Commit Graph

112 Commits

Author SHA1 Message Date
Raven Scott
1f3ff44317 fix token reducer 2023-08-21 21:15:10 -04:00
Raven Scott
a980f3cd43 improving token reducer logic 2023-08-15 01:41:39 -04:00
Raven Scott
05e79cba3a fixing up token reducing if sessions too large 2023-08-15 01:08:56 -04:00
Raven Scott
f91d66b2b3 Adding assistant to Counter 2023-08-14 23:44:06 -04:00
Raven Scott
6efd069b5d counting tokens properly 2023-08-14 23:00:23 -04:00
Raven Scott
bd435ca311 default to disk cache 2023-06-12 18:45:39 +02:00
Raven Scott
61ee13bfbd remove comments 2023-06-12 18:16:16 +02:00
Raven Scott
10473ef702 beginnings of using a tokenizer to check CTX length, remove context if needed 2023-06-12 18:15:11 +02:00
Raven Scott
20c83a656a update dockerfile for GPU 2023-05-31 23:02:07 +02:00
Raven Scott
61c2fed773 update dockerfile for server-non-gpu 2023-05-27 07:35:41 +02:00
Raven Scott
6ed34de804 update readme 2023-05-27 02:02:38 +02:00
Raven Scott
6bf6c1ef28 Dont send anything if time==0 2023-05-23 15:53:53 +02:00
Raven Scott
7102bf32f0 Fix GPU Cache 2023-05-23 14:56:21 +02:00
Raven Scott
099dbf908b update 2023-05-22 18:38:47 +02:00
Raven Scott
bc6157e4a1 bug fix 2023-05-22 18:23:17 +02:00
Raven Scott
da0650d3b6 resetting botMessage to stop crashing 2023-05-22 17:46:59 +02:00
Raven Scott
393999165b adding GPU ENV Var to main compose 2023-05-22 17:35:20 +02:00
Raven Scott
c75926728c comments 2023-05-21 00:20:53 +02:00
Raven Scott
927b5c834d Add warning about caching with cuBLAS 2023-05-20 23:47:16 +02:00
Raven Scott
668b343cbb changing the name scheme for docker-compose 2023-05-20 15:14:25 +02:00
Raven Scott
51a41292e6 bringing back the embed 2023-05-20 15:11:58 +02:00
Raven Scott
2ac55922d2 revert back to older style of workflow until new logic is written 2023-05-20 04:08:27 +02:00
Raven Scott
73636804a5 bug fix 2023-05-20 02:49:17 +02:00
Raven Scott
c7a3316d45 bug fix 2023-05-19 23:49:55 +02:00
Raven Scott
1fe0f20e6f revert 2023-05-19 22:55:25 +02:00
Raven Scott
8ce9e18656 removing non needed code 2023-05-19 22:43:47 +02:00
Raven Scott
2924822a49 ensuring GPU env is set in default .env 2023-05-19 21:45:20 +02:00
Raven Scott
368004f10b adding NVIDIA GPU Support with Stats 2023-05-19 21:32:21 +02:00
Raven Scott
4b090592ad Fix hugging face 2023-05-08 22:15:51 +02:00
Raven Scott
64be911772 adding REPEAT_PENALTY 2023-05-08 22:00:24 +02:00
Raven Scott
83a7bb90ed Fix up env 2023-05-08 21:12:57 +02:00
Raven Scott
14fa3b06ff Remove the abuse of looping functions 2023-05-08 15:45:05 +02:00
Raven Scott
c7d8735c8a Remove any @ symbol to remove spam pings 2023-05-08 15:42:15 +02:00
Raven Scott
012566e93c update model var 2023-05-07 14:18:17 +02:00
Raven Scott
d18cb39ecd update hf 2023-05-06 22:19:08 +02:00
Raven Scott
9bc44c56f8 update hf 2023-05-06 14:34:53 +02:00
Raven Scott
973952aee1 Revert "Urban"
This reverts commit bd5ef4db9a.
2023-05-06 03:41:36 +02:00
Raven Scott
3d2546ebcf Revert "Adding urban dictionary lookups"
This reverts commit f87542132e.
2023-05-06 03:41:16 +02:00
Raven Scott
f87542132e Adding urban dictionary lookups 2023-05-06 02:32:10 +02:00
Raven Scott
bd5ef4db9a Urban 2023-05-06 02:30:46 +02:00
Raven Scott
ee47531d2f HuggingFace Deploy Instructions 2023-05-06 01:49:21 +02:00
Raven Scott
7ea6abb0f8 HuggingFace Deploy Instructions 2023-05-06 01:46:37 +02:00
Raven Scott
4ff16b4fc7 HuggingFace Deploy Instructions 2023-05-06 01:45:01 +02:00
Raven Scott
d1807d37ad HuggingFace Deploy Instructions 2023-05-06 01:41:03 +02:00
Raven Scott
af13ca3717 Move overflow delay from static to .env 2023-05-05 21:07:39 +02:00
Raven Scott
cb880f9bc1 Fix Caching web reading the web: Append last message rather than add to the conversation. 2023-05-05 20:49:29 +02:00
Raven Scott
aec98b576b update default.env 2023-05-05 19:31:17 +02:00
Raven Scott
3b1ec922c4 remove \t 2023-05-05 19:08:13 +02:00
Raven Scott
e54826085d remove redunant code 2023-05-05 18:32:46 +02:00
Raven Scott
18923b7909 cleaner output from web scrapes 2023-05-05 18:31:24 +02:00