The Initial commit for Chat Mode, the nickname part of the UI is missing other than that it should be fully functional. To use Chat Mode effectively you first input a small dialogue (Can be around 6 lines 3 of your own inputs and 3 of the character) formatted as Name : it will then automate the actions needed to chat properly. During this mode single line mode is forced on, and Trim Incomplete Sentences is forced off.
When the prompt is deleted by the user, the topmost remaining chunk of
the story that has at most one non-whitespace character is now made the
new prompt chunk.
Also fixed issues in Chromium-based browsers (desktop and Android) where
selecting all text in the story, typing some new text to replace the
entire story and then defocusing causes the editor to break.
* `print()` and `warn()` now work correctly with `nil` arguments
* Typo: `gpt-neo-1.3M` has been corrected to `gpt-neo-1.3B`
* Regeneration is no longer triggered when writing to `keysecondary` of
a non-selective key
* Handle `genamt` changes in generation modifier properly
* Writing to `kobold.settings.numseqs` from a generation modifier no
longer affects
* Formatting options in `kobold.settings` have been fixed
* Added aliases for setting names
* Fix behaviour of editing story chunks from a generation modifier
* Warnings are now yellow instead of red
* kobold.logits is now the raw logits prior to being filtered, like
the documentation says, rather than after being filtered
* Some erroneous comments and error messages have been corrected
* These parts of the API have now been implemented properly:
* `compute_context()` methods
* `kobold.authorsnote`
* `kobold.restart_generation()`
This makes it so that SocketIO uses long polling to set up the
connection before switching to websocket, instead of immediately using
websocket.
This seems to resolve issues where the browser sometimes can't connect
to the websocket server until the window has been open for a minute.
Allow people to enter a prompt without generating anything by the AI. Combined with the always add prompt this is a very useful feature that allows people to write world information first, and then do a specific action. This mimics the behavior previously seen in AI Dungeon forks where it prompts for world information and then asks an action and can be particularly useful for people who want the prompt to always be part of the generation.
This fixes a problem where WI entries/folders are sometimes able to be
dragged into places they shouldn't be. Steps to reproduce:
1. Start a blank story
2. Refresh the browser
3. Open the W Info screen
4. Add a world info folder
5. Add a world info entry into that folder
6. Drag that world info entry
When Firefox 93.0 was released, they broke the ability to edit text
across multiple chunks or across multiple paragraphs. If you tried,
nothing would happen.
Also, we are no longer using Mutation Observers to detect when a chunk
is modified. We are now using the beforeinput event.
Adds Single Line mode, optimized for things like chatbot testing and other cases where you want to have control over what happens after a paragraph.
This can also be used as a foundation for a chatbot optimized interface mode.
Today I learned that the editor only works properly when the last
<chunk> tag has a <br> inside it at the end. This last <br> is
invisible and is automatically created by all major browsers when you
use the enter key to type a newline at the end of a story to "prevent
the element from collapsing". When there's more than one <br> at the
end of the last <chunk>, only the last of those <br>s is invisible, so
if you have three <br>s, they are rendered as two newlines. This only
applies to the last <chunk>, so if the second last <chunk> has three
<br>s at the end, they are still rendered as three newlines. Since
the browser is really insistent on doing this, this commit mostly deals
with dynamically creating and deleting <br> tags at the ends of <chunk>
tags as needed to provide a consistent experience, and making sure
that all <br> tags actually go inside of <chunk> tags to prevent
breaking the editor. The latter behaviour was exhibited by Chrome and
caused a bug when you added a newline at the end of your story using
the editor.
For some reason the original way only works in Safari after pressing the
refresh button. It did not work if you typed the URL into the address
bar in Safari without refreshing afterwards.
I don't know why Firefox is so weird but we need to do this or else
some chunks don't update properly when you edit multiple chunks at the
same time.
One example of when this happened is when you have a story with at least
one chunk other than the prompt. Then, if you select the entire story
except for the first few characters of the prompt and then delete the
selected characters, and then defocus the story to save your changes,
the last chunk in the story will not register as having been deleted,
which you can verify if you refresh the page.
Another example. If your story has a chunk with no trailing newlines
followed by a chunk with exactly two leading newlines, if you delete
the first newline from the latter chunk and then defocus to save your
edits, the newline will be there again when you refresh the page.
So that if the animation is triggered multiple times you don't have to
wait for all the animations to be performed one after the other before
you can finally manually scroll.
For stories that are long enough for the scroll bar to appear on the
screen, Firefox on desktop would originally only allow you to start
editing if you click on the actual text, i.e. you couldn't click on the
blank part of a line. This behaviour is now fixed.
It wouldn't trigger any events originally when you click on parts of the
story text area that didn't contain any text, e.g. on a blank line or on
the blank part of a line to the right of the actual text.
* Back and Retry buttons no longer pop a story chunk while in the
"Select sequence to keep" menu
* No longer freezes if you retry with no story chunks beyond the initial
prompt chunk
* When "Always Add Prompt" is on, allow Retry even if the prompt is the
only chunk in the story
* Added error messages for Back and Retry buttons
Because we don't need them anymore.
This, combined with the chunk optimization, will allow you to
seamlessly continue editing a different chunk after you've finished
editing your current one; e.g. if you edit a chunk and use arrow keys,
backspace or the mouse to select a different chunk, you can immediately
start editing your new chunk without it defocusing.
This patch really has killed two birds with one stone, eh?
Thanks, Javalar!
This happens when you have a chunk that has no trailing
newlines followed by a chunk with a single newline at the start.
Moving between those chunks would cause the caret to jump to the wrong
position sometimes.
If, with Allow Editing on, you change a chunk and then click directly on
the "Allow Editing" switch again to turn it off without defocusing the
chunk in-between (by clicking outside the chunk, using enter/escape,
etc), all story chunks would still be editable.
That behaviour has now been fixed.
First step towards native Colab support, built in Cloudflare tunnels easily allows players to play KoboldAI on another device. This mode also removes buttons that would get you stuck if you have no local PC access.
Added ability to import World Info files from AI Dungeon.
Added slider for setting World Info scan depth.
Added toggle to control whether prompt is submitted each action.
Added 'Read Only' mode with no AI to startup.
Fixed GPU/CPU choice prompt appearing when GPU isn't an option.
Added error handling to generator calls for CUDA OOM message
Added generator parameter to only return new text
Added in-browser Save/Load/New Story controls
(Force a full refresh in your browser!)
Fixed adding InferKit API key if client.settings already exists
Added cmd calls to bat files so they'll stay open on error
Wait animation now hidden on start state/restart
Help text added to settings items.
Settings now saved to client file when changed.
Separated transformers settings and InferKit settings.
Reorganized model select list.
ebolam
henk717
Gnome Ann
Javalar
KoboldAI Dev
KoboldAI