somebody
b93c339145
Model: Lazyload backends
2023-03-13 20:29:29 -05:00
somebody
3adc67c7a4
RWKV: Move import right before usage
...
So we don't needlessly compile the cuda kernel
2023-03-13 19:37:45 -05:00
somebody
b10b201701
Model: Add basic RWKV implementation
2023-03-13 19:34:38 -05:00
somebody
60793eb121
Make modellist easier to work with
2023-03-13 15:40:24 -05:00
jojorne
c87ef60db1
revert more unrelated code
2023-03-11 13:48:41 -03:00
jojorne
423c022648
Fix save loading between v1 and v2 to v3 with wi features
2023-03-11 13:14:46 -03:00
somebody
38c4edac40
Model: Fix eos/bos padding issue
...
Weird config None assignments
2023-03-10 20:01:01 -06:00
SillyLossy
3ac3e41dd8
Fix setting softprompt via V1 API
2023-03-10 20:00:39 -06:00
somebody
a472bdf6c3
Cleanup
2023-03-10 20:00:39 -06:00
whjms
b5000b35d9
fix 404 handler
2023-03-09 18:17:30 -05:00
whjms
ad2c2b6722
move aetherroom import to separate module
2023-03-09 18:17:30 -05:00
SillyLossy
0d1771501f
Fix setting softprompt via V1 API
2023-03-09 13:15:50 +02:00
somebody
dc3dd38c4e
Merge branch 'united' of https://github.com/henk717/KoboldAI into model-structure-and-maybe-rwkv
2023-03-07 17:02:28 -06:00
somebody
3b05359e4b
Model: Refuse to serve certain models over the API
2023-03-07 16:54:08 -06:00
henk717
b307cdff57
Merge pull request #301 from jojorne/jojorne-patch-use-str-for-wi-keys
...
Access WI uid keys with str
2023-03-06 21:15:44 +01:00
Henk
099c818c06
Experimental unbanning of <>
2023-03-06 20:56:15 +01:00
Henk
b7f30a41c5
LLama tokenizer workaround
2023-03-06 19:32:28 +01:00
jojorne
4f17bf0c87
Access WI uid keys with str.
2023-03-05 13:22:33 -03:00
Henk
9ee6adb35c
Enable BMS API again under a new name
2023-03-04 19:02:19 -06:00
Henk
55349d84e1
Remote no longer unblocks the port by default
2023-03-04 19:02:15 -06:00
Henk
59e1f74f7c
Enable BMS API again under a new name
2023-03-04 14:17:47 +01:00
henk717
aa558da49c
Merge pull request #297 from aibosan/united-singleline-stopper
...
Created SinglelineStopper...
2023-03-02 20:51:54 +01:00
Henk
f03b538cab
Remote no longer unblocks the port by default
2023-03-02 17:43:23 +01:00
Ondřej Benda
aa124b65db
Fix: incorrect newline evaluation
2023-03-02 08:22:31 +01:00
somebody
54cecd4d5d
Model: And another refactor
2023-03-01 19:16:35 -06:00
Ondřej Benda
0ba7ac96d3
Created SinglelineStopper, which interrupts token generation when a newline is reached if singleline mode is enabled
2023-03-02 00:04:17 +01:00
Henk
93bf023bd7
Use our own horde URL
2023-03-01 17:54:39 +01:00
somebody
225dcf1a0a
Model: Documentation part 1
2023-02-28 20:10:09 -06:00
somebody
ef1155291f
Model: TPU Fixes
2023-02-28 18:05:34 -06:00
somebody
bd3bbdaad8
Model: More tpu fixes
2023-02-27 19:29:03 -06:00
somebody
b99c16f562
Merge branch 'united' of https://github.com/henk717/KoboldAI into model-structure-and-maybe-rwkv
2023-02-26 13:06:30 -06:00
somebody
56bcd32f6d
Model: Fix ReadOnly
...
whoops
2023-02-26 12:40:20 -06:00
somebody
a73804ca1e
Accelerate: Remove HAS_ACCELERATE
...
Accelerate has been a dependency for a while, and as such we probably
shouldn't be lugging around code that assumes it isn't present.
2023-02-26 12:18:06 -06:00
somebody
5e3b0062ee
Model: Tokenizer fix
2023-02-26 12:17:49 -06:00
somebody
8d49b5cce1
Model: More TPU stuff
2023-02-26 12:09:26 -06:00
Henk
cdfc7326e6
Horde API Setting Consolidation
2023-02-26 17:59:36 +01:00
henk717
642d173579
Merge pull request #294 from db0/kaimerge
...
Changes to work with the merged hordes
2023-02-26 16:00:48 +01:00
somebody
ffe4f25349
Model: Work on stoppers and stuff
2023-02-25 17:12:16 -06:00
somebody
6b4905de30
Model: Port rest of models over
...
Generation's still broke but it's a start
2023-02-25 16:05:56 -06:00
Henk
165c7219eb
Frequency Penalty for OAI
2023-02-25 22:15:05 +01:00
somebody
f8c4158ebc
Model: Successful load implementation
...
The goal of this series of commits is to have an implementation-agnostic
interface for models, thus being less reliant on HF Transformers for model
support. A model object will have a method for generation, a list of callbacks
to be run on every token generation, a list of samplers that will modify
probabilities, etc. Basically anything HF can do should be easily
implementable with the new interface :^)
Currently I've tested the loading of pre-downloaded models with
breakmodel between GPUs and that works, though essentially no testing
has been done in the larger scheme of things. Currently this is about
the only supported configuration, and generation isn't very functional.
2023-02-24 21:41:44 -06:00
Divided by Zer0
6426e3ca24
changes
2023-02-23 18:34:46 +01:00
Divided by Zer0
2de9672b95
attempt1
2023-02-23 18:27:11 +01:00
Divided by Zer0
d459bdb1c0
adjust
2023-02-22 23:12:34 +01:00
jojorne
d3bedfcbda
Include koboldai_vars.save_story().
2023-02-22 15:42:56 -03:00
jojorne
d6c9f5f1f5
Save/Load Story with UTF-8 encoding.
2023-02-22 14:40:42 -03:00
Henk
9e6a5db745
UI1 Botname
2023-02-19 16:22:26 +01:00
henk717
93f313d6e3
Merge pull request #291 from pi6am/fix/save-as
...
Fix an exception using Save As from the classic UI
2023-02-19 13:21:10 +01:00
Llama
117f0659c3
Fix exception using Save As from the classic UI
...
The `saveas` method was modified to take a data dict but one of the
else blocks still referred to the previous `name` parameter. Assign
to `name` to fix the `NameError: name 'name' is not defined` exception.
2023-02-18 23:41:17 -08:00
Henk
cd566caf20
Revision Fixes (Removes the workaround)
2023-02-19 00:51:50 +01:00
