turboderp
ca4f36feb2
Merge pull request #75 from LlamaEnjoyer/model_search_and_add_tweaks
...
feat: Improve model search and add debouncing
2025-02-05 23:50:35 +01:00
Llama Enjoyer
531c688eeb
Make DRY sampling affected by the 'App defaults' button
2025-02-05 21:24:22 +01:00
Llama Enjoyer
5e669e491d
feat: Improve model search and add debouncing
...
- Refined model search box appearance:
- Set width and adjusted padding.
- Added space for clear button.
- Added debouncing to model search input for performance optimization.
- Automatically set model name based on directory when creating new models.
- Added general-purpose debounce utility function.
2025-02-05 11:02:27 +01:00
turboderp
6906b144ee
Adjust layout
2025-02-04 22:01:47 +01:00
turboderp
ce57e81982
Add DRY settings in chat mode
2025-02-04 22:01:10 +01:00
turboderp
42e4e871e0
Merge remote-tracking branch 'LlamaEnjoyer/read_sampling_params_from_model_config'
2025-02-04 21:23:48 +01:00
turboderp
dd48d472f1
Handle escape key in model search box
2025-02-04 21:20:51 +01:00
Llama Enjoyer
dfe827b864
minor css cleanup
2025-02-04 13:51:08 +01:00
Llama Enjoyer
8a08bd5580
minor css cleanup
2025-02-04 13:39:04 +01:00
Llama Enjoyer
664fb57fa2
fix: don't show the token counter if no model loaded
2025-02-02 22:28:04 +01:00
Llama Enjoyer
477c0ae2ac
feat: add token counter and model search
...
Added /api/count_tokens endpoint to count tokens using model tokenizer
Implemented token counter in chat UI
Added model search functionality
Updated .gitignore to include .vscode/
2025-02-02 11:15:02 +01:00
Llama Enjoyer
8524c8cc43
feat: Add model parameter management and UI controls
...
Add support for reading generation_config.json for model parameters
Implement model_loaded_callback for parameter updates
Add UI controls to switch between model and app defaults
Add visual highlighting for modified parameters
Add new API endpoints for parameter management
Improve session handling of model parameters
Update UI styling for parameter controls
2025-02-01 00:01:13 +01:00
turboderp
f4d9478812
Merge pull request #64 from Downtown-Case/master
...
Add DRY and Banned Strings to Notepad
2025-01-24 20:28:43 +01:00
turboderp
a05e8e004a
Merge pull request #70 from LlamaEnjoyer/qol_tweaks
...
Add auto session naming and context tracking qol features.
2025-01-24 20:25:08 +01:00
Llama Enjoyer
1f1de65b0d
Merge branch 'more_qol_features' into qol_tweaks
2025-01-24 08:39:51 +01:00
Llama Enjoyer
3e0f293e94
Add cancel button to model loading overlay
2025-01-24 08:39:02 +01:00
Llama Enjoyer
7c577b448f
Add auto session naming and context tracking qol features.
2025-01-14 22:51:05 +01:00
turboderp
36e488c183
Add Mistral-specific templates
2024-12-27 00:09:07 +01:00
turboderp
95c63efbae
Merge remote-tracking branch 'github/master'
2024-10-10 00:17:43 +02:00
turboderp
206f25eb47
Bump to exllamav2 0.2.3 and fix for API changes
2024-10-10 00:17:32 +02:00
Downtown-Case
a2b8cf16f6
Add DRY and Banned Strings
2024-09-28 02:25:33 -04:00
Downtown-Case
729ebb933a
Add DRY and Banned Strings
2024-09-28 02:24:10 -04:00
turboderp
bb017b72f5
Merge pull request #63 from LlamaEnjoyer/copy_to_clipboard
...
Add option to copy block of text to clipboard.
2024-09-23 18:19:01 +02:00
Llama Enjoyer
6b55c54e00
Add option to copy block of text to clipboard.
2024-09-22 11:57:42 +02:00
turboderp
cb0982152d
Update reqs
2024-08-30 20:55:36 +02:00
turboderp
0d58ca8b31
Add switch to enable TP mode
2024-08-30 20:55:09 +02:00
turboderp
396a56f037
Add Q6 and Q8 cache modes
2024-06-09 21:56:20 +02:00
turboderp
47f3831cae
Update exllamav2 to 0.1.5
2024-06-09 21:56:03 +02:00
turboderp
a529749be0
Include MarkedJS
2024-06-01 15:08:22 +02:00
turboderp
7f6f307a81
Assume subdirectories in app path rather than current path
2024-05-31 00:11:11 +02:00
turboderp
5e002a6a94
Add Granite prompt format
2024-05-30 20:55:49 +02:00
turboderp
81d9638a8f
Raise maximum min/max/chunk tokens to 8k
2024-05-17 09:06:44 +02:00
turboderp
ee27aa01a4
Fix regression
2024-05-12 00:30:10 +02:00
turboderp
73d63761bf
Add min tokens and banned strings
2024-05-11 23:21:05 +02:00
turboderp
cc9c9e3074
Bump exllamav2 to 0.0.21
2024-05-11 23:20:47 +02:00
turboderp
11aec2abf4
Add BOS to ChatML template
2024-05-05 17:09:01 +02:00
turboderp
426fbca344
Add fallback stop condition for Phi3
2024-04-24 22:24:31 +02:00
turboderp
af6021d147
Add Phi3 template
2024-04-24 21:15:43 +02:00
turboderp
b33d45d45b
Add Llama3 template
2024-04-19 17:41:01 +02:00
turboderp
0e919cb1a1
Bump exllamav2 to 0.0.19
2024-04-19 14:04:06 +02:00
turboderp
444457af26
Add Cohere template
2024-04-06 09:46:02 +02:00
turboderp
984097f93c
fix for some broken tokenizers
2024-03-30 11:06:38 +01:00
turboderp
86b6352461
ctrl+enter shortcut to regenerate last response
2024-03-27 06:55:53 +01:00
turboderp
2a23cb83f0
Add context options: delete-from-here, regenerate, complete
2024-03-27 06:46:24 +01:00
turboderp
a04c201ea8
Add save button to code blocks.
2024-03-26 05:28:25 +01:00
turboderp
84be9a4b4f
Escape HTML in user blocks
2024-03-26 04:57:19 +01:00
turboderp
94ba8817d2
Prevent HTML tags from rendering in code blocks
2024-03-26 04:47:36 +01:00
turboderp
d3e2ccc782
Fix typo in Gemma prompt format
2024-03-26 04:35:30 +01:00
turboderp
415f49ffa4
Remove some console output
2024-03-21 02:00:24 +01:00
turboderp
64c201d66a
Temp fix for exllamav2 tokenizer bug
2024-03-20 16:36:37 +01:00
