Skip to content

Actions: cheald/sd-scripts

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
19 workflow runs
19 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

Add support for positive_captions and negative_captions per data subs… Typos #19: Commit b132409 pushed by cheald
July 2, 2024 18:00 14s training-embedding-control
training-embedding-control
July 2, 2024 18:00 14s
Add missing embedding_weights utility file Typos #18: Commit 08ed0fa pushed by cheald
July 2, 2024 17:52 14s training-embedding-control
training-embedding-control
July 2, 2024 17:52 14s
Add support for positive_captions and negative_captions per data subs… Typos #17: Commit 8be734a pushed by cheald
July 2, 2024 17:51 13s training-embedding-control
training-embedding-control
July 2, 2024 17:51 13s
Perform stats collection directly on the model via forward noising pr… Typos #16: Commit 4e2a731 pushed by cheald
June 21, 2024 04:57 13s autostats
autostats
June 21, 2024 04:57 13s
Fix datatypes Typos #15: Commit 8827c5b pushed by cheald
June 18, 2024 20:33 16s autostats
autostats
June 18, 2024 20:33 16s
Support batching for autostats, add decay for the loss weight Typos #14: Commit 8b9617d pushed by cheald
June 18, 2024 20:02 15s autostats
autostats
June 18, 2024 20:02 15s
Tweaks to noise mean weighting Typos #13: Commit 3f64f55 pushed by cheald
June 15, 2024 07:52 12s autostats
autostats
June 15, 2024 07:52 12s
Add command line switches, store timesteps with the noise stats, samp… Typos #12: Commit 9bcd9b9 pushed by cheald
June 15, 2024 00:39 17s autostats
autostats
June 15, 2024 00:39 17s
Add support for dynamic_dim for lora and lora_fa Typos #11: Commit 3aa8915 pushed by cheald
June 14, 2024 05:22 12s autostats
autostats
June 14, 2024 05:22 12s
Comment out the experimental unmasked portion contributions Typos #10: Commit ca1fdf1 pushed by cheald
May 10, 2024 05:29 12s add-std-loss
add-std-loss
May 10, 2024 05:29 12s
Quick add of the latest experimental loss routine Typos #9: Commit 7c7d967 pushed by cheald
May 10, 2024 04:47 13s add-std-loss
add-std-loss
May 10, 2024 04:47 13s
Add support for --std_loss_weight, which adds a second loss objective… Typos #8: Commit 1b7efa2 pushed by cheald
April 26, 2024 06:14 12s add-std-loss
add-std-loss
April 26, 2024 06:14 12s
Remove the kl_div_loss term. Its gradients are too expensive, and it'… Typos #7: Commit a059cd9 pushed by cheald
April 24, 2024 07:57 17s kl_div_loss
kl_div_loss
April 24, 2024 07:57 17s
Remove the kl_div_loss term. Its gradients are too expensive, and it'… Typos #6: Commit b387137 pushed by cheald
April 24, 2024 07:56 18s kl_div_loss
kl_div_loss
April 24, 2024 07:56 18s
Remove the skew and kurtosis loss objectives; they hurt more than the… Typos #5: Commit 357aa44 pushed by cheald
April 24, 2024 07:21 14s kl_div_loss
kl_div_loss
April 24, 2024 07:21 14s
Use masked stat loss Typos #4: Commit 6bf5dae pushed by cheald
April 24, 2024 07:07 14s kl_div_loss
kl_div_loss
April 24, 2024 07:07 14s
Add support for std, skew, and kurtosis loss weighting. Fix kl_div so… Typos #3: Commit f2f25c9 pushed by cheald
April 22, 2024 20:59 15s kl_div_loss
kl_div_loss
April 22, 2024 20:59 15s
Add support for kl_div_loss_weight Typos #2: Commit 6036cc5 pushed by cheald
April 22, 2024 19:01 13s kl_div_loss
kl_div_loss
April 22, 2024 19:01 13s
RReLU for the multires schedule seems to work better. Typos #1: Commit 19bc066 pushed by cheald
April 7, 2024 21:33 16s learnable-hyperparams
learnable-hyperparams
April 7, 2024 21:33 16s