Commits · e3fb1fd3f1a963033745c7229300c1cdaeb7d15d · orangecat / ollama

"configs/seko_talk/L40s/2gpu/seko_talk_bf16.json" did not exist on "af5105c7b1111e2031e2d44d1bede83d30bd32be"

03 Aug, 2023 1 commit
- update llama.cpp · c5bcf328
  Michael Yang authored Aug 03, 2023
  
  c5bcf328
02 Aug, 2023 1 commit
- no gpu for 70B model · 74a5f7e6
  Michael Yang authored Aug 01, 2023
  
  74a5f7e6
01 Aug, 2023 2 commits
- remove -Werror · 319f078d
  Michael Yang authored Jul 31, 2023
```
there are compile warnings on Linux which -Werror elevates to errors,
preventing compile
```
  319f078d
- only build metal for `darwin,arm` target · 7da249fc
  Jeffrey Morgan authored Jul 31, 2023
  
  7da249fc
28 Jul, 2023 3 commits
- allow specifying stop conditions in modelfile · 184ad8f0
  Bruce MacDonald authored Jul 27, 2023
  
  184ad8f0
- embed ggml-metal.metal · 35496766
  Michael Yang authored Jul 26, 2023
  
  35496766
- add stop conditions · fadf75f9
  Michael Yang authored Jul 27, 2023
  
  fadf75f9
27 Jul, 2023 6 commits
- add NumGQA · ad3a7d0e
  Michael Yang authored Jul 27, 2023
  
  ad3a7d0e
- sample metrics · cca61181
  Michael Yang authored Jul 25, 2023
  
  cca61181
- lock on llm.lock(); decrease batch size · c4904161
  Michael Yang authored Jul 20, 2023
  
  c4904161
- add session expiration · f62a8827
  Michael Yang authored Jul 19, 2023
  
  f62a8827
- update predict code · 3003fc03
  Michael Yang authored Jul 19, 2023
  
  3003fc03
- session id · 35af37a2
  Michael Yang authored Jul 18, 2023
  
  35af37a2
25 Jul, 2023 2 commits
- enable k quants · 726bc647
  Michael Yang authored Jul 25, 2023
  
  726bc647
- enable accelerate · cb55fa92
  Michael Yang authored Jul 24, 2023
  
  cb55fa92
22 Jul, 2023 1 commit
- allocate a large enough tokens slice · b71c67b6
  Michael Yang authored Jul 21, 2023
  
  b71c67b6
15 Jul, 2023 2 commits
- fix multibyte responses · 40c9dc0a
  Michael Yang authored Jul 14, 2023
  
  40c9dc0a
- size_t · 0142660b
  Michael Yang authored Jul 14, 2023
  
  0142660b
14 Jul, 2023 1 commit
- continue conversation · 1775647f
  Michael Yang authored Jul 13, 2023
```
feed responses back into the llm
```
  1775647f
13 Jul, 2023 1 commit
- return more info in generate response · 05e08d23
  Michael Yang authored Jul 12, 2023
  
  05e08d23
12 Jul, 2023 3 commits
- fix eof error in generate · e1f0a0dc
  Michael Yang authored Jul 12, 2023
  
  e1f0a0dc
- return error if model fails to load · c63f8119
  Jeffrey Morgan authored Jul 11, 2023
  
  c63f8119
- error checking new model · c5f7eadd
  Michael Yang authored Jul 11, 2023
  
  c5f7eadd
11 Jul, 2023 2 commits
- vendor llama.cpp · 442dec1c
  Michael Yang authored Jul 11, 2023
  
  442dec1c
- call llama.cpp directly from go · fd4792ec
  Michael Yang authored Jul 07, 2023
  
  fd4792ec
09 Jul, 2023 1 commit
- llama: remove unused helper functions · 5fb96255
  Jeffrey Morgan authored Jul 09, 2023
  
  5fb96255
07 Jul, 2023 2 commits
- pass model and predict options · 3f1b7177
  Patrick Devine authored Jul 06, 2023
  
  3f1b7177
- more free · 5dc9c8ff
  Michael Yang authored Jul 06, 2023
  
  5dc9c8ff
06 Jul, 2023 8 commits
- remove prompt cache · da74384a
  Bruce MacDonald authored Jul 06, 2023
  
  da74384a
- more free · 2c80eddd
  Michael Yang authored Jul 06, 2023
  
  2c80eddd
- use `Makefile` for dependency building instead of `go generate` · 9fe01867
  Jeffrey Morgan authored Jul 06, 2023
  
  9fe01867
- remove debug print statements · 0998d4f0
  Jeffrey Morgan authored Jul 05, 2023
  
  0998d4f0
- fix crash in bindings · 79a999e9
  Jeffrey Morgan authored Jul 05, 2023
  
  79a999e9
- client updates · fd962a36
  Jeffrey Morgan authored Jul 04, 2023
  
  fd962a36
- fix llama.cpp build · 02401653
  Jeffrey Morgan authored Jul 03, 2023
  
  02401653
- add llama.cpp go bindings · 6093a88c
  Jeffrey Morgan authored Jul 03, 2023
  
  6093a88c