minimaxir/gpt-2-simple issues and pull requests

#316 - GPT-2 implementation problem

Issue - State: open - Opened by sanhai77 6 months ago

#315 - how to run only generation on windows pls

Issue - State: open - Opened by Sandy4321 8 months ago

#314 - ValueError: Can't load save_path when it is None.

Issue - State: open - Opened by Mariotim3 over 1 year ago

#313 - doesnt work

Issue - State: closed - Opened by flzpfnove over 1 year ago - 1 comment

#312 - gpt

Issue - State: open - Opened by fotoberia over 1 year ago

#311 - How to use the same trained model with C++ and make questions to it?

Issue - State: open - Opened by paulocoutinhox over 1 year ago

#310 - Resource exhausted error during training

Issue - State: open - Opened by DNJ-JINI over 1 year ago

#309 - any way to clear up vram after running the model

Issue - State: open - Opened by FlashlightET over 1 year ago

#308 - Can you include code for this to handle GPT-Neo-125M

Issue - State: open - Opened by thistleknot over 1 year ago - 1 comment

#307 - Can you force the framework to use CPU?

Issue - State: open - Opened by Leli1024 over 1 year ago

#306 - Response generated in console, but can't retrieve it?

Issue - State: open - Opened by sneezeparty over 1 year ago - 2 comments

#305 - CVE-2007-4559 Patch

Pull Request - State: open - Opened by TrellixVulnTeam almost 2 years ago

#304 - Simple question

Issue - State: open - Opened by Dsanthosh2006 almost 2 years ago - 1 comment

#303 - Results of finetuned model is different in Google Colab and on local machine

Issue - State: closed - Opened by mushtaqpatel0505 about 2 years ago - 1 comment

#302 - Thank you for making this repo.

Issue - State: open - Opened by Tylersuard about 2 years ago

#301 - gpt

Issue - State: open - Opened by symebaline about 2 years ago - 1 comment

#300 - Converting generated data to not tokenized default version

Issue - State: open - Opened by erenarkangil about 2 years ago

#299 - (not an issue) training 124m model works perfectly on 3060

Issue - State: closed - Opened by FlashlightET about 2 years ago

#298 - How to load pretrained model in a new notebook without retraining?

Issue - State: open - Opened by Tylersuard over 2 years ago - 1 comment

#297 - Newbie issue

Issue - State: open - Opened by Sadrahel over 2 years ago - 3 comments

#296 - README Tutorial fails

Issue - State: open - Opened by mrx23dot over 2 years ago - 1 comment

#295 - can't encode character

Issue - State: open - Opened by mrx23dot over 2 years ago - 2 comments

#294 - Remove space from "My Drive" folder path

Pull Request - State: closed - Opened by leetfin over 2 years ago

#293 - How to finetune it for a multivariate feature space?

Issue - State: open - Opened by MOvi996 over 2 years ago

#292 - How to train on text with fewer than 1,024 tokens?

Issue - State: open - Opened by Mennaruuk over 2 years ago

#291 - Impossible to finetune model bigger than 124M or 125M parameters (colab)

Issue - State: open - Opened by Blazeolmo over 2 years ago - 1 comment

#290 - Added UTF-8 support for special characters

Pull Request - State: closed - Opened by sobisonator over 2 years ago

#289 - New transformer architecture

Issue - State: open - Opened by aletote over 2 years ago - 1 comment

#288 - Finetuning multiple datasets

Issue - State: open - Opened by kobel240 over 2 years ago - 1 comment

#287 - Minimaxir master

Pull Request - State: closed - Opened by BorkforceOne over 2 years ago

#286 - Error when finetuning my model : ValueError: Variable model/wpe already exists

Issue - State: open - Opened by Fqlox over 2 years ago

#285 - Resource exhausted: OOM when allocating tensor with shape on 125M with 8GB VRAM

Issue - State: open - Opened by Fqlox over 2 years ago - 1 comment

#284 - Is there a way to have more than 1024 tokens?

Issue - State: open - Opened by aletote over 2 years ago - 1 comment

#283 - Changed hard coded length value for variable

Pull Request - State: closed - Opened by hdkmraf over 2 years ago

#282 - Consistently receiving resource exhaustion errors while training on Colab

Issue - State: open - Opened by mackyufo almost 3 years ago

#281 - How to convert PT (Pytorch) or H5 (Tensorflow) models for gpt-2-simple?

Issue - State: open - Opened by MaxGodTier almost 3 years ago

#280 - auto save to google drive every ex: 1000 step

Issue - State: open - Opened by medozeus almost 3 years ago

#279 - about <|startoftext|> <|endoftext|> every line or every article

Issue - State: open - Opened by medozeus almost 3 years ago - 1 comment

#278 - how much is steps=1000?

Issue - State: closed - Opened by ratneshjoshi almost 3 years ago

#277 - ResourceExhaustedError: failed to allocate memory

Issue - State: open - Opened by Merzmensch almost 3 years ago - 2 comments

#276 - Revert

Pull Request - State: closed - Opened by BenjaminWegener about 3 years ago

#275 - Supporting Tensorflow 2

Pull Request - State: closed - Opened by duhaime about 3 years ago - 19 comments

#274 - indices[0,1024] = 1024 is not in [0, 1024)

Issue - State: open - Opened by DragonDmoney about 3 years ago

#273 - Fixes "List index out of range" error

Pull Request - State: open - Opened by bearsharktopus-dev about 3 years ago - 1 comment

#272 - Provide reuse option when creating TF Session

Pull Request - State: open - Opened by danielenricocahall over 3 years ago

#271 - Are there any parameters to prevent/discourage repetition?

Issue - State: open - Opened by googleooer over 3 years ago

#270 - How to properly process a dataset

Issue - State: open - Opened by pavel-avilov over 3 years ago

#269 - tensorflow.contrib error when using Python 3.8+

Issue - State: open - Opened by Theelx over 3 years ago - 1 comment

#268 - Update gpt_2.py

Pull Request - State: open - Opened by tibrado over 3 years ago

#267 - Local using model trained in google colab

Issue - State: open - Opened by lolomap over 3 years ago

#266 - why gpt encoder norm is before mlp while the original transformer is mlp before norm?

Issue - State: open - Opened by githubutilities over 3 years ago

#265 - added encoding utf-8 specification

Pull Request - State: open - Opened by fastflair over 3 years ago

#264 - How to train another language?

Issue - State: open - Opened by pendekarcode over 3 years ago

#263 - 335M model: JSONDecodeError: Unexpected UTF-8 BOM (decode using utf-8-sig)

Issue - State: closed - Opened by drusepth over 3 years ago - 2 comments

#262 - Setting random seed when finetuning

Issue - State: open - Opened by mesotron over 3 years ago

#261 - MyDrive vs My Drive when trying to copy shakespeare file in notebook

Issue - State: open - Opened by demongolem-biz over 3 years ago

#260 - Add a new token to the tokenizer before fine-tuning

Issue - State: open - Opened by FatemehMashhadi over 3 years ago

#259 - Loss is NaN when using multiple GPUs

Issue - State: open - Opened by kwrobert over 3 years ago

#258 - Is it possible to use a trained model as base for a new fine tuning?

Issue - State: closed - Opened by lucianosilvi over 3 years ago - 1 comment

#257 - Out Of Memory On Google Colab When Training With Big Dataset

Issue - State: open - Opened by nikhilanayak over 3 years ago - 1 comment

#256 - Unable to generate Texts with GPU

Issue - State: open - Opened by AdaUchendu over 3 years ago

#255 - Sort results by loss

Issue - State: open - Opened by Honyant over 3 years ago

#254 - Bump tensorflow from 1.15.2 to 2.4.0

Pull Request - State: closed - Opened by dependabot[bot] over 3 years ago - 1 comment
Labels: dependencies

#253 - Update model download URL

Pull Request - State: closed - Opened by mangtronix over 3 years ago

#252 - download_gpt2 no longer seems to be downloading models

Issue - State: open - Opened by mesotron over 3 years ago - 11 comments

#251 - Question: Should I artificially pad with spaces to generate more efficient tokens?

Issue - State: open - Opened by flesler over 3 years ago - 1 comment

#250 - le 'tensorflow.python.util.compat' has no attribute 'v1'

Issue - State: closed - Opened by yangboz over 3 years ago - 2 comments

#249 - sess returned as None type when loaded back into an environment

Issue - State: closed - Opened by Travis-Barton over 3 years ago - 2 comments

#248 - Distributed gpt-2-simple project (suggestion)

Issue - State: open - Opened by simsim314 over 3 years ago

#247 - New to GPT-2

Issue - State: open - Opened by sodapopwithanh over 3 years ago

#243 - ModuleNotFoundError: No module named 'gpt_2_simple'

Issue - State: open - Opened by dotjavafile almost 4 years ago - 3 comments

#239 - Bump tensorflow from 1.15.2 to 2.3.1

Pull Request - State: closed - Opened by dependabot[bot] almost 4 years ago - 2 comments
Labels: dependencies

#238 - Training is using my CPU not my GPU

Issue - State: open - Opened by EtchProject almost 4 years ago - 5 comments

#234 - Train in colab on local runtime?

Issue - State: open - Opened by mackyufo almost 4 years ago - 1 comment

#227 - Created a fix for my issue https://github.com/minimaxir/gpt-2-simple/…

Pull Request - State: open - Opened by nameofname about 4 years ago - 2 comments

#223 - Error generate() - During handling of the above exception, another exception occurred

Issue - State: open - Opened by skorotkiewicz about 4 years ago - 1 comment

#221 - setup.py is missing tensorflow dependency

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

#220 - KeyError when replacing loss

Issue - State: closed - Opened by aclifton314 about 4 years ago

#219 - generate_samples method producing random results in finetune.

Issue - State: closed - Opened by aclifton314 about 4 years ago

#213 - UnicodeEncodeError: 'charmap' codec can't encode character. character maps to <undefined>

Issue - State: open - Opened by ahmadalli over 4 years ago - 7 comments

#211 - Generating a small amount of tokens takes just as long as generating a whole bunch

Issue - State: open - Opened by flarn2006 over 4 years ago - 1 comment

#198 - GPT2 on flask

Issue - State: closed - Opened by GodSaveTheDucks over 4 years ago - 3 comments

#184 - generated texts are the same as input

Issue - State: open - Opened by shenmeguimlf over 4 years ago - 6 comments

#178 - Cannot convert to pytorch using huggingface

Issue - State: open - Opened by designgrande over 4 years ago - 6 comments

#171 - Multi-gpu for gpt2.generate

Issue - State: open - Opened by picocms over 4 years ago - 2 comments

#156 - InvalidArgumentError during text generation

Issue - State: open - Opened by DGaffney over 4 years ago - 1 comment

#153 - include_prefix=False does not exclude the prefix

Issue - State: open - Opened by InconsolableCellist over 4 years ago - 6 comments

#150 - Support TF2.0

Issue - State: open - Opened by arlyon almost 5 years ago - 6 comments

#148 - add model listing capability

Pull Request - State: closed - Opened by jqueguiner almost 5 years ago

#145 - finetune json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)

Issue - State: open - Opened by thiagorova almost 5 years ago - 9 comments

#139 - perplexity of a sentence

Issue - State: open - Opened by RakshaAg almost 5 years ago - 4 comments

#135 - Add 1558M to GPU finetuning blacklist

Issue - State: closed - Opened by WAUthethird almost 5 years ago - 4 comments

#128 - Is there a parameter for human prompt?

Issue - State: closed - Opened by lenhhoxung86 almost 5 years ago - 2 comments

GitHub / minimaxir/gpt-2-simple issues and pull requests