The TrOCR model was proposed in TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei. TrOCR consists of an image Transformer encoder and an autoregressive text Transformer decoder to perform optical character recognition (OCR).
The abstract from the paper is the following:
Text recognition is a long-standing research problem for document digitalization. Existing approaches for text recognition are usually built based on CNN for image understanding and RNN for char-level text generation. In addition, another language model is usually needed to improve the overall accuracy as a post-processing step. In this paper, we propose an end-to-end text recognition approach with pre-trained image Transformer and text Transformer models, namely TrOCR, which leverages the Transformer architecture for both image understanding and wordpiece-level text generation. The TrOCR model is simple but effective, and can be pre-trained with large-scale synthetic data and fine-tuned with human-labeled datasets. Experiments show that the TrOCR model outperforms the current state-of-the-art models on both printed and handwritten text recognition tasks.
TrOCR architecture. Taken from the original paper. Please refer to the VisionEncoderDecoder class on how to use this model.
This model was contributed by nielsr. The original code can be found here.
A list of official Hugging Face and community (indicated by 🌎) resources to help you get started with TrOCR. If you’re interested in submitting a resource to be included here, please feel free to open a Pull Request and we’ll review it! The resource should ideally demonstrate something new instead of duplicating an existing resource.
⚡️ Inference
TrOCR’s VisionEncoderDecoder model accepts images as input and makes use of
generate() to autoregressively generate text given the input image.
The [ViTImageProcessor/DeiTImageProcessor] class is responsible for preprocessing the input image and
[RobertaTokenizer/XLMRobertaTokenizer] decodes the generated target tokens to the target string. The
TrOCRProcessor wraps [ViTImageProcessor/DeiTImageProcessor] and [RobertaTokenizer/XLMRobertaTokenizer]
into a single instance to both extract the input features and decode the predicted token ids.
>>> from transformers import TrOCRProcessor, VisionEncoderDecoderModel
>>> import requests
>>> from PIL import Image
>>> processor = TrOCRProcessor.from_pretrained("microsoft/trocr-base-handwritten")
>>> model = VisionEncoderDecoderModel.from_pretrained("microsoft/trocr-base-handwritten")
>>> # load image from the IAM dataset
>>> url = "https://fki.tic.heia-fr.ch/static/img/a01-122-02.jpg"
>>> image = Image.open(requests.get(url, stream=True).raw).convert("RGB")
>>> pixel_values = processor(image, return_tensors="pt").pixel_values
>>> generated_ids = model.generate(pixel_values)
>>> generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]See the model hub to look for TrOCR checkpoints.
( vocab_size = 50265 d_model = 1024 decoder_layers = 12 decoder_attention_heads = 16 decoder_ffn_dim = 4096 activation_function = 'gelu' max_position_embeddings = 512 dropout = 0.1 attention_dropout = 0.0 activation_dropout = 0.0 decoder_start_token_id = 2 init_std = 0.02 decoder_layerdrop = 0.0 use_cache = True scale_embedding = False use_learned_position_embeddings = True layernorm_embedding = True pad_token_id = 1 bos_token_id = 0 eos_token_id = 2 **kwargs )
Parameters
int, optional, defaults to 50265) —
Vocabulary size of the TrOCR model. Defines the number of different tokens that can be represented by the
inputs_ids passed when calling TrOCRForCausalLM. int, optional, defaults to 1024) —
Dimensionality of the layers and the pooler layer. int, optional, defaults to 12) —
Number of decoder layers. int, optional, defaults to 16) —
Number of attention heads for each attention layer in the Transformer decoder. int, optional, defaults to 4096) —
Dimensionality of the “intermediate” (often named feed-forward) layer in decoder. str or function, optional, defaults to "gelu") —
The non-linear activation function (function or string) in the pooler. If string, "gelu", "relu",
"silu" and "gelu_new" are supported. int, optional, defaults to 512) —
The maximum sequence length that this model might ever be used with. Typically set this to something large
just in case (e.g., 512 or 1024 or 2048). float, optional, defaults to 0.1) —
The dropout probability for all fully connected layers in the embeddings, and pooler. float, optional, defaults to 0.0) —
The dropout ratio for the attention probabilities. float, optional, defaults to 0.0) —
The dropout ratio for activations inside the fully connected layer. float, optional, defaults to 0.02) —
The standard deviation of the truncated_normal_initializer for initializing all weight matrices. float, optional, defaults to 0.0) —
The LayerDrop probability for the decoder. See the [LayerDrop paper](see https://arxiv.org/abs/1909.11556)
for more details. bool, optional, defaults to True) —
Whether or not the model should return the last key/values attentions (not used by all models). bool, optional, defaults to False) —
Whether or not to scale the word embeddings by sqrt(d_model). bool, optional, defaults to True) —
Whether or not to use learned position embeddings. If not, sinusoidal position embeddings will be used. bool, optional, defaults to True) —
Whether or not to use a layernorm after the word + position embeddings. This is the configuration class to store the configuration of a TrOCRForCausalLM. It is used to instantiate an TrOCR model according to the specified arguments, defining the model architecture. Instantiating a configuration with the defaults will yield a similar configuration to that of the TrOCR microsoft/trocr-base-handwritten architecture.
Configuration objects inherit from PretrainedConfig and can be used to control the model outputs. Read the documentation from PretrainedConfig for more information.
Example:
>>> from transformers import TrOCRConfig, TrOCRForCausalLM
>>> # Initializing a TrOCR-base style configuration
>>> configuration = TrOCRConfig()
>>> # Initializing a model (with random weights) from the TrOCR-base style configuration
>>> model = TrOCRForCausalLM(configuration)
>>> # Accessing the model configuration
>>> configuration = model.config( image_processor = None tokenizer = None **kwargs )
Parameters
ViTImageProcessor/DeiTImageProcessor], optional) —
An instance of [ViTImageProcessor/DeiTImageProcessor]. The image processor is a required input. RobertaTokenizer/XLMRobertaTokenizer], optional) —
An instance of [RobertaTokenizer/XLMRobertaTokenizer]. The tokenizer is a required input. Constructs a TrOCR processor which wraps a vision image processor and a TrOCR tokenizer into a single processor.
TrOCRProcessor offers all the functionalities of [ViTImageProcessor/DeiTImageProcessor] and
[RobertaTokenizer/XLMRobertaTokenizer]. See the call() and decode() for
more information.
When used in normal mode, this method forwards all its arguments to AutoImageProcessor’s
__call__() and returns its output. If used in the context
as_target_processor() this method forwards all its arguments to TrOCRTokenizer’s
~TrOCRTokenizer.__call__. Please refer to the doctsring of the above two methods for more information.
( pretrained_model_name_or_path: Union cache_dir: Union = None force_download: bool = False local_files_only: bool = False token: Union = None revision: str = 'main' **kwargs )
Parameters
str or os.PathLike) —
This can be either:
./my_model_directory/../my_model_directory/preprocessor_config.json.
**kwargs —
Additional keyword arguments passed along to both
from_pretrained() and
~tokenization_utils_base.PreTrainedTokenizer.from_pretrained.Instantiate a processor associated with a pretrained model.
This class method is simply calling the feature extractor
from_pretrained(), image processor
ImageProcessingMixin and the tokenizer
~tokenization_utils_base.PreTrainedTokenizer.from_pretrained methods. Please refer to the docstrings of the
methods above for more information.
( save_directory push_to_hub: bool = False **kwargs )
Parameters
str or os.PathLike) —
Directory where the feature extractor JSON file and the tokenizer files will be saved (directory will
be created if it does not exist). bool, optional, defaults to False) —
Whether or not to push your model to the Hugging Face model hub after saving it. You can specify the
repository you want to push to with repo_id (will default to the name of save_directory in your
namespace). Dict[str, Any], optional) —
Additional key word arguments passed along to the push_to_hub() method. Saves the attributes of this processor (feature extractor, tokenizer…) in the specified directory so that it can be reloaded using the from_pretrained() method.
This class method is simply calling save_pretrained() and save_pretrained(). Please refer to the docstrings of the methods above for more information.
This method forwards all its arguments to TrOCRTokenizer’s batch_decode(). Please refer to the docstring of this method for more information.
This method forwards all its arguments to TrOCRTokenizer’s decode(). Please refer to the docstring of this method for more information.
( config )
Parameters
The TrOCR Decoder with a language modeling head. Can be used as the decoder part of EncoderDecoderModel and VisionEncoderDecoder.
This model inherits from 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.
( input_ids: Optional = None attention_mask: Optional = None encoder_hidden_states: Optional = None encoder_attention_mask: Optional = None head_mask: Optional = None cross_attn_head_mask: Optional = None past_key_values: Optional = None inputs_embeds: Optional = None labels: Optional = None use_cache: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → transformers.modeling_outputs.CausalLMOutputWithCrossAttentions or tuple(torch.FloatTensor)
Parameters
torch.LongTensor of shape (batch_size, sequence_length)) —
Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you
provide it.
Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.
torch.Tensor of shape (batch_size, sequence_length), optional) —
Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
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. 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]: torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) —
Mask to nullify selected heads of the attention modules. Mask values selected in [0, 1]:
torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) —
Mask to nullify selected heads of the cross-attention modules. Mask values selected in [0, 1]:
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)) and 2 additional tensors of
shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head). The two additional
tensors are only required when the model is used as a decoder in a Sequence to Sequence model.
Contains pre-computed hidden-states (key and values in the self-attention blocks and in the
cross-attention blocks) that can be used (see past_key_values input) to speed up sequential 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).
torch.LongTensor of shape (batch_size, sequence_length), optional) —
Labels for computing the masked language modeling loss. Indices should either be in [0, ..., 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.vocab_size]. 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).
bool, optional) —
Whether or not to return the attentions tensors of all attention layers. See attentions under
returned tensors for more detail. bool, optional) —
Whether or not to return the hidden states of all layers. See hidden_states under returned tensors
for more detail. bool, optional) —
Whether or not to return a ModelOutput instead of a plain tuple. Returns
transformers.modeling_outputs.CausalLMOutputWithCrossAttentions or tuple(torch.FloatTensor)
A transformers.modeling_outputs.CausalLMOutputWithCrossAttentions or a tuple of
torch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various
elements depending on the configuration (TrOCRConfig) and inputs.
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.vocab_size)) — Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
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.
cross_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).
Cross attentions weights after the attention softmax, used to compute the weighted average in the cross-attention heads.
past_key_values (tuple(tuple(torch.FloatTensor)), optional, returned when use_cache=True is passed or when config.use_cache=True) — Tuple of torch.FloatTensor tuples of length config.n_layers, with each tuple containing the cached key,
value states of the self-attention and the cross-attention layers if model is used in encoder-decoder
setting. Only relevant if config.is_decoder = True.
Contains pre-computed hidden-states (key and values in the attention blocks) that can be used (see
past_key_values input) to speed up sequential decoding.
Example:
>>> from transformers import (
... TrOCRConfig,
... TrOCRProcessor,
... TrOCRForCausalLM,
... ViTConfig,
... ViTModel,
... VisionEncoderDecoderModel,
... )
>>> import requests
>>> from PIL import Image
>>> # TrOCR is a decoder model and should be used within a VisionEncoderDecoderModel
>>> # init vision2text model with random weights
>>> encoder = ViTModel(ViTConfig())
>>> decoder = TrOCRForCausalLM(TrOCRConfig())
>>> model = VisionEncoderDecoderModel(encoder=encoder, decoder=decoder)
>>> # If you want to start from the pretrained model, load the checkpoint with `VisionEncoderDecoderModel`
>>> processor = TrOCRProcessor.from_pretrained("microsoft/trocr-base-handwritten")
>>> model = VisionEncoderDecoderModel.from_pretrained("microsoft/trocr-base-handwritten")
>>> # load image from the IAM dataset
>>> url = "https://fki.tic.heia-fr.ch/static/img/a01-122-02.jpg"
>>> image = Image.open(requests.get(url, stream=True).raw).convert("RGB")
>>> pixel_values = processor(image, return_tensors="pt").pixel_values
>>> text = "industry, ' Mr. Brown commented icily. ' Let us have a"
>>> # training
>>> model.config.decoder_start_token_id = processor.tokenizer.cls_token_id
>>> model.config.pad_token_id = processor.tokenizer.pad_token_id
>>> model.config.vocab_size = model.config.decoder.vocab_size
>>> labels = processor.tokenizer(text, return_tensors="pt").input_ids
>>> outputs = model(pixel_values, labels=labels)
>>> loss = outputs.loss
>>> round(loss.item(), 2)
5.30
>>> # inference
>>> generated_ids = model.generate(pixel_values)
>>> generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
>>> generated_text
'industry, " Mr. Brown commented icily. " Let us have a'