Datasets:

Modalities:
Image
Text
Formats:
parquet
Languages:
English
ArXiv:
Libraries:
Datasets
Dask
License:
Llabres commited on
Commit
c56be29
·
1 Parent(s): 8c6662d

Updated Readme

Browse files
Files changed (1) hide show
  1. README.md +4 -5
README.md CHANGED
@@ -219,7 +219,7 @@ from datasets import load_dataset
219
 
220
  skill = "sequence_filling" # "sequence_filling", "char_coherence", "visual_closure", "text_closure", "caption_relevance"
221
  split = "val" # "val", "test"
222
- dataset = load_dataset("VLR-CVC/ComPAP", skill, split=split)
223
  ```
224
 
225
  <details>
@@ -405,7 +405,7 @@ from datasets import load_dataset
405
 
406
  skill = "sequence_filling" # "sequence_filling", "char_coherence", "visual_closure", "text_closure", "caption_relevance"
407
  split = "val" # "val", "test"
408
- dataset = load_dataset("VLR-CVC/ComPAP", skill, split=split)
409
 
410
  processor = SingleImagePickAPanel()
411
  dataset = dataset.map(
@@ -414,7 +414,7 @@ dataset = dataset.map(
414
  batch_size=32,
415
  remove_columns=['context', 'options']
416
  )
417
- dataset.save_to_disk(f"ComPAP_{skill}_{split}_single_images")
418
  ```
419
 
420
  </details>
@@ -441,7 +441,6 @@ Where `sample_id` is the id of the sample, `correct_panel_id` is the prediction
441
  <summary>Pseudocode for the evaluation on val set, adapt for your model:</summary>
442
 
443
  ```python
444
- split = "val"
445
  skills = {
446
  "sequence_filling": {
447
  "num_examples": 262
@@ -461,7 +460,7 @@ skills = {
461
  }
462
 
463
  for skill in skills:
464
- dataset = load_dataset("VLR-CVC/ComPAP", skill, split=split)
465
  correct = 0
466
  total = 0
467
  for example in dataset:
 
219
 
220
  skill = "sequence_filling" # "sequence_filling", "char_coherence", "visual_closure", "text_closure", "caption_relevance"
221
  split = "val" # "val", "test"
222
+ dataset = load_dataset("VLR-CVC/ComicsPAP", skill, split=split)
223
  ```
224
 
225
  <details>
 
405
 
406
  skill = "sequence_filling" # "sequence_filling", "char_coherence", "visual_closure", "text_closure", "caption_relevance"
407
  split = "val" # "val", "test"
408
+ dataset = load_dataset("VLR-CVC/ComicsPAP", skill, split=split)
409
 
410
  processor = SingleImagePickAPanel()
411
  dataset = dataset.map(
 
414
  batch_size=32,
415
  remove_columns=['context', 'options']
416
  )
417
+ dataset.save_to_disk(f"ComicsPAP_{skill}_{split}_single_images")
418
  ```
419
 
420
  </details>
 
441
  <summary>Pseudocode for the evaluation on val set, adapt for your model:</summary>
442
 
443
  ```python
 
444
  skills = {
445
  "sequence_filling": {
446
  "num_examples": 262
 
460
  }
461
 
462
  for skill in skills:
463
+ dataset = load_dataset("VLR-CVC/ComicsPAP", skill, split="val")
464
  correct = 0
465
  total = 0
466
  for example in dataset: