Commit History

Try to change output to stream instead of wait for completion
82af3b6
verified

andreska commited on

Removed chat history
fd25069
verified

andreska commited on

Removed trying to remove text from input box
9428c8a
verified

andreska commited on

Revert back to qwen.
d218ef0
verified

andreska commited on

Try earlier llama model
4e0c6dc
verified

andreska commited on

fixed syntax error in previous commit
31daeee
verified

andreska commited on

Switch model to llama
8932ece
verified

andreska commited on

Update app.py
63728fb
verified

andreska commited on

Changed to mistral model
9391655
verified

andreska commited on

Changed model and alterned default context to remove location.
100673e
verified

andreska commited on

Try fix problem with clearing user input box
c9b44c4
verified

andreska commited on

Clear input box after use
9898aa3
verified

andreska commited on

Update app.py
2249329
verified

andreska commited on

Rearranged help checkbox after ask button
30f53de
verified

andreska commited on

Rearranged columns below input box
187d5dc
verified

andreska commited on

Try to realign items
634e250
verified

andreska commited on

Try to fix problems with method not being initialized in time
a7a6cbd
verified

andreska commited on

Try to realign elements to take less vertical space.
11a6534
verified

andreska commited on

Remove header to make it fit better inside Adrega
f5e81c3
verified

andreska commited on

Reduced max tokens further to make sure it accepts a larger question
ca90665
verified

andreska commited on

Changed max tokens to take Help into consideration for max total tokens
33c3d9e
verified

andreska commited on

Change token limit to Qwen max
5616271
verified

andreska commited on

Changed output text to always display last question and answer at top
5b8c9a0
verified

andreska commited on

Updated context for non-help related questions. Try to add answers in scrollable div, with better overview of question and answer.
52ef193
verified

andreska commited on

Try to replace textarea with markdown to display nice html
c28c28f
verified

andreska commited on

Updated to use session state variable instead of just variable for include_context
c119414
verified

andreska commited on

Removed code for include_context. It should always have a value
07d6d5e
verified

andreska commited on

Test update with invalid information to test context with and without checkbox
9e74383
verified

andreska commited on

Write text to screen, to see if the formatting is better
42fd51d
verified

andreska commited on

Reverted trying to not load dataset, as it causes problem with requests
e7169f6
verified

andreska commited on

Rearranged the submit button.
b0c57d4
verified

andreska commited on

Prevent loading dataset if Help checkbox is not checked
9dfb04d
verified

andreska commited on

Fix session_state try2
cbd1365
verified

andreska commited on

Initialize include_context before usage
442c909
verified

andreska commited on

Fixed minor syntax error
3a24c45
verified

andreska commited on

Moved checkbox up, added possibility to press enter to submit
d943ab1
verified

andreska commited on

Fix problem with conversation not being initialized
ec395fa
verified

andreska commited on

Keep and display chat history
fed0941
verified

andreska commited on

Update to yellow background #3
165ae3e
verified

andreska commited on

Make background yellow try2
811ba5a
verified

andreska commited on

Test setting a specific background color
2e57e8b
verified

andreska commited on

Fixed syntax error
7ed9b7b
verified

andreska commited on

Added button to turn help on or off
88e2d62
verified

andreska commited on

Removed chunks, as it didn't work anyway
82909b4
verified

andreska commited on

Try to split query into chunks to prevent overloading model
2fc80c5
verified

andreska commited on

Test qwen 72B
7498c16
verified

andreska commited on

Test what happens if switch to cold inference model
85fad1d
verified

andreska commited on

One more try to get inference llama 3.2 model working
14ef7cc
verified

andreska commited on

Try to switch to llama model to see if it handles more tokens
c06b155
verified

andreska commited on

Switched from using specific .txt file to referencing whole dataset
d1b814f
verified

andreska commited on