This repository has been archived by the owner on Oct 31, 2022. It is now read-only.

Flaskapp#90

Open

napalmj wants to merge 90 commits intonshepperd:finetuningfrom napalmj:flaskapp

+4,395-41,883

Commits on Feb 18, 2019

update README
WuTheFWasThat
authored

Commits on Feb 19, 2019

reorganize and add temp 0.7
WuTheFWasThat
committed

Commits on Feb 20, 2019

add license
WuTheFWasThat
committed
add conditional samples
WuTheFWasThat
committed
separate out tensorflow install
WuTheFWasThat
committed
shuffle headings
WuTheFWasThat
committed
more warning
WuTheFWasThat
committed
instructinos mention git clone
WuTheFWasThat
committed
Add a Dockerfile and document usage in README

madisonmay
authored and
WuTheFWasThat
committed
fixed unconditional sampling reproducibility issue

Ignacio Lopez-Francos
authored and
WuTheFWasThat
committed
fixed seed arg to ensure reproducibility in conditional-samples model

Ignacio Lopez-Francos
authored and
WuTheFWasThat
committed
update readme
WuTheFWasThat
committed

Commits on Feb 21, 2019

add conditional samples with default settings
WuTheFWasThat
committed
add .gitattributes file to ensure files copied to docker container have LF line endings and all files stay unix on commit

stephan orlowsky
authored and
WuTheFWasThat
committed
Minor: update readme

natemurthy
authored and
WuTheFWasThat
committed

Commits on Feb 27, 2019

Minor: update readme

natemurthy
authored and
WuTheFWasThat
committed
Add documentation for help flags (nshepperd#81 )

ArmaanBhullar
authored and
WuTheFWasThat
committed
slight fix to batch size description
WuTheFWasThat
committed

Commits on Feb 28, 2019

updates
WuTheFWasThat
committed

Commits on Mar 3, 2019

Add finetuning code.
Neil Shepperd
committed
chmod +x
Neil Shepperd
committed
Add finetuning instructions
Neil Shepperd
committed
Fix sample generation with batch_size greater than 1.
Neil Shepperd
committed

Commits on Mar 4, 2019

Python download script (nshepperd#89 )

webproduktion01
authored and
WuTheFWasThat
committed
update download stuff
WuTheFWasThat
committed

Commits on Mar 6, 2019

Commits on Mar 7, 2019

fix for windows (thanks to chrothenbach)
WuTheFWasThat
committed

Commits on Mar 19, 2019

Merge pull request nshepperd#2 from tlkh/finetuning
nshepperd
authored
autoformat
Neil Shepperd
committed
Combine input text files with <|endoftext|> delimiter to ensure there are enough contiguous tokens to sample from.
Neil Shepperd
committed

Commits on Mar 20, 2019

Write losses to summary file for tensorboard.
Neil Shepperd
committed
Add learning rate as command line flag.
Neil Shepperd
committed
Use argparse instead of fire in train.py.
Neil Shepperd
committed

Commits on Mar 21, 2019

Fix encode.py
Neil Shepperd
committed
Add gradient accumulation with default of 5 minibatches
Neil Shepperd
committed

Commits on Mar 25, 2019

Merge remote-tracking branch 'origin/master' into finetuning
Neil Shepperd
committed

Commits on May 2, 2019

Turn off gradient accumulation by default, it shouldn't be needed.
Neil Shepperd
committed

Commits on May 3, 2019

Commits on May 4, 2019

Add gradient checkpointing and another optimization necessary to allow training the 345M model.
Neil Shepperd
committed
Add "validation" loss calculation.
Neil Shepperd
committed

Commits on May 5, 2019

Add toposort to requirements
Tenoke
committed

Commits on May 6, 2019

Merge pull request nshepperd#3 from Tenoke/finetuning
Neil Shepperd
authored

Commits on May 14, 2019

Add option to use SGD for optimizer
Neil Shepperd
committed
Record learning rate in tensorboard logs
Neil Shepperd
committed
Add text in README for --optimizer flag
Neil Shepperd
committed
Reduce default learning rate of train.py.
Neil Shepperd
committed
Merge remote-tracking branch 'origin/master' into finetuning
Neil Shepperd
committed

Commits on May 15, 2019

New feature: add noise to network inputs to regularize against overreacting to typos.
Neil Shepperd
committed
Add top-p sampling
Neil Shepperd
committed
Add top_p to interactive_conditional_samples.py and generate_unconditional_samples.py.
Neil Shepperd
committed
fix typo in top_p
Neil Shepperd
committed
Fix top_p sampling for batch_size>1
Neil Shepperd
committed

Commits on May 16, 2019

allow models to be in a separate folder via models_dir argument (openai#129 )

memo
authored and
WuTheFWasThat
committed

Commits on May 31, 2019

Fix TODO in sample.sample_sequences- Avoid 'leaving last token calculation to while loop' (openai#119 )

albertwujj
authored and
WuTheFWasThat
committed

Commits on Jul 27, 2019

Commits on Aug 15, 2019

Updated README.md
biranchi2018
authored

Commits on Aug 20, 2019

add model card

jackclarksf
authored and
WuTheFWasThat
committed
push 774M model
WuTheFWasThat
committed

Commits on Aug 27, 2019

nucleus sampling
WuTheFWasThat
committed
Merge pull request nshepperd#22 from biranchi2018/biranchi2018-patch-1
N Shepperd
authored
Add note to install cudnn, re nshepperd#8
Neil Shepperd
committed
Add flag to set encoding for text reading and writing, defaulting to utf-8.
Neil Shepperd
committed

Commits on Nov 5, 2019

update readmes
WuTheFWasThat
committed
update model card

jackclarksf
authored and
WuTheFWasThat
committed

Commits on Nov 26, 2019

updated g_form contact
jackclarksf
authored

Commits on Jan 3, 2020

Update LICENSE

cookee12
authored and
WuTheFWasThat
committed
Update README.md
WuTheFWasThat
authored

Commits on Jan 4, 2020

delete
WuTheFWasThat
committed

Commits on Dec 2, 2020

move to azure
WuTheFWasThat
committed

Commits on Mar 2, 2021

Merge branch 'master' into finetuning
nshepperd
committed

Commits on Mar 6, 2021

Fix models_dir issue nshepperd#76
nshepperd
committed
Delete train-horovod.py, which is unmaintained
nshepperd
committed
Fixes to support tensorflow v2.x. Training should still work in v1.x.
nshepperd
committed

Commits on Mar 16, 2021

Add tensor rematerialization.
nshepperd
committed

Commits on Apr 1, 2021

Update twremat.cabal for ghc 9.0
nshepperd
committed

Commits on Apr 13, 2022