Fampai
39541d7725
Fixes race condition in training when VAE is unloaded
...
set_current_image can attempt to use the VAE when it is unloaded to
the CPU while training
2022-11-04 04:50:22 -04:00
AUTOMATIC1111
4918eb6ce4
Merge branch 'master' into hn-activation
2022-11-04 09:02:15 +03:00
AUTOMATIC1111
17a2076f72
Merge pull request #3928 from R-N/validate-before-load
...
Optimize training a little
2022-10-30 09:51:36 +03:00
Muhammad Rizqi Nur
3d58510f21
Fix dataset still being loaded even when training will be skipped
2022-10-30 00:54:59 +07:00
Muhammad Rizqi Nur
a07f054c86
Add missing info on hypernetwork/embedding model log
...
Mentioned here: https://github.com/AUTOMATIC1111/stable-diffusion-webui/discussions/1528#discussioncomment-3991513
Also group the saving into one
2022-10-30 00:49:29 +07:00
Muhammad Rizqi Nur
ab05a74ead
Revert "Add cleanup after training"
...
This reverts commit 3ce2bfdf95
2022-10-30 00:32:02 +07:00
Muhammad Rizqi Nur
3ce2bfdf95
Add cleanup after training
2022-10-29 19:43:21 +07:00
Muhammad Rizqi Nur
ab27c111d0
Add input validations before loading dataset for training
2022-10-29 18:09:17 +07:00
timntorres
e98f72be33
Merge branch 'AUTOMATIC1111:master' into 3825-save-hypernet-strength-to-info
2022-10-29 00:31:23 -07:00
AUTOMATIC1111
810e6a407d
Merge pull request #3858 from R-N/log-csv
...
Fix log off by 1 #3847
2022-10-29 07:55:20 +03:00
AUTOMATIC1111
d3b4b9d7ec
Merge pull request #3717 from benkyoujouzu/master
...
Add missing support for linear activation in hypernetwork
2022-10-29 07:30:14 +03:00
AngelBottomless
f361e804eb
Re enable linear
2022-10-29 08:36:50 +09:00
Muhammad Rizqi Nur
9ceef81f77
Fix log off by 1
2022-10-28 20:48:08 +07:00
timntorres
db5a354c48
Always ignore "None.pt" in the hypernet directory.
2022-10-28 01:41:57 -07:00
benkyoujouzu
b2a8b263b2
Add missing support for linear activation in hypernetwork
2022-10-28 12:54:59 +08:00
AngelBottomless
462e6ba667
Disable unavailable or duplicate options
2022-10-27 15:40:24 +09:00
AngelBottomless
029d7c7543
Revert unresolved changes in Bias initialization
...
it should be zeros_ or parameterized in future properly.
2022-10-27 14:44:53 +09:00
guaneec
cc56df996e
Fix dropout logic
2022-10-27 14:38:21 +09:00
AngelBottomless
85fcccc105
Squashed commit of fixing dropout silently
...
fix dropouts for future hypernetworks
add kwargs for Hypernetwork class
hypernet UI for gradio input
add recommended options
remove as options
revert adding options in ui
2022-10-27 14:38:21 +09:00
guaneec
b6a8bb123b
Fix merge
2022-10-26 15:15:19 +08:00
timntorres
a524d137d0
patch bug (SeverianVoid's comment on 5245c7a)
2022-10-26 10:12:46 +03:00
guaneec
91bb35b1e6
Merge fix
2022-10-26 15:00:03 +08:00
guaneec
649d79a8ec
Merge branch 'master' into hn-activation
2022-10-26 14:58:04 +08:00
guaneec
877d94f97c
Back compatibility
2022-10-26 14:50:58 +08:00
AngelBottomless
7207e3bf49
remove duplicate keys and lowercase
2022-10-26 09:17:01 +03:00
AngelBottomless
de096d0ce7
Weight initialization and More activation func
...
add weight init
add weight init option in create_hypernetwork
fstringify hypernet info
save weight initialization info for further debugging
fill bias with zero for He/Xavier
initialize LayerNorm with Normal
fix loading weight_init
2022-10-26 09:17:01 +03:00
guaneec
c702d4d0df
Fix off-by-one
2022-10-26 13:43:04 +08:00
guaneec
2f4c91894d
Remove activation from final layer of HNs
2022-10-26 12:10:30 +08:00
AngelBottomless
e9a410b535
check length for variance
2022-10-24 09:07:39 +03:00
AngelBottomless
0d2e1dac40
convert deque -> list
...
I don't feel this being efficient
2022-10-24 09:07:39 +03:00
AngelBottomless
348f89c8d4
statistics for pbar
2022-10-24 09:07:39 +03:00
AngelBottomless
40b56c9289
cleanup some code
2022-10-24 09:07:39 +03:00
AngelBottomless
b297cc3324
Hypernetworks - fix KeyError in statistics caching
...
Statistics logging has changed to {filename : list[losses]}, so it has to use loss_info[key].pop()
2022-10-24 09:07:39 +03:00
DepFA
1fbfc052eb
Update hypernetwork.py
2022-10-23 08:34:33 +03:00
AngelBottomless
48dbf99e84
Allow tracking real-time loss
...
Someone had 6000 images in their dataset, and it was shown as 0, which was confusing.
This will allow tracking real time dataset-average loss for registered objects.
2022-10-22 22:24:19 +03:00
AngelBottomless
24694e5983
Update hypernetwork.py
2022-10-22 20:25:32 +03:00
discus0434
6a4fa73a38
small fix
2022-10-22 13:44:39 +00:00
discus0434
97749b7c7d
Merge branch 'AUTOMATIC1111:master' into master
2022-10-22 22:00:59 +09:00
discus0434
7912acef72
small fix
2022-10-22 13:00:44 +00:00
discus0434
fccba4729d
add an option to avoid dying relu
2022-10-22 12:02:41 +00:00
AUTOMATIC
7fd90128eb
added a guard for hypernet training that will stop early if weights are getting no gradients
2022-10-22 14:48:43 +03:00
discus0434
dcb45dfecf
Merge branch 'master' of upstream
2022-10-22 11:14:46 +00:00
discus0434
0e8ca8e7af
add dropout
2022-10-22 11:07:00 +00:00
timntorres
272fa527bb
Remove unused variable.
2022-10-21 16:52:24 +03:00
timntorres
19818f023c
Match hypernet name with filename in all cases.
2022-10-21 16:52:24 +03:00
timntorres
51e3dc9cca
Sanitize hypernet name input.
2022-10-21 16:52:24 +03:00
AUTOMATIC
03a1e288c4
turns out LayerNorm also has weight and bias and needs to be pre-multiplied and trained for hypernets
2022-10-21 10:13:24 +03:00
AUTOMATIC1111
0c5522ea21
Merge branch 'master' into training-help-text
2022-10-21 09:57:55 +03:00
timntorres
4ff274e1e3
Revise comments.
2022-10-21 09:55:00 +03:00
timntorres
5245c7a493
Issue #2921-Give PNG info to Hypernet previews.
2022-10-21 09:55:00 +03:00
