flax-vision-encoder-decoder-vit-gpt2-coco-en / run_image_captioning_flax.py

Commit History

use custom TrainingArguments
0fc714d

ydshieh commited on

use jax rng
9ca46fa

ydshieh commited on

fix speed 2
a231b72

ydshieh commited on

fix speed
efe62b9

ydshieh commited on

more general
283180e

ydshieh commited on

keep in memory for feature extraction
cb2b435

ydshieh commited on

workaround for tokenizer
fde7d97

ydshieh commited on

better epoch desc on terminal
36f2642

ydshieh commited on

cleanup
c32f83d

ydshieh commited on

cleanup
e70a466

ydshieh commited on

fix
4450aac

ydshieh commited on

separate tokenization and feature extraction
8f31d11

ydshieh commited on

fix tokenization hang on TPU VM when worker > 1
9f6265f

ydshieh commited on

cleanup
c2bb67b

ydshieh commited on

remove incorrect comment
a552802

ydshieh commited on

init
6984f2b

ydshieh commited on