pancake
|
7057b5c0d7
|
Add hint for gpt-4-turbo model support
|
2023-12-30 00:26:16 +01:00 |
pancake
|
c614d85e65
|
Add llm.maxmsglen config var
|
2023-12-30 00:23:48 +01:00 |
pancake
|
416654ba5d
|
Remove some dupped imports and fix r2 plugin integration
|
2023-12-29 21:27:29 +01:00 |
pancake
|
089032a55c
|
Add support for zephyr models and improve ferret/dolphin one
|
2023-12-29 20:28:13 +01:00 |
pancake
|
312791b48b
|
Fix chat.reply
|
2023-12-29 20:27:31 +01:00 |
pancake
|
ed10c51683
|
Add support for the utopiaXL and the dolphin-PHI models
|
2023-12-29 20:10:33 +01:00 |
pancake
|
0c02a63411
|
Add support for the Apple's Ferret model
|
2023-12-29 19:52:12 +01:00 |
pancake
|
e730e7a102
|
Native and python implementations can't coexist, force likewise uninstall
|
2023-12-29 19:32:17 +01:00 |
pancake
|
f824af387c
|
One less backtrace when the model is not found
|
2023-12-29 19:30:22 +01:00 |
pancake
|
14f51a39ec
|
Fix null deref when model cant be found in the r2ai-native
|
2023-12-29 10:34:46 +01:00 |
pancake
|
5405e69b65
|
First working version of r2ai-native, still far from clean
|
2023-12-20 21:13:51 +01:00 |
pancake
|
5d87f1f174
|
Add uninstall target
|
2023-12-19 20:16:06 +01:00 |
pancake
|
6d2d855dcf
|
Fix build on Linux
|
2023-12-19 20:14:52 +01:00 |
pancake
|
50051433af
|
A bit more work integrating llama with the r2ai cxx plugin
|
2023-12-19 19:35:37 +01:00 |
pancake
|
e0b69fc07e
|
Initial rewrite of llama's main example for the native
|
2023-12-19 11:55:51 +01:00 |
pancake
|
7579175cb9
|
Fix -M
|
2023-12-19 11:09:10 +01:00 |
pancake
|
7891957a37
|
Initial import of the r2ai http client plugin
|
2023-12-12 20:31:31 +01:00 |
pancake
|
f51e4c33b1
|
Implement the -w command to start the http server
|
2023-12-12 20:06:06 +01:00 |
pancake
|
0cbf4fd9d4
|
Add the mistral 0.2 in the listing
|
2023-12-12 11:41:24 +01:00 |
pancake
|
83708433a1
|
Improve error messages and fix missing import for the voice support
|
2023-12-11 01:10:13 +01:00 |
pancake
|
b593ceb399
|
Initial import of the C rewrite of r2ai (very WIP just a PoC)
|
2023-12-07 12:42:32 +01:00 |
pancake
|
b69ef96c04
|
Improved r2pipe with argument support
* `#!pipe r2pm -r r2ai examples/r2test.py arg1 arg2`
|
2023-12-06 18:38:36 +01:00 |
pancake
|
d8eb08b0ef
|
Improved support for venv and rlang command handling
|
2023-12-06 16:01:21 +01:00 |
pancake
|
95a2abe010
|
Initial support to index json files (WIP PoC)
|
2023-12-06 15:54:27 +01:00 |
pancake
|
49dbbf28ae
|
Vim integration through adding support to use r2ai as a hashbang to run scripts
|
2023-12-06 15:12:14 +01:00 |
pancake
|
a9f4a9d280
|
Initial Wikipedia integration via `wikit`
|
2023-12-06 14:56:04 +01:00 |
pancake
|
159af0d76a
|
use ai to split relevant keywords for mastodon search
|
2023-12-06 13:27:22 +01:00 |
pancake
|
af7665f143
|
Initial support for magicoder model and fix chat.reply
|
2023-12-06 10:51:40 +01:00 |
pancake
|
a5c6ecb38d
|
Initial support for Mastodon in vectordb
|
2023-12-06 10:51:09 +01:00 |
pancake
|
53b6273c8d
|
Fix bug in -t
|
2023-12-06 10:50:04 +01:00 |
pancake
|
4b471cfa0b
|
Fix -t
|
2023-12-05 19:26:35 +01:00 |
pancake
|
84ca584e6e
|
Fix deepseek prompt template, as it's alpaca based
|
2023-12-05 12:14:23 +01:00 |
pancake
|
c549a3ff6b
|
Add git commit script example
|
2023-12-05 11:36:00 +01:00 |
pancake
|
08ecc7930b
|
Add extra check when no model is selected
|
2023-12-05 11:33:37 +01:00 |
pancake
|
c0ea855773
|
Fix "inr2" checks, now r2ai works via r2pipe, rlang and standalone
|
2023-12-04 19:47:13 +01:00 |
pancake
|
24d76dfef7
|
Fix systemprompt varname
|
2023-12-01 17:32:11 +01:00 |
pancake
|
b1ba080e39
|
Make code compression algorithm recursive
|
2023-12-01 15:40:44 +01:00 |
pancake
|
0908d07072
|
Add function call example
|
2023-12-01 15:40:26 +01:00 |
pancake
|
74e9d15997
|
Support parsing large code (> 4KB)
|
2023-11-30 18:16:56 +01:00 |
pancake
|
8f9b7a89c8
|
Fix prompt and internal chat log for llama models
|
2023-11-29 15:39:28 +01:00 |
pancake
|
f5354006f9
|
Also expose the self.model (WIP)
|
2023-11-29 12:24:21 +01:00 |
pancake
|
95bf393d8d
|
Protect env and report better errors when using unexistent ones
|
2023-11-29 12:20:15 +01:00 |
pancake
|
2cf08f4949
|
Make context window, max tokens and temperature configurable
|
2023-11-29 12:15:55 +01:00 |
pancake
|
1f850701da
|
Add support for scripting
|
2023-11-29 12:00:52 +01:00 |
pancake
|
a6f8cabdd1
|
Initial PoC for the `aikeywords` implementation
|
2023-11-29 11:37:15 +01:00 |
pancake
|
a50b163373
|
vectordb fails if nothing is indexed :{
|
2023-11-29 11:36:59 +01:00 |
pancake
|
1fee6b29ae
|
Makefile fails if not r2 is available
|
2023-11-29 11:36:42 +01:00 |
pancake
|
efb473a884
|
Fix the role prompt for alpaca models
|
2023-11-28 15:38:42 +01:00 |
pancake
|
399fccc0dd
|
Deprecate self.withresponse and just use env.chat.reply
|
2023-11-28 15:38:25 +01:00 |
pancake
|
5b88cbae9d
|
Fix vectordb usage when mprt is not available (macOS)
|
2023-11-28 11:58:48 +01:00 |