#celeba256#ValueError: cannot reshape array of size 12116 into shape (63,63,3)

[sumorday]

Hello, I'm trying to run the Celeba256x256 dataset, which has already been processed into LMDB format. I use the following command:

CUDA_VISIBLE_DEVICES=0 bash ./bash_scripts/run.sh ./test_args/celeb256_adm.txt

However, I keep receiving an error: ValueError: cannot reshape array of size 12116 into shape (63,63,3), CalledProcessError: Command '['/home/edward/local/anaconda3/envs/torch/bin/python', 'train_flow_latent.py', '--exp', 'celeb256_f8_adm', '--dataset', 'celeba_256', '--datadir', '../data/celeba/celeba-lmdb/', '--batch_size', '112', '--num_epoch', '500', '--image_size', '256', '--f', '8', '--num_in_channels', '4', '--num_out_channels', '4', '--nf', '256', '--ch_mult', '1', '2', '2', '2', '--attn_resolution', '16', '8', '--num_res_blocks', '2', '--lr', '2e-5', '--scale_factor', '0.18215', '--save_content', '--save_content_every', '10', '--use_origin_adm']' returned non-zero exit status 1.

Do I need to modify the channels=3 in line 49 of this file: [lmdb_datasets.py]? I noticed in the run script, there is --image_size 256 --f 8 --num_in_channels 4 --num_out_channels 4, so I'm curious if I should modify the lmdb_datasets.py file, or is there an issue with how I'm loading the dataset? Thanks!

ps: This segment is used to test if the code runs successfully: bash ./run_test.sh test_args/celeb256_dit.txt. Therefore, I believe my celeba256 dataset is not the issue, as it generates a sample from Celeba.

[sumorday]

is this correctly?

[sumorday]

---

[sumorday]

wget https://openaipublic.azureedge.net/glow-demo/data/celeba-tfr.tar tar -xvf celeb-tfr.tar python convert_tfrecord_to_lmdb.py --dataset=celeba --tfr_path=$DATA_DIR/celeba/celeba-tfr --lmdb_path=$DATA_DIR/celeba/celeba-lmdb --split=train