tarrade proj_multilingual_text_classification issues

tarrade / proj_multilingual_text_classification

Explore multilingal text classification using embedding, bert and deep learning architecture

Apache License 2.0

5 stars 2 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

TPU for TensorFlow 2.1 not yet available for AI Platform training

#43 tarrade closed 4 years ago
3
implement example with python client library for model training instead of gcloud

#42 tarrade closed 4 years ago
1
region parameter not use in the yaml file when using ai platform training

#41 tarrade closed 4 years ago
1
hyperparameter tuning with keras on GCP: failing for no reason

#40 tarrade closed 4 years ago
2
see how to use TPU and keras with minimal modification

#39 tarrade closed 4 years ago
1
move the model part in a function for ai platform training

#38 tarrade closed 4 years ago
1
switch from gcloud to python client library for model serving

#37 tarrade closed 4 years ago
1
clean up def batch_translate_text

#36 tarrade closed 4 years ago
1
Store tf.data after the transformation needed for BERT in TFRecord files (preprocessing done only one)

#35 tarrade closed 4 years ago
1
Batch Translation on GCP creates problems

#34 vluechinger closed 4 years ago
1
be able to use any kind of model like xml-roberta

#33 tarrade closed 4 years ago
0
max length with BERT: 512 or it was only in the original paper from Google ?

#32 tarrade closed 4 years ago
1
Benchmark classical ML pipeline for the chosen datasets

#31 vluechinger opened 4 years ago
1
Implement LASER and see if it is useful

#30 vluechinger closed 4 years ago
1
compute the multilingual BERT model (512 sequence length) with a GPU and compare computation time and costs to the CPU run

#29 vluechinger closed 4 years ago
1
compare multilingual BERT model to alternatives like XLM-RoBERTa

#28 vluechinger closed 4 years ago
1
print_info_data (function in preprocessing) is incompatible with other data formats

#27 vluechinger closed 4 years ago
1
add path when downloading pre trained model or tokenizer to not download them all the time

#26 tarrade closed 4 years ago
1
add profiler in TensorBoard

#25 tarrade closed 4 years ago
1
hyperparameter tunning: keras with Keras or GCP

#24 tarrade closed 4 years ago
1
model interpretation with LIME and SHAP

#23 tarrade opened 4 years ago
3
Find imdb-like movie review datasets in FR, DE and IT

#22 tarrade closed 4 years ago
3
Dashboard migration for GCP notebook

#21 tarrade opened 4 years ago
0
TensorBoard: it dispaly only 1 point instead of 5 batches * 2 epoch = 10 steps (not always!)

#20 tarrade closed 4 years ago
1
Callback: get loss and accuracy for both training and testing and per step (NOT epoch)

#19 tarrade closed 4 years ago
1
Migration of the code to Tensorflow 2.2.0

#18 tarrade closed 4 years ago
6
Train the model with English Imdb (sentiment analysis movies review) but do inference with other datasets (DE, IT or FR)

#17 tarrade closed 4 years ago
1
Use PCA and display embedding for few identique words but in 4 differents languages

#16 tarrade opened 4 years ago
1
Save and serve the model using AI Platform serving

#15 tarrade closed 4 years ago
2
Use custom image and AI Platform training GCP service to train the model

#14 tarrade closed 4 years ago
1
Use AI Platform trainig GCP service to train the model

#13 tarrade closed 4 years ago
2
Use GPU for training with stable model

#12 tarrade closed 4 years ago
1
How to use TensorBoard with jupyter lab extension ?

#11 tarrade closed 4 years ago
1
Issue with tf.data.experimental.cardinality

#10 tarrade closed 4 years ago
3
How to inspect a BERT model ?

#9 tarrade closed 4 years ago
1
TensorBoard: display plot as a function of nb epoch and as nb of steps

#8 tarrade closed 4 years ago
1
What preprocessing steps are required / possible when using BERT?

#7 vluechinger closed 4 years ago
2
How can we plot the model history in steps?

#6 vluechinger closed 4 years ago
1
How are the pretrained models structured and how can we access the structure?

#5 vluechinger closed 4 years ago
1
What entails fine-tuning in NLP transfer learning?

#4 vluechinger closed 4 years ago
1
How to choose the batch size for training and validation datasets?

#3 vluechinger closed 4 years ago
1
Flake8 error message in the AI Platform Notebook

#2 vluechinger closed 4 years ago
2
TensorBoard blocked by corporate proxy ?

#1 tarrade closed 4 years ago
2