Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / huggingface/pytorch-openai-transformer-lm issues and pull requests

#60 - Implementation of Similarity Head

Pull Request - State: open - Opened by TEGELB about 3 years ago

#59 - Training from scratch: Repeated and mangled words

Issue - State: open - Opened by maruker about 5 years ago

#58 - Instructions for encoding own sentences

Issue - State: open - Opened by izaskr over 5 years ago - 1 comment

#57 - Running on new dataset similar to rocstories

Issue - State: open - Opened by priyanka-chaudhary over 5 years ago

#55 - ConvAI

Issue - State: open - Opened by bkj over 5 years ago

#54 - vocab = n_vocab + n_special + n_ctx means?

Issue - State: open - Opened by JiahangOK over 5 years ago - 1 comment

#53 - Why do we need to apply mask while fine tuning?

Issue - State: closed - Opened by pranoy-k over 5 years ago - 4 comments

#52 - Implementation of Seq2Seq with Transformer

Issue - State: open - Opened by bhedayat over 5 years ago

#51 - why it didn't use softmax in computing multichoice loss

Issue - State: closed - Opened by eveliao over 5 years ago - 1 comment

#50 - problems about DoubleHeadModel implementation

Issue - State: closed - Opened by eveliao over 5 years ago - 1 comment

#49 - Use the embedding dropout

Pull Request - State: closed - Opened by sgugger over 5 years ago - 1 comment

#48 - Potentially incorrect regex in text_utils.py

Issue - State: open - Opened by schmmd over 5 years ago

#47 - Why is output vocab including positional embeddings?

Issue - State: open - Opened by OanaMariaCamburu almost 6 years ago - 2 comments

#46 - Retrain the LM on new dataset?

Issue - State: open - Opened by fabrahman almost 6 years ago

#45 - Encoder paddings influence results?

Issue - State: open - Opened by OanaMariaCamburu almost 6 years ago

#44 - How does position embedding implementation work?

Issue - State: closed - Opened by bcserna almost 6 years ago - 2 comments

#43 - How should one modify the code to successfully run text classification?

Issue - State: closed - Opened by davidefiocco almost 6 years ago - 7 comments

#42 - help to understand bpe logic

Issue - State: closed - Opened by BogdanDidenko about 6 years ago - 2 comments

#40 - How to create transforms for entailment task?

Issue - State: open - Opened by lordzuko about 6 years ago - 12 comments

#39 - Confused by multiply n to both clf_logits and clf_losses?

Issue - State: open - Opened by Vimos about 6 years ago - 3 comments

#38 - Vocabulary size code explanation and occasionally shape error

Issue - State: closed - Opened by Vimos about 6 years ago - 2 comments

#37 - Removing unused decoder for TransformedModel class.

Pull Request - State: closed - Opened by rodgzilla about 6 years ago - 1 comment

#36 - Results and questions on text generation experiments with pretrained LM model

Issue - State: closed - Opened by xiaoda99 about 6 years ago - 10 comments

#35 - Text generation with pretrained LM model

Pull Request - State: closed - Opened by xiaoda99 about 6 years ago - 2 comments

#34 - So we can not change the word embedding with the pretrained LM?

Issue - State: open - Opened by herbertchen1 about 6 years ago - 1 comment

#33 - How to use Inference?

Issue - State: open - Opened by masati91 about 6 years ago - 4 comments

#32 - Redundant decoder

Issue - State: closed - Opened by joshim5 about 6 years ago - 1 comment

#31 - Avoid model overfitting

Issue - State: open - Opened by BangLiu about 6 years ago - 9 comments

#29 - Instructions on how to train a language model from scratch

Issue - State: open - Opened by Froskekongen about 6 years ago - 9 comments

#28 - loading pretrained open ai model

Issue - State: open - Opened by mehdimashayekhi about 6 years ago - 3 comments

#27 - Add comment explaining the encoding of the position information.

Pull Request - State: closed - Opened by sharpsy about 6 years ago - 1 comment

#26 - Clarifying last step of the 'transform_roc' function

Issue - State: open - Opened by sharpsy about 6 years ago - 7 comments

#25 - Simplifying the use of the model to perform different tasks

Pull Request - State: closed - Opened by rodgzilla about 6 years ago

#24 - Having various network heads

Issue - State: open - Opened by rodgzilla about 6 years ago - 1 comment

#23 - Removing unused field of the Transformer class.

Pull Request - State: closed - Opened by rodgzilla about 6 years ago - 1 comment

#22 - Noise shape dropout

Pull Request - State: closed - Opened by rodgzilla about 6 years ago - 3 comments

#21 - Can someone explain this line?

Issue - State: open - Opened by teucer about 6 years ago - 4 comments

#20 - Removing python 2.7 check from Travis.

Pull Request - State: closed - Opened by rodgzilla about 6 years ago - 1 comment

#19 - what is the use of dropout in the Transformer?

Issue - State: open - Opened by teucer about 6 years ago - 2 comments

#18 - Using the ellipsis to allow the function to work with dim 3 or 4 tensors

Pull Request - State: closed - Opened by rodgzilla about 6 years ago - 1 comment

#17 - Cleaning the data encoding code.

Pull Request - State: closed - Opened by rodgzilla about 6 years ago - 1 comment

#16 - Updating README.md

Pull Request - State: closed - Opened by rodgzilla about 6 years ago - 1 comment

#15 - Fixing issue #13 and #14.

Pull Request - State: closed - Opened by rodgzilla about 6 years ago - 1 comment

#14 - DoubleHeadModel is not instanciated when n_gpu <= 1

Issue - State: closed - Opened by rodgzilla about 6 years ago - 1 comment

#13 - Object is not specified

Issue - State: closed - Opened by Oktai15 about 6 years ago - 4 comments

#12 - In `transform_roc`, why do we need `xmb[:, :, :, 1] `?

Issue - State: open - Opened by FrankWork about 6 years ago - 3 comments

#11 - How does Dropout2d help in cloze task?

Issue - State: closed - Opened by sai-prasanna about 6 years ago - 12 comments

#10 - [WIP] Add multi-gpu support

Pull Request - State: closed - Opened by nottombrown about 6 years ago - 2 comments

#9 - Remove globals to make code easier to follow

Pull Request - State: closed - Opened by nottombrown about 6 years ago - 1 comment

#8 - Using negative index to reshape the input tensor.

Pull Request - State: closed - Opened by rodgzilla about 6 years ago - 1 comment

#7 - Pre-trained LMHead

Issue - State: closed - Opened by rodgzilla over 6 years ago - 3 comments

#6 - fix the scope of optimizer

Pull Request - State: closed - Opened by soskek over 6 years ago - 3 comments

#5 - Create .travis.yml

Pull Request - State: closed - Opened by cclauss over 6 years ago - 1 comment

#4 - fix arguments for initializing a model

Pull Request - State: closed - Opened by soskek over 6 years ago - 1 comment

#3 - a git clone is large because history has the "add model" commit

Issue - State: closed - Opened by soumith over 6 years ago - 1 comment

#2 - dimensioning bug?

Issue - State: closed - Opened by jtatusko over 6 years ago - 4 comments

#1 - Fixes to allow the model to load

Pull Request - State: closed - Opened by DeNeutoy over 6 years ago - 1 comment