Bug fix for: "ImportError: cannot import name 'FileWriter' from 'tensorboard'"#29

Open

Redcof wants to merge 90 commits intohanzhanggit:masterfrom Redcof:master

+10,035-262

Commits on May 5, 2023

Bug fix for
soumensardar
committed
Support for Python3.8 and Python 2.7
soumensardar
committed

Commits on May 9, 2023

- A script to generate custom dataset for StackGAN
soumensardar
committed

Commits on May 10, 2023

- generate custome dataset: filenames, classids
soumentechnoexponent
committed
- filenames and embedding is created
soumensardar
committed

Commits on May 11, 2023

- Proper embeddings, classids, and filenames are generated
soumensardar
committed

Commits on May 12, 2023

- Stage-1 and Stage-2 ran successfully
soumensardar
committed

Commits on May 15, 2023

- code cleanup
soumensardar
committed
- install fasttext through conda
soumensardar
committed
install pytorch through conda
soumensardar
committed

Commits on May 16, 2023

- Doc update
soumensardar
committed
- script to check CUDA
soumensardar
committed
- script to check CUDA support version
soumensardar
committed

Commits on May 19, 2023

- Cuda checks
soumensardar
committed
- Generate data form sqlite
soumensardar
committed
-pascal voc tool update
soumensardar
committed
- generate dataset form sqlite
soumensardar
committed
configuration for Sixray500 stage-1
soumensardar
committed
configuration for Sixray500 stage-2
soumensardar
committed
Retraining configuration for stage-1 with 1000 epoch
soumensardar
committed
Retraining configuration for stage-2 with 1000 epoch
soumensardar
committed
Retraining configuration for stage-2 with 1000 epoch
soumensardar
committed
configuration for 600 captions
soumensardar
committed
Stop config for finetune
soumensardar
committed
Config for stage-2 600 captions
soumensardar
committed

Commits on May 22, 2023

Experiment-3 1107 captions and 432 files
soumensardar
committed
Experiment-3 1107 captions and 432 files, stage-2
soumensardar
committed
Trainer image logging update
soumensardar
committed
Setup config for experiment-3
soumensardar
committed

Commits on May 23, 2023

- support to train fasttext on caption dataset
soumensardar
committed
- Updating scripts and configuration
soumensardar
committed
- save image at last epoch
soumensardar
committed
- config for stage -1
soumensardar
committed
- config for stage -2
soumensardar
committed

Commits on May 24, 2023

- config for stage -1 for 1347 captions
soumensardar
committed
- one script for entire pipeline
soumensardar
committed

Commits on May 25, 2023

- generate test dataset
soumentechnoexponent
committed
- testing script
soumentechnoexponent
committed
- script update to generate data with fasttext encoding
soumensardar
committed
- config update for express train and test
soumensardar
committed
- increase epoch form 1 to 500
soumensardar
committed

Commits on May 26, 2023

- script generation update
soumensardar
committed
Preparing for training with 2183 caption 1000D
soumensardar
committed

Commits on May 29, 2023

- config for 2500 captions
soumensardar
committed
- Fasttext algorithm change
soumensardar
committed
- Aspect ratio resize transform
soumensardar
committed
- cbow 2048D, 200 epoch, zdim 200
soumensardar
committed
- resize bug fix
soumensardar
committed

Commits on May 30, 2023

- config for cbow1024D no crop 2500 caption and zdim 100
soumensardar
committed
- minimum caption length=3
soumensardar
committed

Commits on Jun 1, 2023

- script update foe experiment 11: sixray_2500_ftt_1024D_skipgram_nocrop
soumensardar
committed
WIP for openai access
soumensardar
committed
Bulk OpenAPI embedding text processor
soumensardar
committed

Commits on Jun 2, 2023

- bulk caption loading
soumensardar
committed
- single caption loading
soumensardar
committed
- Embedding and POC with OpenAI is done
soumensardar
committed
- file check in progress
soumensardar
committed
- config update for openai embedding
soumensardar
committed

Commits on Jun 3, 2023

- FP32 to FP64 auto adaption
soumensardar
committed
- experiment12
soumensardar
committed
texts to captions
soumensardar
committed
texts to captions
soumensardar
committed
- ready for experiment12
soumensardar
committed
run with 1000 epoch openai embeddings
soumensardar
committed
ignore pycache
soumensardar
committed
floating point precision control
soumensardar
committed

Commits on Jun 8, 2023

11805 caption and 2381 images is available
soumensardar
committed
config update for 2381 images
soumensardar
committed
Prepare for SQLite data generation
soumensardar
committed
32 batch size
soumensardar
committed
16 batch size
soumensardar
committed
64 batch size
soumensardar
committed
epoch 1000
soumensardar
committed
- Generate test images during training
soumensardar
committed
- CUDA clean in test phase
soumensardar
committed
- PYTORCH_CUDA_ALLOC_CONF
soumensardar
committed
- bug fix during memory cleaning
soumensardar
committed
Fixing CUDA PyTorch Future deprecation waring
soumensardar
committed

Commits on Jun 9, 2023

- capping captions to 4 as a few files are inconsistently contains up to 6 captions
soumentechnoexponent
committed

Commits on Jun 12, 2023

- Hyperparameter double
soumensardar
committed
- Hyperparameter double
soumensardar
committed
- copy config
soumensardar
committed
- copy config
soumensardar
committed

Commits on Jun 13, 2023

- Finetune Experiment15.
soumensardar
committed
- fix memory count
soumensardar
committed

Commits on Jul 7, 2023

- mild code update for WorkRNN models
soumensardar
committed

Commits on Jul 8, 2023

- saving git checksum and generating a config
soumensardar
committed
- config update
soumensardar
committed
buf gix
soumensardar
committed

Commits on Jul 10, 2023

config for cnn-rnn word embedding
soumensardar
committed