device=input_ids.device |
|
) |
|
input_ids = torch.cat([input_ids, dummy_token], dim=1) |
|
|
|
return {"input_ids": input_ids, "attention_mask": attention_mask} |
|
|
|
|
|
@add_start_docstrings( |
|
"""RoFormer Model with a `language modeling` head on top for CLM fine-tuning.""", ROFORMER_START_DOCSTRING |
|
) |
|
class RoFormerForCausalLM(RoFormerPreTrainedModel): |
|
_keys_to_ignore_on_load_missing = ["cls.predictions.decoder.bias", "cls.predictions.decoder.weight"] |
|
|
|
def __init__(self, config): |
|
super().__init__(config) |
|
|
|
if not config.is_decoder: |
|
logger.warning("If you want to use `RoFormerForCausalLM` as a standalone, add `is_decoder=True.`") |
|
|
|
self.roformer = RoFormerModel(config) |
|
self.cls = RoFormerOnlyMLMHead(config) |
|
|
|
|
|
self.post_init() |
|
|
|
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(ROFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length")) |
|
@replace_return_docstrings(output_type=CausalLMOutputWithCrossAttentions, config_class=_CONFIG_FOR_DOC) |
|
def forward( |
|
self, |
|
input_ids: Optional[torch.LongTensor] = None, |
|
attention_mask: Optional[torch.FloatTensor] = None, |
|
token_type_ids: Optional[torch.LongTensor] = None, |
|
inputs_embeds: Optional[torch.FloatTensor] = None, |
|
encoder_hidden_states: Optional[torch.FloatTensor] = None, |
|
encoder_attention_mask: Optional[torch.FloatTensor] = None, |
|
head_mask: Optional[torch.FloatTensor] = None, |
|
cross_attn_head_mask: Optional[torch.Tensor] = None, |
|
past_key_values: Optional[Tuple[Tuple[torch.FloatTensor]]] = None, |
|
labels: Optional[torch.LongTensor] = None, |
|
use_cache: Optional[bool] = None, |
|
output_attentions: Optional[bool] = None, |
|
output_hidden_states: Optional[bool] = None, |
|
return_dict: Optional[bool] = None, |
|
) -> Union[CausalLMOutputWithCrossAttentions, Tuple[torch.Tensor]]: |
|
r""" |
|
encoder_hidden_states (`torch.FloatTensor` of shape `(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 (`torch.FloatTensor` of shape `(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 (`tuple(tuple(torch.FloatTensor))` of length `config.n_layers` with each tuple having 4 tensors of shape `(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 `past_key_values` are used, the user can optionally input only the last `decoder_input_ids` (those that |
|
don't have their past key value states given to this model) of shape `(batch_size, 1)` instead of all |
|
`decoder_input_ids` of shape `(batch_size, sequence_length)`. |
|
labels (`torch.LongTensor` of shape `(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]`. |
|
use_cache (`bool`, *optional*): |
|
If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see |
|
`past_key_values`). |
|
|
|
Returns: |
|
|
|
Example: |
|
|
|
```python |
|
>>> from transformers import RoFormerTokenizer, RoFormerForCausalLM, RoFormerConfig |
|
>>> import torch |
|
|
|
>>> tokenizer = RoFormerTokenizer.from_pretrained("junnyu/roformer_chinese_base") |
|
>>> config = RoFormerConfig.from_pretrained("junnyu/roformer_chinese_base") |
|
>>> config.is_decoder = True |
|
>>> model = RoFormerForCausalLM.from_pretrained("junnyu/roformer_chinese_base", config=config) |
|
|
|
>>> inputs = tokenizer("今天天气非常好。", 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 |
|
|
|
outputs = self.roformer( |
|
input_ids, |
|
attention_mask=attention_mask, |
|
token_type_ids=token_type_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, |
|
) |
|
|
|
sequence_output = outputs[0] |
|
prediction_scores = self.cls(sequence_output) |
|
|
|
lm_loss = None |
|
if labels is not None: |
|
|
|
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 not return_dict: |
|
output = (prediction_scores,) + outputs[1:] |
|
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 attention_mask is None: |
|
attention_mask = input_ids.new_ones(input_shape) |
|
|
|
|
|
if past is not None: |
|
input_ids = input_ids[:, -1:] |
|
|
|
return {"input_ids": input_ids, "attention_mask": attention_mask, "past_key_values": past} |
|
|
|
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[:2]) + layer_past[2:], |
|
) |
|
return reordered_past |
|
|
|
|
|
class RoFormerClassificationHead(nn.Module): |
|
"""Head for sentence-level classification tasks.""" |
|
|
|
def __init__(self, config): |
|
super().__init__() |
|
self.dense = nn.Linear(config.hidden_size, config.hidden_size) |
|
self.dropout = nn.Dropout(config.hidden_dropout_prob) |
|
self.out_proj = nn.Linear(config.hidden_size, config.num_labels) |
|
|
|
self.config = config |
|
|
|
def forward(self, features, **kwargs): |
|
x = features[:, 0, :] |
|
x = self.dropout(x) |
|
x = self.dense(x) |
|
x = ACT2FN[self.config.hidden_act](x) |
|
x = self.dropout(x) |
|
x = self.out_proj(x) |
|
return x |
|
|
|
|
|
@add_start_docstrings( |
|
""" |
|
RoFormer Model transformer with a sequence classification/regression head on top (a linear layer on top of the |
|
pooled output) e.g. for GLUE tasks. |
|
""", |
|
ROFORMER_START_DOCSTRING, |
|
) |
|
class RoFormerForSequenceClassification(RoFormerPreTrainedModel): |
|
def __init__(self, config): |
|
super().__init__(config) |
|
self.num_labels = config.num_labels |
|
self.roformer = RoFormerModel(config) |
|
self.classifier = RoFormerClassificationHead(config) |
|
|
|
|
|
self.post_init() |
|
|
|
@add_start_docstrings_to_model_forward(ROFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length")) |
|
@add_code_sample_docstrings( |
|
processor_class=_TOKENIZER_FOR_DOC, |
|
checkpoint=_CHECKPOINT_FOR_DOC, |
|
output_type=SequenceClassifierOutput, |
|
config_class=_CONFIG_FOR_DOC, |
|
) |
|
def forward( |
|
self, |
|
input_ids: Optional[torch.LongTensor] = None, |
|
attention_mask: Optional[torch.FloatTensor] = None, |
|
token_type_ids: Optional[torch.LongTensor] = None, |
|
head_mask: Optional[torch.FloatTensor] = None, |
|
inputs_embeds: Optional[torch.FloatTensor] = None, |
|
labels: Optional[torch.LongTensor] = None, |
|
output_attentions: Optional[bool] = None, |
|
output_hidden_states: Optional[bool] = None, |
|
return_dict: Optional[bool] = None, |
|
) -> Union[SequenceClassifierOutput, Tuple[torch.Tensor]]: |
|
r""" |
|
labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): |
|
Labels for computing the sequence classification/regression loss. Indices should be in `[0, ..., |
|
config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If |
|
`config.num_labels > 1` a classification loss is computed (Cross-Entropy). |
|
""" |
|
return_dict = return_dict if return_dict is not None else self.config.use_return_dict |
|
|
|
outputs = self.roformer( |
|
input_ids, |
|
attention_mask=attention_mask, |
|
token_type_ids=token_type_ids, |
|
head_mask=head_mask, |
|
inputs_embeds=inputs_embeds, |
|
output_attentions=output_attentions, |
|
output_hidden_states=output_hidden_states, |
|
return_dict=return_dict, |
|
) |
|
|
|
sequence_output = outputs[0] |
|
logits = self.classifier(sequence_output) |
|
|
|
loss = None |
|
if labels is not None: |
|
if self.config.problem_type is None: |
|
if self.num_labels == 1: |
|
self.config.problem_type = "regression" |
|
elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int): |
|
self.config.problem_type = "single_label_classification" |
|
else: |
|
self.config.problem_type = "multi_label_classification" |
|
|
|
if self.config.problem_type == "regression": |
|
loss_fct = MSELoss() |
|
if self.num_labels == 1: |
|
loss = loss_fct(logits.squeeze(), labels.squeeze()) |
|
else: |
|
loss = loss_fct(logits, labels) |
|
elif self.config.problem_type == "single_label_classification": |
|
loss_fct = CrossEntropyLoss() |
|
loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1)) |
|
elif self.config.problem_type == "multi_label_classification": |
|
loss_fct = BCEWithLogitsLoss() |
|
loss = loss_fct(logits, labels) |
|
if not return_dict: |
|
output = (logits,) + outputs[1:] |
|
return ((loss,) + output) if loss is not None else output |
|
|
|
return SequenceClassifierOutput( |
|
loss=loss, |
|
logits=logits, |
|
hidden_states=outputs.hidden_states, |
|
attentions=outputs.attentions, |
|
) |
|
|
|
|
|
@add_start_docstrings( |
|
""" |
|
RoFormer Model with a multiple choice classification head on top (a linear layer on top of the pooled output and a |
|
softmax) e.g. for RocStories/SWAG tasks. |
|
""", |
|
ROFORMER_START_DOCSTRING, |
|
) |
|
class RoFormerForMultipleChoice(RoFormerPreTrainedModel): |
|
def __init__(self, config): |
|
super().__init__(config) |
|
|
|
self.roformer = RoFormerModel(config) |
|
self.sequence_summary = SequenceSummary(config) |
|
self.classifier = nn.Linear(config.hidden_size, 1) |
|
|
|
|
|
self.post_init() |
|
|
|
@add_start_docstrings_to_model_forward( |
|
ROFORMER_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length") |
|
) |
|
@add_code_sample_docstrings( |
|
processor_class=_TOKENIZER_FOR_DOC, |
|
checkpoint=_CHECKPOINT_FOR_DOC, |
|
output_type=MultipleChoiceModelOutput, |
|
config_class=_CONFIG_FOR_DOC, |
|
) |
|
def forward( |
|
self, |
|
input_ids: Optional[torch.LongTensor] = None, |
|
attention_mask: Optional[torch.FloatTensor] = None, |
|
token_type_ids: Optional[torch.LongTensor] = None, |
|
head_mask: Optional[torch.FloatTensor] = None, |
|
inputs_embeds: Optional[torch.FloatTensor] = None, |
|
labels: Optional[torch.LongTensor] = None, |
|
output_attentions: Optional[bool] = None, |
|
output_hidden_states: Optional[bool] = None, |
|
return_dict: Optional[bool] = None, |
|
) -> Union[MultipleChoiceModelOutput, Tuple[torch.Tensor]]: |
|
r""" |
|
labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): |
|
Labels for computing the multiple choice classification loss. Indices should be in `[0, ..., |
|
num_choices-1]` where `num_choices` is the size of the second dimension of the input tensors. (See |
|
`input_ids` above) |
|
""" |
|
return_dict = return_dict if return_dict is not None else self.config.use_return_dict |
|
num_choices = input_ids.shape[1] if input_ids is not None else inputs_embeds.shape[1] |
|
|
|
input_ids = input_ids.view(-1, input_ids.size(-1)) if input_ids is not None else None |
|
attention_mask = attention_mask.view(-1, attention_mask.size(-1)) if attention_mask is not None else None |
|
token_type_ids = token_type_ids.view(-1, token_type_ids.size(-1)) if token_type_ids is not None else None |
|
|
|
inputs_embeds = ( |
|
inputs_embeds.view(-1, inputs_embeds.size(-2), inputs_embeds.size(-1)) |
|
if inputs_embeds is not None |
|
else None |
|
) |
|
|
|
outputs = self.roformer( |
|
input_ids, |
|
attention_mask=attention_mask, |
|
token_type_ids=token_type_ids, |
|
head_mask=head_mask, |
|
inputs_embeds=inputs_embeds, |
|
output_attentions=output_attentions, |
|
output_hidden_states=output_hidden_states, |
|
return_dict=return_dict, |
|
) |
|
|
|
sequence_output = outputs[0] |
|
|
|
pooled_output = self.sequence_summary(sequence_output) |
|
logits = self.classifier(pooled_output) |
|
reshaped_logits = logits.view(-1, num_choices) |
|
|
|
loss = None |
|
if labels is not None: |
|
loss_fct = CrossEntropyLoss() |
|
loss = loss_fct(reshaped_logits, labels) |
|
|
|
if not return_dict: |
|
output = (reshaped_logits,) + outputs[1:] |
|
return ((loss,) + output) if loss is not None else output |
|
|
|
return MultipleChoiceModelOutput( |
|
loss=loss, |
|
logits=reshaped_logits, |
|
hidden_states=outputs.hidden_states, |
|
attentions=outputs.attentions, |
|
) |
|
|
|
|
|
@add_start_docstrings( |
|
""" |
|
RoFormer Model with a token classification head on top (a linear layer on top of the hidden-states output) e.g. for |
|
Named-Entity-Recognition (NER) tasks. |
|
""", |
|
ROFORMER_START_DOCSTRING, |
|
) |
|
class RoFormerForTokenClassification(RoFormerPreTrainedModel): |
|
def __init__(self, config): |
|
super().__init__(config) |
|
self.num_labels = config.num_labels |
|
|
|
self.roformer = RoFormerModel(config) |
|
self.dropout = nn.Dropout(config.hidden_dropout_prob) |
|
self.classifier = nn.Linear(config.hidden_size, config.num_labels) |
|
|
|
|
|
self.post_init() |
|
|
|
@add_start_docstrings_to_model_forward(ROFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length")) |
|
@add_code_sample_docstrings( |
|
processor_class=_TOKENIZER_FOR_DOC, |
|
checkpoint=_CHECKPOINT_FOR_DOC, |
|
output_type=TokenClassifierOutput, |
|
config_class=_CONFIG_FOR_DOC, |
|
) |
|
def forward( |
|
self, |
|
input_ids: Optional[torch.LongTensor] = None, |
|
attention_mask: Optional[torch.FloatTensor] = None, |
|
token_type_ids: Optional[torch.LongTensor] = None, |
|
head_mask: Optional[torch.FloatTensor] = None, |
|
inputs_embeds: Optional[torch.FloatTensor] = None, |
|
labels: Optional[torch.LongTensor] = None, |
|
output_attentions: Optional[bool] = None, |
|
output_hidden_states: Optional[bool] = None, |
|
return_dict: Optional[bool] = None, |
|
) -> Union[TokenClassifierOutput, Tuple[torch.Tensor]]: |
|
r""" |
|
labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): |
|
Labels for computing the token classification loss. Indices should be in `[0, ..., config.num_labels - 1]`. |
|
""" |
|
return_dict = return_dict if return_dict is not None else self.config.use_return_dict |
|
|
|
outputs = self.roformer( |
|
input_ids, |
|
attention_mask=attention_mask, |
|
token_type_ids=token_type_ids, |
|
head_mask=head_mask, |
|
inputs_embeds=inputs_embeds, |
|
output_attentions=output_attentions, |
|
output_hidden_states=output_hidden_states, |
|
return_dict=return_dict, |
|
) |
|
|
|
sequence_output = outputs[0] |
|
|
|
sequence_output = self.dropout(sequence_output) |
|
logits = self.classifier(sequence_output) |
|
|
|
loss = None |
|
if labels is not None: |
|
loss_fct = CrossEntropyLoss() |
|
loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1)) |
|
|
|
if not return_dict: |
|
output = (logits,) + outputs[1:] |
|
return ((loss,) + output) if loss is not None else output |
|
|
|
return TokenClassifierOutput( |
|
loss=loss, |
|
logits=logits, |
|
hidden_states=outputs.hidden_states, |
|
attentions=outputs.attentions, |
|
) |
|
|
|
|
|
@add_start_docstrings( |
|
""" |
|
RoFormer Model with a span classification head on top for extractive question-answering tasks like SQuAD (a linear |
|
layers on top of the hidden-states output to compute `span start logits` and `span end logits`). |
|
""", |
|
ROFORMER_START_DOCSTRING, |
|
) |
|
class RoFormerForQuestionAnswering(RoFormerPreTrainedModel): |
|
def __init__(self, config): |
|
super().__init__(config) |
|
|
|
config.num_labels = 2 |
|
self.num_labels = config.num_labels |
|
|
|
self.roformer = RoFormerModel(config) |
|
self.qa_outputs = nn.Linear(config.hidden_size, config.num_labels) |
|
|
|
|
|
self.post_init() |
|
|
|
@add_start_docstrings_to_model_forward(ROFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length")) |
|
@add_code_sample_docstrings( |
|
processor_class=_TOKENIZER_FOR_DOC, |
|
checkpoint=_CHECKPOINT_FOR_DOC, |
|
output_type=QuestionAnsweringModelOutput, |
|
config_class=_CONFIG_FOR_DOC, |
|
) |
|
def forward( |
|
self, |
|
input_ids: Optional[torch.LongTensor] = None, |
|
attention_mask: Optional[torch.FloatTensor] = None, |
|
token_type_ids: Optional[torch.LongTensor] = None, |
|
head_mask: Optional[torch.FloatTensor] = None, |
|
inputs_embeds: Optional[torch.FloatTensor] = None, |
|
start_positions: Optional[torch.LongTensor] = None, |
|
end_positions: Optional[torch.LongTensor] = None, |
|
output_attentions: Optional[bool] = None, |
|
output_hidden_states: Optional[bool] = None, |
|
return_dict: Optional[bool] = None, |
|
) -> Union[QuestionAnsweringModelOutput, Tuple[torch.Tensor]]: |
|
r""" |
|
start_positions (`torch.LongTensor` of shape `(batch_size,)`, *optional*): |
|
Labels for position (index) of the start of the labelled span for computing the token classification loss. |
|
Positions are clamped to the length of the sequence (`sequence_length`). Position outside of the sequence |
|
are not taken into account for computing the loss. |
|
end_positions (`torch.LongTensor` of shape `(batch_size,)`, *optional*): |
|
Labels for position (index) of the end of the labelled span for computing the token classification loss. |
|
Positions are clamped to the length of the sequence (`sequence_length`). Position outside of the sequence |
|
are not taken into account for computing the loss. |
|
""" |
|
return_dict = return_dict if return_dict is not None else self.config.use_return_dict |
|
|
|
outputs = self.roformer( |
|
input_ids, |
|
attention_mask=attention_mask, |
|
token_type_ids=token_type_ids, |
|
head_mask=head_mask, |
|
inputs_embeds=inputs_embeds, |
|
output_attentions=output_attentions, |
|
output_hidden_states=output_hidden_states, |
|
return_dict=return_dict, |
|
) |
|
|
|
sequence_output = outputs[0] |
|
|
|
logits = self.qa_outputs(sequence_output) |
|
start_logits, end_logits = logits.split(1, dim=-1) |
|
start_logits = start_logits.squeeze(-1) |
|
end_logits = end_logits.squeeze(-1) |
|
|
|
total_loss = None |
|
if start_positions is not None and end_positions is not None: |
|
|
|
if len(start_positions.size()) > 1: |
|
start_positions = start_positions.squeeze(-1) |
|
if len(end_positions.size()) > 1: |
|
end_positions = end_positions.squeeze(-1) |
|
|
|
ignored_index = start_logits.size(1) |
|
start_positions = start_positions.clamp(0, ignored_index) |
|
end_positions = end_positions.clamp(0, ignored_index) |
|
|
|
loss_fct = CrossEntropyLoss(ignore_index=ignored_index) |
|
start_loss = loss_fct(start_logits, start_positions) |
|
end_loss = loss_fct(end_logits, end_positions) |
|
total_loss = (start_loss + end_loss) / 2 |
|
|
|
if not return_dict: |
|
output = (start_logits, end_logits) + outputs[1:] |
|
return ((total_loss,) + output) if total_loss is not None else output |
|
|
|
return QuestionAnsweringModelOutput( |
|
loss=total_loss, |
|
start_logits=start_logits, |
|
end_logits=end_logits, |
|
hidden_states=outputs.hidden_states, |
|
attentions=outputs.attentions, |
|
) |
|
|