jalammar/ecco

Issues

• Can ecco be applied to tabular pytorch models?

  #117 opened 5 months ago by eyal-trabelsi
  0

• Installing Ecco, unable to install in Colab

  #115 opened 7 months ago by SalvatoreRa
  3

• Decoder hidden states not found

  #113 opened a year ago by herbert-zhou
  3

• baseline for sentence classification task using IG/Deeplift?

  #114 opened 9 months ago by lkqnaruto
  0

• installing ecco

  #111 opened a year ago by shahramzs
  2

• minGPT support

  #32 opened 4 years ago by MasterScrat
  4

• How to serialize the LMOutput object?

  #109 opened a year ago by JamesCao2048
  0

• Support for pre-generated output

  #107 opened a year ago by ThomasDelteil
  0

• Google Colab support only Python 3.10

  #102 opened a year ago by Rolkarolka
  7

• llama support

  #89 opened 2 years ago by srikanthmalla
  4

• Support to LLM like LLAMA-2 and Vicuna?

  #98 opened a year ago by BiEchi
  15

• How to retrieve salience of some specific words?

  #105 opened a year ago by CarhoJohn
  1

• How to add my own model like chinese-bert-wwm-ext

  #104 opened a year ago by Myoontyee
  0

• import ecco ERRO

  #88 opened 2 years ago by nikipi
  9

• Install ecco on Apple M1 Max clip fails

  #101 opened a year ago by verazuo
  1

• AttributeError: 'OutputSeq' object has no attribute 'saliency'

  #86 opened 2 years ago by Claus1
  4

• Support for Distributing Model on Multiple GPU's

  #81 opened 2 years ago by chrispan68
  3

• Adding GENRE model to ecco

  #94 opened 2 years ago by sky-2002
  1

• Error encountered when comparing activation matrices in batch generation loop

  #97 opened a year ago by miguel-kjh
  0

• I can't find jupyter notebook for Input_Saliency.

  #96 opened a year ago by SuperBianC
  0

• Presence of character Ġ before each token in output

  #87 opened 2 years ago by osmanii
  2

• bug -object has no attribute 'lm_head'-

  #91 opened 2 years ago by LeonardRanaldi
  2

• Request for feature addition: Visualization of Neuron Activations and Clustered Neuron Firings

  #93 opened 2 years ago by cristianestojeda
  1

• lm.generate and HuggingFace's generate give different results with do_sample=False

  #92 opened 2 years ago by noeliaferruz
  0

• Bug: Broken links in docs

  #85 opened 2 years ago by keyboardAnt
  0

• token prefix in roberta model?

  #73 opened 2 years ago by sarthusarth
  4

• IndexError when using analysis.pwcca

  #75 opened 2 years ago by demkejon001
  1

• Support for Constrained Beam Search?

  #77 opened 2 years ago by pablogranolabar
  1

• Rankings_watch displaying wrong sequence

  #84 opened 2 years ago by MiriUll
  1

• Save vectorized figure

  #80 opened 2 years ago by jadermcs
  1

• Problem with saliency()

  #78 opened 2 years ago by Poul-Andersson
  1

• How to use ECCO to explain FineTuned T5 for Conditional Generation ?

  #74 opened 3 years ago by tejas20
  3

• Running Eccomap for Pre Trained BertForMaskedLM

  #70 opened 3 years ago by iamakshay1
  1

• Add Primary Attribution support for BERT classification

  #64 opened 3 years ago by jalammar
  4

• KeyError: 'tokenizer_config'

  #66 opened 3 years ago by guustfranssensEY
  2

• Remove `tokenizer_config` usage from the library

  #67 opened 3 years ago by jalammar
  0

• Tokenizer has partial token suffix instead of prefix

  #65 opened 3 years ago by guustfranssensEY
  1

• BART Support

  #31 opened 4 years ago by anshoomehra
  9

• Which french model is supported in ecco library

  #45 opened 3 years ago by Ghani-25
  5

• GPT Generation error on GPU

  #57 opened 3 years ago by jalammar
  1

• T5 Generation Error

  #56 opened 3 years ago by jalammar
  3

• Add a `conda` install option for `ecco`

  #54 opened 3 years ago by sugatoray
  7

• Add models to config

  #52 opened 3 years ago by jalammar
  0

• Merge partial tokens in token visualizations [Generation & Attribution]

  #51 opened 3 years ago by jalammar
  1

• Adding AraBERT

  #42 opened 3 years ago by ay94
  2

• Support for T5-like Seq2SeqLM

  #39 opened 3 years ago by chiragjn
  8

• attention head

  #37 opened 3 years ago by afcarvallo
  5

• Why do we need to construct `one-hot` vectors of the input_ids and then multiply by the embeddings, as opposed to applying the embedding directly?

  #43 opened 3 years ago by vgoklani
  1

• Error when generate using T5

  #44 opened 3 years ago by SSamDav
  0

• How to increase fontsize of saliency output?

  #29 opened 4 years ago by VirajBagal
  2