google-research/albert issues and pull requests

#279 - Albert

Issue - State: closed - Opened by Albertjabal 7 months ago

#278 - Albet

Issue - State: closed - Opened by Albertjabal 7 months ago

#277 - Why do I find inconsistencies between the output of my ALBERT model converted to ONNX format and tested with ONNX Runtime, compared to the original PyTorch format model?

Issue - State: open - Opened by Orient12 9 months ago

#276 - Load in Browser Tensorflow

Issue - State: open - Opened by sysmaya about 1 year ago

#275 - Made README.md more Readable

Pull Request - State: open - Opened by Aryanutkarsh over 1 year ago

#274 - Replace deprecated `tf.contrib.data.map_and_batch(...)` with `tf.data.experimental.map_and_batch(...)`.

Pull Request - State: closed - Opened by copybara-service[bot] over 1 year ago

#273 - Replace deprecated `tf.contrib.data.parallel_interleave(...)` with `tf.data.experimental.parallel_interleave(...)`.

Pull Request - State: closed - Opened by copybara-service[bot] over 1 year ago

#272 - Bump tensorflow from 1.15.2 to 2.11.1

Pull Request - State: open - Opened by dependabot[bot] over 1 year ago
Labels: dependencies

#271 - Bump tensorflow from 1.15.2 to 2.9.3

Pull Request - State: closed - Opened by dependabot[bot] almost 2 years ago - 1 comment
Labels: dependencies

#270 - Improvement to how the `app` and `pages` files conflict is shown. Especially the last log line `"pages/" - "app/"` made it seem like you should remove the `pages` folder altogether. This was a bug in how the `''` case was displayed. After having a look at this I went further and added exactly which file caused the conflict given that `app` allows you to create `app/(home)/page.js` and such it saves some digging for what the actual conflicting file is. Similarly in `pages` both `pages/dashboard/index.js` and `pages/dashboard.js` are possible.

Issue - State: closed - Opened by albrtkinn almost 2 years ago

#269 - The results can't be reproduced

Issue - State: open - Opened by kavin525zhang about 2 years ago - 2 comments

#268 - tokenization: log spm usage only in debug mode to avoid console spamming

Pull Request - State: open - Opened by stefan-it about 2 years ago

#267 - Wrong evaluate result on Squad2.0

Issue - State: closed - Opened by YuxuanJiang1 over 2 years ago

#266 - Difference between v1 and v2 for xxlarge

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

#265 - Bump tensorflow from 1.15.2 to 2.7.2

Pull Request - State: closed - Opened by dependabot[bot] over 2 years ago - 1 comment
Labels: dependencies

#264 - when training in Race , The eval_accuracy is flat , it only has three numbers which are 0.0, 0.33334, 0.66667, 1.0

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

#263 - Bump tensorflow from 1.15.2 to 2.6.4

Pull Request - State: closed - Opened by dependabot[bot] over 2 years ago - 1 comment
Labels: dependencies

#262 - Remove unused comments related to Python 2 compatibility.

Pull Request - State: closed - Opened by copybara-service[bot] over 2 years ago

#261 - Explicitly import estimator from tensorflow as a separate import instead of

Pull Request - State: open - Opened by copybara-service[bot] over 2 years ago

#260 - How to get the test embeddings from output of fine-tuned model (tutorial)

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

#259 - Prediction Fails on default Colab

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

#258 - Bump tensorflow from 1.15.2 to 2.5.3

Pull Request - State: closed - Opened by dependabot[bot] over 2 years ago - 1 comment
Labels: dependencies

#257 - Default Tutorial Not Working - Can't download MRPC data

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

#256 - Probable error on line 306 in `create_pretraining_data.py` for albert

Issue - State: open - Opened by wjdghks950 over 2 years ago

#252 - Minor bug fix

Pull Request - State: closed - Opened by copybara-service[bot] almost 3 years ago

#251 - Internal Google change

Pull Request - State: closed - Opened by copybara-service[bot] almost 3 years ago
Labels: cla: yes

#250 - which word segmentation tool is used for pretraining Chinese ALBERT

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

#249 - Discrepancy in tokenization results using albert's tokenizer and sentencepiece library

Issue - State: open - Opened by anjali-chadha almost 3 years ago

#248 - fine tune on my own English dataset

Issue - State: open - Opened by cxx-cindy about 3 years ago

#247 - Wrong pieces for control symbols after loading SentencepieceProcessor from official model

Issue - State: closed - Opened by JanKaul about 3 years ago - 2 comments

#246 - Bump tensorflow from 1.15.2 to 2.5.1

Pull Request - State: closed - Opened by dependabot[bot] about 3 years ago - 1 comment
Labels: cla: yes, dependencies

#245 - 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: cla: yes, dependencies

#244 - Add an "excluded_tvars" arg in the create_optimizer function, which can be used to forbid updating the selected variables.

Pull Request - State: closed - Opened by copybara-service[bot] over 3 years ago
Labels: cla: yes

#243 - albert base fine-tuned on squad2.0 gets stuck in loop when predicting on new file

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

#242 - The exact English pretraining data and Chinese pretraining data that are exact same to the BERT paper's pretraining data.

Issue - State: open - Opened by guotong1988 over 3 years ago

#221 - TypeError: '<' not supported between instances of 'int' and 'NoneType' while only predicting, try to fix it, but not sure is it make sense

Issue - State: open - Opened by alznn about 4 years ago - 2 comments

#220 - why running run_squad_v2.py in prediction mode with dev-v2.0.json gets zero neg example

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

#219 - Add a signature that exposed the SOP log probabilities.

Pull Request - State: closed - Opened by copybara-service[bot] about 4 years ago
Labels: cla: yes

#218 - Is it possible to load pretrained model's checkpoint and pretrain other corpus with that checkpoint?

Issue - State: open - Opened by sikaleo about 4 years ago

#217 - invalid load key, '\x12'

Issue - State: open - Opened by liu-x-p over 4 years ago - 1 comment

#216 - How to prepare for the environment/data for fine-tuning?

Issue - State: closed - Opened by 517030910405 over 4 years ago - 2 comments

#215 - Allow run_classifier to export fine-tuned model as a SavedModel.

Pull Request - State: closed - Opened by copybara-service[bot] over 4 years ago
Labels: cla: yes

#214 - use custom vocab.txt

Issue - State: open - Opened by 2696120622 over 4 years ago - 1 comment

#213 - is inner_group_num exactly equal to num_hidden_layers/num_hidden_groups?

Issue - State: open - Opened by lsy641 over 4 years ago

#212 - how to do inference for SQuAD style QA task using ALBERT?

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

#211 - fine_tune albert with run_classifier.py on GPU but the GPU-util is 0,why and how?

Issue - State: open - Opened by yuanjun1991 over 4 years ago - 2 comments

#210 - Where is the function of Factorized Embedding Parameterization?

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

#209 - fix the attention mask description error and a cola evaluation calculation error.

Pull Request - State: closed - Opened by copybara-service[bot] over 4 years ago
Labels: cla: yes

#208 - Description problem of attention_mask.

Issue - State: open - Opened by liushaoweihua over 4 years ago

#207 - model isn't learning

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

#206 - different parameter position with api doc

Issue - State: open - Opened by cyjj over 4 years ago

#205 - get Matthew_corr = 0 in eval on CoLA

Issue - State: open - Opened by cyjj over 4 years ago

#204 - Extremely large RAM consumption by create_pretraining.py

Issue - State: open - Opened by sultanovazamat over 4 years ago - 3 comments

#203 - Modify default value of the max_seq_length at the stage of fine-tuning on GLUE

Pull Request - State: closed - Opened by gongel over 4 years ago - 3 comments
Labels: cla: yes

#202 - Training without SOP Loss

Issue - State: open - Opened by divkakwani over 4 years ago

#200 - [question] Cross-layer parameter sharing

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

#199 - ModuleNotFoundError: No module named 'albert'

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

#198 - ALBERT-xxlarge V2 training on TPU V3-512 extremely slow

Issue - State: closed - Opened by agemagician over 4 years ago - 15 comments

#197 - albert large_v2

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

#194 - GPU not used when fine-tuning on SQuAD 2.0

Issue - State: open - Opened by Inception95 over 4 years ago - 4 comments

#188 - TPU training problem

Issue - State: closed - Opened by manueltonneau over 4 years ago - 2 comments

#181 - How to extract the word embedding

Issue - State: open - Opened by 256785 over 4 years ago - 1 comment

#174 - [Question] Is there any fine-tuned ALBERT model on GLUE tasks available?

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

#169 - Running run_classifier.py error

Issue - State: closed - Opened by junrong1 over 4 years ago - 2 comments

#168 - Ran out of memory in memory space hbm on RACE xlarge v3 on TPU v2-8

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

#166 - When I ran the file ‘run_pretraining.py’, there was a problem

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

#165 - I'm going to pretrain Albert on my own data. What am I going to do?What format is the training data?

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

#137 - Exceeding Memory

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

#135 - Squad v2 does not seem to train.

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

#132 - Has anyone reproduced SQuAD 1.1 score(90.2/83.2) on albert-base V2?

Issue - State: closed - Opened by YJYJLee over 4 years ago - 8 comments

#123 - Hyper-params settings for MNLI fine-tuning using Albert-v2

Issue - State: open - Opened by agupta74 over 4 years ago - 2 comments

#106 - SQuAD vocab file for ALBERT

Issue - State: closed - Opened by abdolence almost 5 years ago - 6 comments

#105 - Index Out of Range Error in tokenization using TF Hub for Pretrained Albert Models

Issue - State: closed - Opened by jsmith09 almost 5 years ago - 9 comments

#104 - [ALBERT] Pretraining

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

#103 - [ALBERT] Checkpoints for pretrained models

Issue - State: closed - Opened by SahanaRamnath almost 5 years ago - 3 comments

#102 - [ALBERT] --random_next_sentence option in create_pretraining_data.py

Issue - State: open - Opened by tomohideshibata almost 5 years ago - 1 comment

#101 - [ALBERT] Tokenization crashes while trying to finetune classifier with TF Hub model

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

#100 - [ALBERT]: LookupError: gradient registry has no entry for: AddV2

Issue - State: closed - Opened by wxp16 almost 5 years ago - 8 comments

#99 - [ALBERT]Has anyone reproduced ALBERT a scores on GLUE dataset?

Issue - State: closed - Opened by lonePatient almost 5 years ago - 8 comments

#98 - [ALBERT]: In run_squad_sp, convert_examples_to_features gives error in case sentence piece model is not provided.

Issue - State: open - Opened by Rachnas almost 5 years ago - 10 comments

GitHub / google-research/albert issues and pull requests