3d76ef7b
(HEAD -> master)
Rewritten some examples docs, explaining what these examples show instead of just showing the source code. (#728) by
2024-05-11 04:00:45 +0100
e70db1f3
(refs/pull/676/merge)
Merge 43786b0d9a into 3b0b2ab224 by
2024-05-11 10:13:09 +0800
6293cca1
(refs/pull/717/merge)
Merge 9a68785022 into 3b0b2ab224 by
2024-05-11 10:11:58 +0800
045d6d71
(refs/pull/728/head)
Rewritten some examples docs, explaining what these examples show instead of just showing the source code. by
2024-05-10 19:19:24 +0100
867436b2
(refs/pull/726/merge)
Merge 7b309d7bf6 into 3b0b2ab224 by
2024-05-10 15:07:59 +0000
7b309d7b
(refs/pull/726/head)
KernelMemory bug fix - cleanup nullable refs by
2024-05-10 17:07:53 +0200
9053d1dd
(refs/pull/712/merge)
Merge 703df2e3f3 into 3b0b2ab224 by
2024-05-10 22:30:47 +0800
3b0b2ab2
Merge pull request #721 from martindevans/kv_cache_view by
2024-05-10 15:19:36 +0100
703df2e3
(refs/pull/712/head)
Renamed one of the commented out pre tokenization items, showing the pattern they should all use. by
2024-05-10 15:12:31 +0100
44bd5b31
Merge pull request #715 from martindevans/llama-templater by
2024-05-10 15:10:22 +0100
e3044812
KernelMemory bug fix by
2024-05-10 10:07:34 +0200
b326624a
(refs/pull/715/head)
Split template out to a field, so it can be changed more easily. by
2024-05-10 00:23:57 +0100
c2f30ebe
(refs/pull/688/merge)
Merge d4cc122317 into b25f93b86d by
2024-05-10 01:34:27 +0800
68705c9f
New set of binaries: - llama.cpp: a743d76a01 - https://github.com/SciSharp/LLamaSharp/actions/runs/9017784838 by
2024-05-09 14:42:42 +0100
b25f93b8
Merge pull request #725 from martindevans/fix_cublas_git_ref by
2024-05-09 14:07:50 +0100
c585eb5b
(refs/pull/725/head)
Fixed cublas action always compiling `master` instead of the specific commit by
2024-05-09 14:05:54 +0100
d509105d
ci: add windows benchmark test. (#723) by
2024-05-09 03:31:54 +0800
b2f25496
(refs/pull/723/head)
revert some changes. by
2024-05-09 03:13:06 +0800
c8804a16
fix library loading error. by
2024-05-09 02:52:58 +0800
f21eb4a9
add cuda 12 linux by
2024-05-09 02:42:22 +0800
21c21316
debug by
2024-05-09 02:12:20 +0800
b9d01689
debug by
2024-05-09 02:01:23 +0800
6c9ba12d
debug by
2024-05-09 01:42:10 +0800
df879698
debug by
2024-05-09 01:25:31 +0800
ddbcc2b7
debug by
2024-05-09 01:16:47 +0800
32c97cb3
debug by
2024-05-09 01:13:25 +0800
94303961
debug by
2024-05-09 01:11:03 +0800
a2e9dd37
debug by
2024-05-09 00:49:22 +0800
d03a146c
debug by
2024-05-09 00:32:56 +0800
db3afb6d
debug by
2024-05-09 00:23:51 +0800
2cd0619f
debug by
2024-05-09 00:08:40 +0800
4efc0a9b
ci: add windows benchmark test. by
2024-05-08 23:51:31 +0800
6f9097f2
ci: add benchmark test. (#720) by
2024-05-08 23:39:49 +0800
7788746d
(refs/pull/720/head)
finish the feature. by
2024-05-08 22:56:56 +0800
e982de98
debug by
2024-05-08 21:54:16 +0800
910c1663
debug by
2024-05-08 21:30:17 +0800
194e7140
debug by
2024-05-08 21:23:06 +0800
fbe4fd4e
debug by
2024-05-08 21:17:04 +0800
12e8c59e
debug by
2024-05-08 21:09:36 +0800
0112bb59
debug by
2024-05-08 21:01:09 +0800
f70af0f0
debug by
2024-05-08 20:54:18 +0800
897e223c
debug by
2024-05-08 20:26:47 +0800
03512627
debug by
2024-05-08 20:20:34 +0800
9758cf15
debug by
2024-05-08 19:40:03 +0800
24f59987
debug by
2024-05-08 19:28:11 +0800
7bdb6955
debug by
2024-05-08 19:17:02 +0800
12a9105d
debug by
2024-05-08 18:59:04 +0800
33ccc542
debug by
2024-05-08 18:51:07 +0800
b2d9fc00
debug by
2024-05-08 18:40:27 +0800
4d52e8af
debug by
2024-05-08 18:30:06 +0800
d2e2b1a2
debug by
2024-05-08 17:53:04 +0800
7807dd74
debug by
2024-05-08 17:29:04 +0800
bfa44544
debug by
2024-05-08 17:14:52 +0800
6dfff81f
debug by
2024-05-08 17:05:34 +0800
261a6422
debug by
2024-05-08 16:26:55 +0800
2ddee300
debug by
2024-05-08 15:30:29 +0800
6bb5795c
debug by
2024-05-08 15:25:23 +0800
dcb5cb65
debug by
2024-05-08 15:15:55 +0800
f336a438
fix model path error. by
2024-05-08 14:18:51 +0800
34f1f58d
debug by
2024-05-08 11:50:26 +0800
4f33c0dd
debug by
2024-05-08 04:49:53 +0800
8f3f608b
debug by
2024-05-08 04:40:34 +0800
851b8fa0
debug by
2024-05-08 04:37:50 +0800
ec92b646
debug by
2024-05-08 04:33:32 +0800
0a1f1500
debug by
2024-05-08 04:31:15 +0800
ba7663aa
debug by
2024-05-08 04:21:39 +0800
1fc35ad8
debug by
2024-05-08 04:15:28 +0800
2eae0cd6
debug by
2024-05-08 04:06:27 +0800
9076d441
debug by
2024-05-08 04:02:08 +0800
d6097ac5
fix error by
2024-05-08 03:21:14 +0800
ee55cc29
debug by
2024-05-08 03:16:41 +0800
2a1e1504
Change the way to cache models. by
2024-05-08 03:14:26 +0800
1efce656
fix error. by
2024-05-08 02:38:41 +0800
413c23ca
feat: add benchmark test for prefill. by
2024-05-08 02:15:06 +0800
2117287d
(refs/pull/721/head)
Expanded the `LLamaKvCacheView` to make it usable without unsafe. - Checking indices - Returning span of correct length - Hiding native methods - Hiding native types by
2024-05-07 13:16:13 +0100
4332ab38
Changes based on review feedback: - Returning template for chaining method calls - Returning a `TextMessage` object instead of a tuple by
2024-05-06 14:04:47 +0100
a0335f67
- Added `LLamaTemplate` which efficiently formats a series of messages according to the model template. - Fixed `llama_chat_apply_template` method (wrong entrypoint, couldn't handle null model) by
2024-05-04 15:53:54 +0100
f4d7533e
(refs/pull/718/merge)
Merge dd224a8158 into 05100184f4 by
2024-05-06 21:02:50 +0000
dd224a81
(refs/pull/718/head)
Removed unused code. by
2024-05-07 07:02:45 +1000
c07c4ccf
debug by
2024-05-07 04:12:40 +0800
8e57b56b
debug by
2024-05-07 04:06:18 +0800
5d76857a
debug by
2024-05-07 04:02:27 +0800
a5739325
debug by
2024-05-07 04:00:21 +0800
9d3eb384
change cuda11 docker image by
2024-05-07 03:57:05 +0800
bfa1bc65
Menu tweaks. by
2024-05-07 05:38:30 +1000
e09d8c26
Fixed typo and updated images. by
2024-05-07 04:57:19 +1000
ad4c94c8
Corrected filenames. by
2024-05-07 04:49:50 +1000
e492ea2b
Updated images on web README. by
2024-05-07 04:42:14 +1000
5fe4814d
fix runner error. by
2024-05-07 02:40:17 +0800
ff9e4f2a
debug by
2024-05-07 02:36:57 +0800
9423a8a7
ci: add benchmark test. by
2024-05-07 02:32:38 +0800
05100184
Merge pull request #719 from martindevans/remove-batched-conversation-prompt-with-string by
2024-05-06 16:16:02 +0100
3ba49754
(refs/pull/719/head)
Removed (marked as obsolete) prompting with a string for `Conversation`. Tokenization requires extra parameters (e.g. addBos, special) which require special considersation. For now it's better to tokenize using other tools and pass the tokens directly. by
2024-05-06 15:53:21 +0100
19cac84b
Merge branch 'web-net-8' of https://github.com/Lamothe/LLamaSharp into web-net-8 by
2024-05-06 22:31:22 +1000
7cf5393b
Async model loading. by
2024-05-06 22:31:02 +1000
54668d61
Update LLama.Web/Common/ModelOptions.cs by
2024-05-06 21:27:45 +1000
a2e4f319
Use CDN for bootstrap also. by
2024-05-06 20:54:18 +1000
9057d716
Use CDN for Bootstrap Icons. by
2024-05-06 20:37:23 +1000
1fbdc215
* Updated to .NET 8 * Updated to Bootstrap 5.3.3 * Use libman for client side packages * Enabled razor runtime compilation in dev * Added fav icon * Removed the footer * Added theme switcher by
2024-05-06 16:57:35 +1000
9a687850
(refs/pull/717/head)
add android support by
2024-05-04 23:24:23 +0100