📄 Lmms Eval¶

The doc-vqa Dataset integrates images from the Infographic_vqa dataset sourced from HuggingFaceM4 The Cauldron dataset, as well as images from the dataset AFTDB (Arxiv Figure Table Database) curated by cmarkea. This dataset consists of pairs of images and corresponding text, with each image linked to an average of five questions and answers available in both English and French. These questions and answers were generated using Gemini 1.5 Pro, thereby rendering the dataset well-suited for multimodal tasks involving image-text pairing and multilingual question answering.

Tags: license:apache-2.0, multilinguality:monolingual, modalities:['image', 'text'], size_categories:10K<n<100K, task_categories:question-answering, task_ids:extractive-qa, category:dataset

cards.doc_vqa.lmms_eval

type: TaskCard
loader: 
  type: LoadHF
  path: lmms-lab/DocVQA
  name: DocVQA
  data_classification_policy: 
    - public
preprocess_steps: 
  - type: RenameSplits
    mapper: 
      validation: test
  - type: ToImage
    field: image
    to_field: context
  - type: Set
    fields: 
      context_type: image
task: tasks.qa.with_context.abstractive[metrics=[metrics.anls]]
templates: 
  - type: MultiReferenceTemplate
    input_format: "{context}\n{question}\nAnswer the question using a single word or phrase."
    references_field: answers
  - templates.qa.with_context
  - templates.qa.extractive
  - templates.qa.with_context.simple
  - templates.qa.with_context.simple2
  - templates.qa.with_context.with_type
  - templates.qa.with_context.question_first
  - templates.qa.with_context.ffqa
  - templates.qa.with_context.title
  - templates.qa.with_context.lmms_eval
default_template: 
  type: MultiReferenceTemplate
  input_format: "{context}\n{question}\nAnswer the question using a single word or phrase."
  references_field: answers
[source]

Explanation about TaskCard¶

TaskCard delineates the phases in transforming the source dataset into model input, and specifies the metrics for evaluation of model output.

Args:

loader:
specifies the source address and the loading operator that can access that source and transform it into a unitxt multistream.

preprocess_steps:
list of unitxt operators to process the data source into model input.

task:
specifies the fields (of the already (pre)processed instance) making the inputs, the fields making the outputs, and the metrics to be used for evaluating the model output.

templates:
format strings to be applied on the input fields (specified by the task) and the output fields. The template also carries the instructions and the list of postprocessing steps, to be applied to the model output.

default_template:
a default template for tasks with very specific task dataset specific template

Explanation about Set¶

Sets specified fields in each instance, in a given stream or all streams (default), with specified values. If fields exist, updates them, if do not exist – adds them.

Args:
fields (Dict[str, object]): The fields to add to each instance. Use ‘/’ to access inner fields

use_deepcopy (bool) : Deep copy the input value to avoid later modifications

Examples:
# Set a value of a list consisting of “positive” and “negative” do field “classes” to each and every instance of all streams Set(fields={"classes": ["positive","negatives"]})

# In each and every instance of all streams, field “span” is to become a dictionary containing a field “start”, in which the value 0 is to be set Set(fields={"span/start": 0}

# In all instances of stream “train” only, Set field “classes” to have the value of a list consisting of “positive” and “negative” Set(fields={"classes": ["positive","negatives"], apply_to_stream=["train"]})

# Set field “classes” to have the value of a given list, preventing modification of original list from changing the instance. Set(fields={"classes": alist}), use_deepcopy=True) if now alist is modified, still the instances remain intact.

Explanation about LoadHF¶

Loads datasets from the HuggingFace Hub.

It supports loading with or without streaming, and it can filter datasets upon loading.
Args:

path:
The path or identifier of the dataset on the HuggingFace Hub.

name:
An optional dataset name.

data_dir:
Optional directory to store downloaded data.

split:
Optional specification of which split to load.

data_files:
Optional specification of particular data files to load.

revision:
Optional. The revision of the dataset. Often the commit id. Use in case you want to set the dataset version.

streaming (bool):
indicating if streaming should be used.

filtering_lambda (str, optional):
A lambda function for filtering the data after loading.

num_proc (int, optional):
Specifies the number of processes to use for parallel dataset loading.

Example:
Loading glue’s mrpc dataset
load_hf = LoadHF(path='glue', name='mrpc')

References: templates.qa.with_context.question_first, templates.qa.with_context.lmms_eval, templates.qa.with_context.with_type, tasks.qa.with_context.abstractive, templates.qa.with_context.simple2, templates.qa.with_context.simple, templates.qa.with_context.title, templates.qa.with_context.ffqa, templates.qa.with_context, templates.qa.extractive, metrics.anls