Commits · 7df342a6ea818462e8ddef04776ba6535f42a1fd · OpenDAS / ollama

26 Aug, 2023 2 commits
- allow F16 to use metal · b25dd179
  Michael Yang authored Aug 26, 2023
```
warning F16 uses significantly more memory than quantized model so the
standard requires don't apply.
```
  b25dd179
- add 34b to mem check · 304f2b6c
  Michael Yang authored Aug 26, 2023
  
  304f2b6c
17 Aug, 2023 1 commit
- model and file type as strings · a894cc79
  Michael Yang authored Aug 17, 2023
  
  a894cc79
14 Aug, 2023 1 commit
- close open files · e26085b9
  Michael Yang authored Aug 14, 2023
  
  e26085b9
10 Aug, 2023 4 commits
- implement loading ggml lora adapters through the modelfile · 6de5d032
  Michael Yang authored Aug 03, 2023
  
  6de5d032
- check memory requirements before loading · d791df75
  Michael Yang authored Aug 03, 2023
  
  d791df75
- disable gpu for q5_0, q5_1, q8_0 quants · 020a3b35
  Michael Yang authored Aug 03, 2023
  
  020a3b35
- partial decode ggml bin for more info · fccf8d17
  Michael Yang authored Jul 21, 2023
  
  fccf8d17