dhansmair/flamingo-mini

Issues

• Doubt about MaskedCrossAttention

  #18 opened a year ago by eileenforwhat
  3

• Doubts about training

  #17 opened 2 years ago by TheMrguiller
  11

• About the training script

  #16 opened 2 years ago by kenchan0226
  1

• Text Prompting the model using the cached image

  #14 opened 2 years ago by tutunarsl
  2

• few shot example

  #12 opened 2 years ago by shiv6891
  2

• text generation not working with transformers > v4.25.1

  #11 opened 2 years ago by fionathrill
  4

• Bugs of evaluation / caption generation

  #13 opened 2 years ago by evelinehong
  3

• Error on image_captioning.ipynb

  #9 opened 2 years ago by tomasmadeira
  2

• ImportError: cannot import name 'CLIPImageProcessor' from 'transformers' (/databricks/python/lib/python3.9/site-packages/transformers/__init__.py)

  #10 opened 2 years ago by fionathrill
  2

• text generation incompatible with transformers 4.22.0

  #3 opened 2 years ago by dhansmair
  0

• Prompting example

  #5 opened 2 years ago by samp830
  4

• Video Support

  #8 opened 2 years ago by dhansmair
  0

• FlamingoProcessor detects any "<" character as a media_location

  #7 opened 2 years ago by dhansmair
  0

• generate() only works with use_cache=True

  #6 opened 2 years ago by dhansmair
  0

• use weight tying when copying lm_head from the language model

  #4 opened 2 years ago by dhansmair
  0

• parameters_trainable() and state_dict_trainable() do not include the token embedding matrix

  #2 opened 2 years ago by dhansmair
  0

• Web scraped data pretraining

  #1 opened 2 years ago by edmondja
  4