Add vqvae-celeba-16-512

Files changed (7) hide show

vqvae-celeba-16-512/ckpt/step0499999/meta.pt ADDED Viewed

+version https://git-lfs.github.com/spec/v1
+oid sha256:df99aa5f4769833bb82ff2e829751b4d07f62f0674714325cb3fc3ab673c2dc8
+size 852

vqvae-celeba-16-512/ckpt/step0499999/model.pt ADDED Viewed

+version https://git-lfs.github.com/spec/v1
+oid sha256:f634999092a9d52bff36c51ab7b5d57008a60bcde7da71265e5ba24b2343cbca
+size 19527185

vqvae-celeba-16-512/ckpt/step0499999/optimizer.pt ADDED Viewed

+version https://git-lfs.github.com/spec/v1
+oid sha256:e32f938547d031242c796378cf3393f400d3e50439c1ca11faeaf308619b5f69
+size 39058707

vqvae-celeba-16-512/config-2024-07-21-13-50-42.yaml ADDED Viewed

+seed: 8888
+data:
+  name: celeba
+  root: ~/data/CelebA/
+  img_size: 64
+dataloader:
+  num_workers: 4
+  pin_memory: true
+  prefetch_factor: 2
+encoder:
+  target: models.autoencoder.simple_cnn.Encoder
+  params:
+    img_channels: 3
+    hidden_dim: 256
+    n_resblocks: 2
+decoder:
+  target: models.autoencoder.simple_cnn.Decoder
+  params:
+    img_channels: 3
+    hidden_dim: 256
+    n_resblocks: 2
+quantizer:
+  target: models.quantizer.VectorQuantizer
+  params:
+    codebook_num: 512
+    codebook_dim: 16
+train:
+  n_steps: 500000
+  batch_size: 256
+  coef_commit: 0.25
+  print_freq: 100
+  sample_freq: 10000
+  save_freq: 10000
+  optim:
+    target: torch.optim.Adam
+    params:
+      lr: 0.0004

vqvae-celeba-16-512/output-2024-07-21-13-50-42.log ADDED Viewed

The diff for this file is too large to render. See raw diff

vqvae-celeba-16-512/samples.zip ADDED Viewed

+version https://git-lfs.github.com/spec/v1
+oid sha256:d350df30ba380c56cb873c205afa869afee6dcea22dd1db5b6fe4714ec3d7f8b
+size 23650943

vqvae-celeba-16-512/tensorboard/events.out.tfevents.1721541042.amax-ESC8000A-E11.3195644.0 ADDED Viewed

+version https://git-lfs.github.com/spec/v1
+oid sha256:3bc15e98b638212c5fa05736d64065024a3c3ab6cbc7dc1e67f26fc550917264
+size 133917518