backend für neural zu llama

b64c1e53 · Demian Frister · 65a9d5b8 · b64c1e53 · b64c1e53 · b64c1e53
Commit b64c1e53 authored 1 year ago by Demian Frister
--- a/gpt-3.5-turbo.yaml
+++ b/gpt-3.5-turbo.yaml
@@ -10,6 +10,7 @@ urls:
 config_file: |
  name: neural
  context_size: 8096
+  backend: llama
  f16: true
  mmap: true
  threads: 4

--- a/neural.yaml
+++ b/neural.yaml
@@ -11,6 +11,7 @@ config_file: |
  name: neural
  context_size: 8096
  f16: true
+  backend: llama
  mmap: true
  threads: 4
  parameters:

--- a/neural32.yaml
+++ b/neural32.yaml
@@ -10,6 +10,7 @@ urls:
 config_file: |
  name: neural
  context_size: 32000
+  backend: llama
  f16: true
  mmap: true
  threads: 4