mirror of
https://github.com/huggingface/transformers.git
synced 2025-07-02 12:20:05 +06:00
508a704055
* No more Tuple, List, Dict * make fixup * More style fixes * Docstring fixes with regex replacement * Trigger tests * Redo fixes after rebase * Fix copies * [test all] * update * [test all] * update * [test all] * make style after rebase * Patch the hf_argparser test * Patch the hf_argparser test * style fixes * style fixes * style fixes * Fix docstrings in Cohere test * [test all] --------- Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
614 lines
29 KiB
Python
614 lines
29 KiB
Python
# 🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
|
|
# This file was automatically generated from examples/modular-transformers/modular_new_task_model.py.
|
|
# Do NOT edit this file manually as any edits will be overwritten by the generation of
|
|
# the file from the modular. If any change should be done, please apply the change to the
|
|
# modular_new_task_model.py file directly. One of our CI enforces this.
|
|
# 🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
|
|
from dataclasses import dataclass
|
|
from typing import ClassVar, Optional, Union
|
|
|
|
import torch
|
|
from torch import nn
|
|
|
|
from ...cache_utils import Cache, HybridCache, StaticCache
|
|
from ...generation import GenerationMixin
|
|
from ...modeling_flash_attention_utils import FlashAttentionKwargs
|
|
from ...modeling_outputs import BaseModelOutputWithPast
|
|
from ...modeling_utils import PreTrainedModel
|
|
from ...processing_utils import Unpack
|
|
from ...utils import ModelOutput, auto_docstring, can_return_tuple, is_torchdynamo_compiling
|
|
from ..auto import AutoModel
|
|
from .configuration_new_task_model import NewTaskModelConfig
|
|
|
|
|
|
@dataclass
|
|
class NewTaskModelModelOutputWithPast(BaseModelOutputWithPast):
|
|
"""
|
|
Base class for NewTaskModel outputs, with hidden states and attentions.
|
|
|
|
Args:
|
|
last_hidden_state (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`):
|
|
Sequence of hidden-states at the output of the last layer of the model.
|
|
past_key_values (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`):
|
|
Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of shape
|
|
`(batch_size, num_heads, sequence_length, embed_size_per_head)`)
|
|
|
|
Contains pre-computed hidden-states (key and values in the self-attention blocks) that can be used (see
|
|
`past_key_values` input) to speed up sequential decoding.
|
|
hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
|
|
Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
|
|
one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.
|
|
|
|
Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
|
|
attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
|
|
Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
|
|
sequence_length)`.
|
|
|
|
Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
|
|
heads.
|
|
image_hidden_states (`torch.FloatTensor`, *optional*):
|
|
A `torch.FloatTensor` of size `(batch_size, num_images, sequence_length, hidden_size)`.
|
|
image_hidden_states of the model produced by the vision encoder and after projecting the last hidden state.
|
|
"""
|
|
|
|
image_hidden_states: Optional[torch.FloatTensor] = None
|
|
|
|
|
|
@dataclass
|
|
class NewTaskModelCausalLMOutputWithPast(ModelOutput):
|
|
"""
|
|
Base class for NewTaskModel causal language model (or autoregressive) outputs.
|
|
|
|
Args:
|
|
loss (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` is provided):
|
|
Language modeling loss (for next-token prediction).
|
|
logits (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.text_config.vocab_size)`):
|
|
Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
|
|
past_key_values (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`):
|
|
Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of shape
|
|
`(batch_size, num_heads, sequence_length, embed_size_per_head)`)
|
|
|
|
Contains pre-computed hidden-states (key and values in the self-attention blocks) that can be used (see
|
|
`past_key_values` input) to speed up sequential decoding.
|
|
hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
|
|
Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
|
|
one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.
|
|
|
|
Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
|
|
attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
|
|
Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
|
|
sequence_length)`.
|
|
|
|
Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
|
|
heads.
|
|
image_hidden_states (`torch.FloatTensor`, *optional*):
|
|
A `torch.FloatTensor` of size `(batch_size, num_images, sequence_length, hidden_size)`.
|
|
image_hidden_states of the model produced by the vision encoder after projecting last hidden state.
|
|
"""
|
|
|
|
loss: Optional[torch.FloatTensor] = None
|
|
logits: Optional[torch.FloatTensor] = None
|
|
past_key_values: Optional[Union[list[torch.FloatTensor], Cache]] = None
|
|
hidden_states: Optional[tuple[torch.FloatTensor]] = None
|
|
attentions: Optional[tuple[torch.FloatTensor]] = None
|
|
image_hidden_states: Optional[torch.FloatTensor] = None
|
|
|
|
|
|
class NewTaskModelMultiModalProjector(nn.Module):
|
|
def __init__(self, config: NewTaskModelConfig):
|
|
super().__init__()
|
|
self.linear = nn.Linear(config.vision_config.hidden_size, config.vision_config.projection_dim, bias=True)
|
|
|
|
def forward(self, image_features):
|
|
hidden_states = self.linear(image_features)
|
|
|
|
return hidden_states
|
|
|
|
|
|
@auto_docstring
|
|
class NewTaskModelPreTrainedModel(PreTrainedModel):
|
|
config_class = NewTaskModelConfig
|
|
base_model_prefix = ""
|
|
supports_gradient_checkpointing = True
|
|
_no_split_modules = ["NewTaskModelMultiModalProjector"]
|
|
_skip_keys_device_placement = "past_key_values"
|
|
_supports_cache_class = True
|
|
_supports_quantized_cache = True
|
|
_supports_static_cache = True
|
|
_supports_flash_attn_2 = True
|
|
_supports_sdpa = True
|
|
_supports_flex_attn = True
|
|
_supports_attention_backend = True
|
|
|
|
def _init_weights(self, module):
|
|
# important: this ported version of NewTaskModelisn't meant for training from scratch - only
|
|
# inference and fine-tuning
|
|
std = getattr(self.config, "initializer_range", self.config.get_text_config().initializer_range)
|
|
|
|
if isinstance(module, nn.Linear):
|
|
module.weight.data.normal_(mean=0.0, std=std)
|
|
if module.bias is not None:
|
|
module.bias.data.zero_()
|
|
|
|
|
|
@auto_docstring(
|
|
custom_intro="""
|
|
The Base NewTaskModel model which consists of a vision backbone and a language model withou language modeling head.,
|
|
"""
|
|
)
|
|
class NewTaskModelModel(NewTaskModelPreTrainedModel):
|
|
_checkpoint_conversion_mapping = {"language_model.model": "language_model"}
|
|
|
|
def __init__(self, config: NewTaskModelConfig):
|
|
super().__init__(config)
|
|
self.vision_tower = AutoModel.from_config(config=config.vision_config)
|
|
self.multi_modal_projector = NewTaskModelMultiModalProjector(config)
|
|
self.vocab_size = config.text_config.vocab_size
|
|
|
|
language_model = AutoModel.from_config(config=config.text_config)
|
|
self.language_model = language_model
|
|
|
|
self.pad_token_id = self.config.pad_token_id if self.config.pad_token_id is not None else -1
|
|
self.post_init()
|
|
|
|
def get_input_embeddings(self):
|
|
return self.language_model.get_input_embeddings()
|
|
|
|
def set_input_embeddings(self, value):
|
|
self.language_model.set_input_embeddings(value)
|
|
|
|
def _update_causal_mask(
|
|
self,
|
|
attention_mask,
|
|
token_type_ids=None,
|
|
past_key_values=None,
|
|
cache_position=None,
|
|
input_tensor=None,
|
|
is_training: Optional[bool] = None,
|
|
):
|
|
if self.config.text_config._attn_implementation == "flash_attention_2":
|
|
if attention_mask is not None and 0.0 in attention_mask:
|
|
return attention_mask
|
|
return None
|
|
is_training = is_training if is_training is not None else self.training
|
|
using_static_cache = isinstance(past_key_values, StaticCache)
|
|
min_dtype = torch.finfo(self.dtype).min
|
|
if input_tensor is None:
|
|
input_tensor = attention_mask
|
|
|
|
inputs_lead_dim, sequence_length = input_tensor.shape[:2]
|
|
if using_static_cache:
|
|
target_length = past_key_values.get_max_cache_shape()
|
|
elif isinstance(past_key_values, HybridCache):
|
|
target_length = past_key_values.get_max_cache_shape()
|
|
else:
|
|
target_length = (
|
|
attention_mask.shape[-1]
|
|
if isinstance(attention_mask, torch.Tensor)
|
|
else cache_position[0] + sequence_length + 1
|
|
)
|
|
|
|
if attention_mask is not None and attention_mask.dim() == 4:
|
|
# In this case we assume that the mask comes already in inverted form and requires no inversion or slicing.
|
|
return attention_mask
|
|
|
|
causal_mask = torch.full(
|
|
(sequence_length, target_length), fill_value=min_dtype, dtype=self.dtype, device=cache_position.device
|
|
)
|
|
# Causal diagonal mask only if training, otherwise attend to the whole prefix. Training-specific attn for prefix is handled below
|
|
if sequence_length != 1:
|
|
if is_training:
|
|
causal_mask = torch.triu(causal_mask, diagonal=1)
|
|
else:
|
|
causal_mask[:, :sequence_length] = 0.0
|
|
|
|
causal_mask *= torch.arange(target_length, device=cache_position.device) > cache_position.reshape(-1, 1)
|
|
causal_mask = causal_mask[None, None, :, :].expand(inputs_lead_dim, 1, -1, -1)
|
|
if attention_mask is not None:
|
|
causal_mask = causal_mask.clone() # copy to contiguous memory for in-place edit
|
|
mask_length = attention_mask.shape[-1]
|
|
|
|
# First unmask prefix tokens during training
|
|
if is_training:
|
|
if token_type_ids is None:
|
|
raise ValueError("Token type ids must be provided during training")
|
|
causal_mask[:, :, :, :mask_length] = causal_mask[:, :, :, :mask_length].masked_fill(
|
|
token_type_ids[:, None, None, :].to(causal_mask.device) == 0, 0
|
|
)
|
|
|
|
# Then apply padding mask (will mask pad tokens)
|
|
padding_mask = causal_mask[:, :, :, :mask_length] + attention_mask[:, None, None, :].to(causal_mask.device)
|
|
padding_mask = padding_mask == 0
|
|
causal_mask[:, :, :, :mask_length] = causal_mask[:, :, :, :mask_length].masked_fill(
|
|
padding_mask, min_dtype
|
|
)
|
|
|
|
return causal_mask
|
|
|
|
def get_image_features(self, pixel_values: torch.FloatTensor):
|
|
"""
|
|
Obtains image last hidden states from the vision tower and apply multimodal projection.
|
|
|
|
Args:
|
|
pixel_values (`torch.FloatTensor]` of shape `(batch_size, channels, height, width)`)
|
|
The tensors corresponding to the input images.
|
|
Returns:
|
|
image_features (`torch.Tensor`): Image feature tensor of shape `(num_images, image_length, embed_dim)`).
|
|
"""
|
|
image_outputs = self.vision_tower(pixel_values)
|
|
selected_image_feature = image_outputs.last_hidden_state
|
|
image_features = self.multi_modal_projector(selected_image_feature)
|
|
image_features = image_features / (self.config.text_config.hidden_size**0.5)
|
|
return image_features
|
|
|
|
@can_return_tuple
|
|
@auto_docstring
|
|
def forward(
|
|
self,
|
|
input_ids: torch.LongTensor = None,
|
|
pixel_values: torch.FloatTensor = None,
|
|
attention_mask: Optional[torch.Tensor] = None,
|
|
position_ids: Optional[torch.LongTensor] = None,
|
|
past_key_values: Optional[Union[list[torch.FloatTensor], Cache]] = None,
|
|
token_type_ids: Optional[torch.LongTensor] = None,
|
|
cache_position: Optional[torch.LongTensor] = None,
|
|
inputs_embeds: Optional[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,
|
|
**kwargs: Unpack[FlashAttentionKwargs],
|
|
) -> Union[tuple, NewTaskModelModelOutputWithPast]:
|
|
r"""
|
|
labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
|
|
Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
|
|
config.text_config.vocab_size]` or -100 (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.text_config.vocab_size]`.
|
|
|
|
Example:
|
|
|
|
```python
|
|
>>> from PIL import Image
|
|
>>> import requests
|
|
>>> from transformers import AutoProcessor, NewTaskModelForConditionalGeneration
|
|
|
|
>>> model = NewTaskModelForConditionalGeneration.from_pretrained("google/new_task_model2-3b-mix-224")
|
|
>>> processor = AutoProcessor.from_pretrained("google/new_task_model2-3b-mix-224")
|
|
|
|
>>> prompt = "Where is the cat standing?"
|
|
>>> url = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/pipeline-cat-chonk.jpeg"
|
|
>>> image = Image.open(requests.get(url, stream=True).raw)
|
|
|
|
>>> inputs = processor(images=image, text=prompt, return_tensors="pt")
|
|
|
|
>>> # Generate
|
|
>>> generate_ids = model.generate(**inputs,)
|
|
>>> processor.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
|
|
"Where is the cat standing?\nsnow"
|
|
```"""
|
|
|
|
if (input_ids is None) ^ (inputs_embeds is not None):
|
|
raise ValueError("You must specify exactly one of input_ids or inputs_embeds")
|
|
|
|
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
|
|
|
|
is_training = token_type_ids is not None and labels is not None
|
|
|
|
# Replace image id woth PAD if the image token if OOV, to avoid index-errors
|
|
if input_ids is not None and self.config.image_token_id >= self.vocab_size:
|
|
special_image_mask = input_ids == self.config.image_token_id
|
|
llm_input_ids = input_ids.clone()
|
|
llm_input_ids[special_image_mask] = 0
|
|
else:
|
|
llm_input_ids = input_ids
|
|
|
|
if inputs_embeds is None:
|
|
inputs_embeds = self.get_input_embeddings()(llm_input_ids)
|
|
|
|
if cache_position is None:
|
|
past_seen_tokens = past_key_values.get_seq_length() if past_key_values is not None else 0
|
|
cache_position = torch.arange(
|
|
past_seen_tokens, past_seen_tokens + inputs_embeds.shape[1], device=inputs_embeds.device
|
|
)
|
|
|
|
if position_ids is None:
|
|
position_ids = cache_position.unsqueeze(0) + 1 # NewTaskModel positions are 1-indexed
|
|
|
|
# Merge text and images
|
|
if pixel_values is not None:
|
|
image_features = self.get_image_features(pixel_values)
|
|
|
|
if input_ids is None:
|
|
special_image_mask = inputs_embeds == self.get_input_embeddings()(
|
|
torch.tensor(self.config.image_token_id, dtype=torch.long, device=inputs_embeds.device)
|
|
)
|
|
else:
|
|
special_image_mask = (input_ids == self.config.image_token_id).unsqueeze(-1)
|
|
special_image_mask = special_image_mask.expand_as(inputs_embeds).to(inputs_embeds.device)
|
|
|
|
if not is_torchdynamo_compiling() and inputs_embeds[special_image_mask].numel() != image_features.numel():
|
|
image_tokens_in_text = (special_image_mask).sum(dim=1).sum(dim=0)[0]
|
|
raise ValueError(
|
|
f"Number of images does not match number of special image tokens in the input text. "
|
|
f"Got {image_tokens_in_text} image tokens in the text but {image_features.shape[0] * image_features.shape[1]} "
|
|
"tokens from image embeddings."
|
|
)
|
|
image_features = image_features.to(inputs_embeds.device, inputs_embeds.dtype)
|
|
inputs_embeds = inputs_embeds.masked_scatter(special_image_mask, image_features)
|
|
|
|
causal_mask = self._update_causal_mask(
|
|
attention_mask, token_type_ids, past_key_values, cache_position, inputs_embeds, is_training
|
|
)
|
|
outputs = self.language_model(
|
|
attention_mask=causal_mask,
|
|
position_ids=position_ids,
|
|
past_key_values=past_key_values,
|
|
inputs_embeds=inputs_embeds,
|
|
use_cache=use_cache,
|
|
output_attentions=output_attentions,
|
|
output_hidden_states=output_hidden_states,
|
|
return_dict=True,
|
|
cache_position=cache_position,
|
|
**kwargs,
|
|
)
|
|
|
|
return NewTaskModelModelOutputWithPast(
|
|
last_hidden_state=outputs.last_hidden_state,
|
|
past_key_values=outputs.past_key_values,
|
|
hidden_states=outputs.hidden_states,
|
|
attentions=outputs.attentions,
|
|
image_hidden_states=image_features if pixel_values is not None else None,
|
|
)
|
|
|
|
|
|
@auto_docstring(
|
|
custom_intro="""
|
|
The Base NewTaskModel model which consists of a vision backbone and a language model without language modeling head.,
|
|
"""
|
|
)
|
|
class NewTaskModelForNewTask(NewTaskModelPreTrainedModel, GenerationMixin):
|
|
_checkpoint_conversion_mapping = {
|
|
"^language_model.model": "model.language_model",
|
|
"^vision_tower": "model.vision_tower",
|
|
"^multi_modal_projector": "model.multi_modal_projector",
|
|
"^language_model.lm_head": "lm_head",
|
|
}
|
|
_tied_weights_keys = ["lm_head.weight"]
|
|
main_input_name: ClassVar[str] = "doc_input_ids" # transformers-related
|
|
|
|
def __init__(self, config):
|
|
super().__init__(config)
|
|
self.model = NewTaskModelModel(config)
|
|
self.lm_head = nn.Linear(config.text_config.hidden_size, config.text_config.vocab_size, bias=False)
|
|
|
|
self.embedding_dim = self.config.embedding_dim
|
|
self.custom_text_proj = nn.Linear(self.config.text_config.hidden_size, self.embedding_dim)
|
|
|
|
if self.language_model._tied_weights_keys is not None:
|
|
self._tied_weights_keys = [f"model.language_model.{k}" for k in self.language_model._tied_weights_keys]
|
|
self.post_init()
|
|
|
|
def get_input_embeddings(self):
|
|
return self.model.get_input_embeddings()
|
|
|
|
def set_input_embeddings(self, value):
|
|
self.model.set_input_embeddings(value)
|
|
|
|
def get_output_embeddings(self):
|
|
return self.lm_head
|
|
|
|
def set_output_embeddings(self, new_embeddings):
|
|
self.lm_head = new_embeddings
|
|
|
|
def set_decoder(self, decoder):
|
|
self.model = decoder
|
|
|
|
def get_decoder(self):
|
|
return self.model
|
|
|
|
# Make modules available throught conditional class for BC
|
|
@property
|
|
def language_model(self):
|
|
return self.model.language_model
|
|
|
|
@property
|
|
def vision_tower(self):
|
|
return self.model.vision_tower
|
|
|
|
@property
|
|
def multi_modal_projector(self):
|
|
return self.model.multi_modal_projector
|
|
|
|
@can_return_tuple
|
|
@auto_docstring
|
|
def forward(
|
|
self,
|
|
input_ids: torch.LongTensor = None,
|
|
pixel_values: torch.FloatTensor = None,
|
|
attention_mask: Optional[torch.Tensor] = None,
|
|
position_ids: Optional[torch.LongTensor] = None,
|
|
past_key_values: Optional[Union[list[torch.FloatTensor], Cache]] = None,
|
|
token_type_ids: Optional[torch.LongTensor] = None,
|
|
cache_position: Optional[torch.LongTensor] = None,
|
|
inputs_embeds: Optional[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,
|
|
num_logits_to_keep: int = 0,
|
|
) -> Union[tuple, NewTaskModelCausalLMOutputWithPast]:
|
|
r"""
|
|
labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
|
|
Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
|
|
config.text_config.vocab_size]` or -100 (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.text_config.vocab_size]`.
|
|
|
|
Example:
|
|
|
|
```python
|
|
>>> from PIL import Image
|
|
>>> import requests
|
|
>>> from transformers import AutoProcessor, NewTaskModelForNewTask
|
|
|
|
>>> model = NewTaskModelForNewTask.from_pretrained("google/new_task_model2-3b-mix-224")
|
|
>>> processor = AutoProcessor.from_pretrained("google/new_task_model2-3b-mix-224")
|
|
|
|
>>> prompt = "Where is the cat standing?"
|
|
>>> url = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/pipeline-cat-chonk.jpeg"
|
|
>>> image = Image.open(requests.get(url, stream=True).raw)
|
|
|
|
>>> inputs = processor(images=image, text=prompt, return_tensors="pt")
|
|
|
|
>>> # Generate
|
|
>>> generate_ids = model.generate(**inputs,)
|
|
>>> processor.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
|
|
"Where is the cat standing?\nsnow"
|
|
```
|
|
Returns:
|
|
"""
|
|
vlm_outputs = super().forward(
|
|
input_ids=input_ids,
|
|
pixel_values=pixel_values,
|
|
attention_mask=attention_mask,
|
|
position_ids=position_ids,
|
|
past_key_values=past_key_values,
|
|
token_type_ids=token_type_ids,
|
|
cache_position=cache_position,
|
|
inputs_embeds=inputs_embeds,
|
|
labels=labels,
|
|
use_cache=use_cache,
|
|
output_attentions=output_attentions,
|
|
output_hidden_states=True,
|
|
return_dict=True,
|
|
num_logits_to_keep=num_logits_to_keep,
|
|
)
|
|
last_hidden_states = vlm_outputs.hidden_states[-1] # (batch_size, sequence_length, hidden_size)
|
|
proj = self.custom_text_proj(last_hidden_states) # (batch_size, sequence_length, dim)
|
|
|
|
# L2 normalization
|
|
embeddings = proj / proj.norm(dim=-1, keepdim=True) # (batch_size, sequence_length, dim)
|
|
|
|
if attention_mask is not None:
|
|
embeddings = embeddings * attention_mask.unsqueeze(-1) # (batch_size, sequence_length, dim)
|
|
|
|
return (embeddings,) + vlm_outputs
|
|
|
|
def prepare_inputs_for_generation(
|
|
self,
|
|
input_ids,
|
|
past_key_values=None,
|
|
inputs_embeds=None,
|
|
cache_position=None,
|
|
position_ids=None,
|
|
pixel_values=None,
|
|
attention_mask=None,
|
|
token_type_ids=None,
|
|
use_cache=True,
|
|
logits_to_keep=None,
|
|
labels=None,
|
|
**kwargs,
|
|
):
|
|
# Overwritten -- custom `position_ids` and `pixel_values` handling
|
|
model_inputs = super().prepare_inputs_for_generation(
|
|
input_ids,
|
|
past_key_values=past_key_values,
|
|
inputs_embeds=inputs_embeds,
|
|
attention_mask=attention_mask,
|
|
position_ids=position_ids,
|
|
cache_position=cache_position,
|
|
use_cache=use_cache,
|
|
logits_to_keep=logits_to_keep,
|
|
token_type_ids=token_type_ids,
|
|
**kwargs,
|
|
)
|
|
|
|
# position_ids in NewTaskModel are 1-indexed
|
|
if model_inputs.get("position_ids") is not None:
|
|
model_inputs["position_ids"] += 1
|
|
# If we're in cached decoding stage, pixel values should be None because input ids do not contain special image token anymore
|
|
# Otherwise we need pixel values to be passed to model. NOTE: use_cache=False needs pixel_values always
|
|
if cache_position[0] == 0:
|
|
model_inputs["pixel_values"] = pixel_values
|
|
is_training = token_type_ids is not None and labels is not None
|
|
if cache_position[0] == 0 and isinstance(past_key_values, HybridCache):
|
|
input_tensor = inputs_embeds if inputs_embeds is not None else input_ids
|
|
causal_mask = self.model._update_causal_mask(
|
|
attention_mask, token_type_ids, past_key_values, cache_position, input_tensor, is_training
|
|
)
|
|
model_inputs["attention_mask"] = causal_mask
|
|
|
|
return model_inputs
|
|
|
|
@staticmethod
|
|
def _prepare_4d_causal_attention_mask_with_cache_position(
|
|
attention_mask: torch.Tensor,
|
|
sequence_length: int,
|
|
target_length: int,
|
|
dtype: torch.dtype,
|
|
cache_position: torch.Tensor,
|
|
batch_size: int,
|
|
**kwargs,
|
|
):
|
|
"""
|
|
Creates a causal 4D mask of shape `(batch_size, 1, query_length, key_value_length)` from a 2D mask of shape
|
|
`(batch_size, key_value_length)`, or if the input `attention_mask` is already 4D, do nothing.
|
|
|
|
Args:
|
|
attention_mask (`torch.Tensor`):
|
|
A 2D attention mask of shape `(batch_size, key_value_length)` or a 4D attention mask of shape
|
|
`(batch_size, 1, query_length, key_value_length)`.
|
|
sequence_length (`int`):
|
|
The sequence length being processed.
|
|
target_length (`int`):
|
|
The target length: when generating with static cache, the mask should be as long as the static cache,
|
|
to account for the 0 padding, the part of the cache that is not filled yet.
|
|
dtype (`torch.dtype`):
|
|
The dtype to use for the 4D attention mask.
|
|
cache_position (`torch.Tensor`):
|
|
Indices depicting the position of the input sequence tokens in the sequence.
|
|
batch_size (`torch.Tensor`):
|
|
Batch size.
|
|
"""
|
|
if attention_mask is not None and attention_mask.dim() == 4:
|
|
# In this case we assume that the mask comes already in inverted form and requires no inversion or slicing.
|
|
causal_mask = attention_mask
|
|
else:
|
|
min_dtype = torch.finfo(dtype).min
|
|
causal_mask = torch.full(
|
|
(sequence_length, target_length), fill_value=min_dtype, dtype=dtype, device=cache_position.device
|
|
)
|
|
if sequence_length != 1:
|
|
causal_mask = torch.triu(causal_mask, diagonal=1)
|
|
causal_mask *= torch.arange(target_length, device=cache_position.device) > cache_position.reshape(-1, 1)
|
|
causal_mask = causal_mask[None, None, :, :].expand(batch_size, 1, -1, -1)
|
|
if attention_mask is not None:
|
|
causal_mask = causal_mask.clone() # copy to contiguous memory for in-place edit
|
|
mask_length = attention_mask.shape[-1]
|
|
padding_mask = causal_mask[:, :, :, :mask_length] + attention_mask[:, None, None, :].to(
|
|
causal_mask.device
|
|
)
|
|
padding_mask = padding_mask == 0
|
|
causal_mask[:, :, :, :mask_length] = causal_mask[:, :, :, :mask_length].masked_fill(
|
|
padding_mask, min_dtype
|
|
)
|
|
|
|
return causal_mask
|
|
|
|
def resize_token_embeddings(
|
|
self, new_num_tokens: Optional[int] = None, pad_to_multiple_of=None, mean_resizing=True
|
|
) -> nn.Embedding:
|
|
model_embeds = self.language_model.resize_token_embeddings(new_num_tokens, pad_to_multiple_of, mean_resizing)
|
|
|
|
# Update vocab size
|
|
self.config.text_config.vocab_size = model_embeds.num_embeddings
|
|
self.config.vocab_size = model_embeds.num_embeddings
|
|
self.vocab_size = model_embeds.num_embeddings
|
|
|
|
return model_embeds
|