This website requires JavaScript.
e98048093c
Fix typo in js script, add printing to console for preview node, remove deprecated config option
main
Zuellni
2024-12-06 15:22:11 +01:00
63c7394d2c
Merge pull request #27 from ricklove/main
Zuellni
2024-09-22 08:44:30 +02:00
fbbe8bc910
FIX serious performance issue on every prompt
Rick Love
2024-09-21 21:29:02 -05:00
bef6a0d3cb
Fix default values not showing in preview
Zuellni
2024-08-23 22:21:47 +02:00
97ec4b551c
Update README.md
Zuellni
2024-08-07 20:06:26 +02:00
9f9d25dd49
Rename the convert node, some fixes and cleanup
Zuellni
2024-08-04 12:57:22 +02:00
4d58d29f8e
Fix attention size typo
Zuellni
2024-07-31 01:25:02 +02:00
5c63d0609f
Remove the hidden inputs since they don't seem to be working anyway (no prompt in workflow), touch-up the convert node a bit
Zuellni
2024-07-29 19:22:39 +02:00
1b02390317
Update workflow
Zuellni
2024-07-29 16:53:02 +02:00
42d621d4fd
Merge pull request #25 from Zuellni/dev
Zuellni
2024-07-29 16:52:19 +02:00
66a538e316
Update README.md
Zuellni
2024-07-29 15:51:34 +02:00
2db2d93844
Added a convert node, a string node and fixed some js (probably)
Zuellni
2024-07-29 15:41:49 +02:00
2acc730ad3
Update README.md
Zuellni
2024-07-28 21:31:21 +02:00
7ae7ef9a24
Split some nodes, added a format node that uses tokenizer config template
Zuellni
2024-07-28 17:33:07 +02:00
8311351697
Cache should be a multiple of 256 for paged attention
Zuellni
2024-06-26 16:01:08 +02:00
d5feb5a293
Clean up cache
Zuellni
2024-06-26 13:07:37 +02:00
0bd03f4dae
Clean up imports
Zuellni
2024-06-24 12:33:30 +02:00
cc6c50e2ff
Merge pull request #23 from Zuellni/dev
Zuellni
2024-06-14 18:52:24 +02:00
5a90464b6c
Update README.md
Zuellni
2024-06-14 18:48:05 +02:00
51728c0102
Update requirements.txt
Zuellni
2024-06-14 13:28:48 +02:00
cac9d1fdf4
Update README.md
Zuellni
2024-06-14 13:24:38 +02:00
f54093078e
Clean up README
Zuellni
2024-06-14 13:21:42 +02:00
add411a7c3
Merge remote-tracking branch 'origin/main' into dev
Zuellni
2024-06-14 13:15:47 +02:00
49ace07a0e
Update README.md
Zuellni
2024-06-14 12:51:36 +02:00
4754f78c70
Add option to enable/disable flash attention and fast tensors
Zuellni
2024-06-14 11:50:22 +02:00
f988703614
Add custom stop conditions
Zuellni
2024-06-14 11:33:10 +02:00
e91a1ce2c6
Switch to dynamic generator
Zuellni
2024-06-14 10:57:28 +02:00
e5e13c40ed
Update README.md
Zuellni
2024-06-05 09:53:12 +02:00
d7413b2ebc
Fix vram usage with higher context
Zuellni
2024-05-17 22:55:50 +02:00
c005e3c637
Update config init
Zuellni
2024-05-04 09:13:01 +02:00
4422f0b0d0
Update readme
Zuellni
2024-04-22 18:30:33 +02:00
4633de563a
Update workflow with llama-3 format
Zuellni
2024-04-22 16:18:18 +02:00
87810044ea
Change default cache bits to 16, reorganize imports, update readme
Zuellni
2024-04-22 15:50:50 +02:00
8c2f17829a
Clarify instructions
Zuellni
2024-04-08 03:03:53 +02:00
c5765ffb8c
Update with a cascade workflow
Zuellni
2024-04-07 20:19:55 +02:00
fb7b7ff683
Update readme
Zuellni
2024-04-07 20:00:44 +02:00
fb5b6d95ef
Update link
Zuellni
2024-04-07 19:55:34 +02:00
d2543cbd59
Simplify instructions
Zuellni
2024-04-07 19:52:53 +02:00
baed69aba5
Bump minimum exllamav2 version, add autosplit and q4 cache, some reformatting
Zuellni
2024-04-07 19:35:32 +02:00
d2b4d7a796
More sanity checks for model unloading, bump exllama
Zuellni
2024-02-26 13:24:03 +01:00
f7a108315e
Update readme
Zuellni
2024-02-21 23:06:35 +01:00
217304e0ef
Only unload other models if unload is checked
Zuellni
2024-02-17 16:53:00 +01:00
8c4b9d058f
Update workflow
Zuellni
2024-02-17 15:14:46 +01:00
861aa6fc51
Update readme
Zuellni
2024-02-17 15:13:53 +01:00
8b7b957b3e
Add gc again
Zuellni
2024-02-17 14:53:54 +01:00
e1c8e61291
Unload other models first before loading llms, rename some params, update requirements
Zuellni
2024-02-17 14:40:46 +01:00
ca41704f38
Update readme
Zuellni
2024-02-14 20:26:46 +01:00
bf64bcc0e4
Split requirements to file, neutralize samplers by default
Zuellni
2024-02-14 20:23:27 +01:00
9ffbacf080
Update requirements
Zuellni
2024-02-12 16:38:52 +01:00
6cc728d7cb
Update readme
Zuellni
2024-02-09 16:13:30 +01:00
ec6443816e
Update readme
Zuellni
2024-02-06 14:51:02 +01:00
6831b951c5
Update readme
Zuellni
2024-02-03 15:53:52 +01:00
54bedbed53
Update workflow
Zuellni
2024-02-03 15:52:58 +01:00
88a826fa30
Add top_a sampling
Zuellni
2024-02-03 15:14:38 +01:00
18c7854d76
Add missing import, look for paths recursively
Zuellni
2024-02-03 15:13:30 +01:00
261a73a119
Merge pull request #14 from ScottNealon/patch-1
Zuellni
2024-02-03 15:11:39 +01:00
348aeea258
Parse all possible folders mapped to "llm"
Scott Nealon
2024-02-02 22:12:50 -05:00
73e6a0b0e4
Bump exllama
Zuellni
2024-01-26 15:39:58 +01:00
331b594f00
Bump flash-attn
Zuellni
2023-12-27 13:13:16 +01:00
de401d308d
Bump exllama
Zuellni
2023-12-21 19:51:05 +01:00
04f0c87e31
Fix error when model isn't fully unloaded and bump requirements
Zuellni
2023-12-02 23:27:41 +01:00
eca815e45a
Fix linux requirements
Zuellni
2023-11-25 20:16:56 +01:00
15a224ccbb
Even better table
Zuellni
2023-11-25 14:55:51 +01:00
485e2c4c51
Prettify table
Zuellni
2023-11-25 13:48:56 +01:00
98ec700a5c
Remove the tensor numel check, should be redundant
Zuellni
2023-11-25 13:21:19 +01:00
69b70b1b41
Clarify gpu split should be in GB, not MB
Zuellni
2023-11-25 13:09:40 +01:00
1b9803ae34
Update wheels
Zuellni
2023-11-25 12:15:44 +01:00
677bc6c388
Update readme
Zuellni
2023-11-25 12:10:53 +01:00
1bfa440ef1
Add progress bar to model loader
Zuellni
2023-11-24 18:49:50 +01:00
09de8351b9
Update flash attention wheel
Zuellni
2023-11-24 18:09:04 +01:00
3e2c6d6759
Update link
Zuellni
2023-11-24 17:42:03 +01:00
125cdf51cb
Prettify readme
Zuellni
2023-11-24 14:58:06 +01:00
8ca924691f
Update readme
Zuellni
2023-11-23 22:47:49 +01:00
68f563eb4b
Add field descriptions
Zuellni
2023-11-23 22:44:15 +01:00
d881503d4e
Add some info about model downloading
Zuellni
2023-11-23 22:15:10 +01:00
9cbbcb1e72
Update readme
Zuellni
2023-11-23 20:22:45 +01:00
4d65f3c4db
Update readme
Zuellni
2023-11-23 20:12:02 +01:00
2f1affecb4
Simplify model loader Load from custom 'llm' dir specified in 'extra_model_paths.yaml' or from 'models/llm'
Zuellni
2023-11-23 20:04:14 +01:00
8b75af13b3
Update workflow
Zuellni
2023-11-22 16:15:23 +01:00
72a44cbcca
Fix preview output
Zuellni
2023-11-22 16:14:28 +01:00
51bb1bcf84
Revert autosplit
Zuellni
2023-11-22 16:04:01 +01:00
996fdd012d
Update readme
Zuellni
2023-11-22 14:28:23 +01:00
51ff962f14
Add autosplit, temperature_last
Zuellni
2023-11-22 14:15:00 +01:00
957fe3e6a5
Bump exllamav2
Zuellni
2023-11-22 12:38:08 +01:00
05a30fce3f
Remove unnecessary line in preview
Zuellni
2023-11-21 01:56:57 +01:00
b6cd5f0de8
Update workflow
Zuellni
2023-11-20 23:37:17 +01:00
a33a1ac395
Add gpu split, add 8bit cache toggle, add min_p, encode specal tokens, update requirements versions
Zuellni
2023-11-20 21:44:52 +01:00
9d20724d02
Update readme
Zuellni
2023-11-07 14:04:16 +01:00
47e38dabbc
Update workflow
Zuellni
2023-11-07 13:04:37 +01:00
cb7f195068
Merge pull request #9 from Zuellni/0.0.7
Zuellni
2023-11-07 13:03:56 +01:00
7e97b26af9
Fix preview size
Zuellni
2023-11-07 12:55:39 +01:00
f48ebd1c68
Some minor changes
Zuellni
2023-11-07 12:46:50 +01:00
55871388b6
Move for clarity
Zuellni
2023-10-27 15:54:24 +02:00
dcaade09c2
Remove loras for now, there seems to be a memory leak and idk how to fix it Remove allowed strings, they don't seem very useful
Zuellni
2023-10-27 15:51:15 +02:00
2b3ddde76b
Update README.md
Zuellni
2023-10-25 19:59:58 +02:00
0b94a076ab
Update README.md
Zuellni
2023-10-25 19:57:16 +02:00
d2c554b69d
Add loras, 8bit cache, fix random seed, unloading
Zuellni
2023-10-25 19:53:44 +02:00
420b1fbd2d
Update workflow
Zuellni
2023-10-10 22:21:58 +02:00
ebb32253a9
Allow only some tokens in output, add condition node save prompt without having to preview, auto set seq len and tokens if 0 and some other stuff I already forgot, it should all work, but likely won't
Zuellni
2023-10-10 22:16:12 +02:00
f16a2ab515
Update readme link
Zuellni
2023-10-08 20:09:23 +02:00