fix: fix llama 3.1 70b (#402)

Signed-off-by: Sertac Ozercan <[email protected]>
sozercan · Sep 27, 2024 · 1b0a7f1 · 1b0a7f1
1 parent b4d52b5
commit 1b0a7f1
Showing 1 changed file with 4 additions and 4 deletions.
diff --git a/models/llama-3.1-70b-instruct.yaml b/models/llama-3.1-70b-instruct.yaml
@@ -4,8 +4,8 @@ debug: true
 runtime: cuda
 models:
  - name: llama-3.1-70b-instruct
- source: https://huggingface.co/lmstudio-community/Meta-Llama-3.1-70B-Instruct-GGUF/resolve/main/Meta-Llama-3.1-70B-Instruct-Q4_K_M.gguf
- sha256: "34d1d88b70a67dc19088ca84e226673962766607791882a7a85959b92857bcb3"
+ source: https://huggingface.co/MaziyarPanahi/Meta-Llama-3.1-70B-Instruct-GGUF/resolve/main/Meta-Llama-3.1-70B-Instruct.Q4_K_M.gguf
+ sha256: "3f16ab17da4521fe3ed7c5d7beed960d3fe7b5b64421ee9650aa53d6b649ccab"
  promptTemplates:
  - name: chatMsg
  template: |
@@ -63,7 +63,7 @@ config: |
  response_regex:
  - <function=(?P<name>\w+)>(?P<arguments>.*)</function>
  parameters:
- model: Meta-Llama-3.1-70B-Instruct-Q4_K_M.gguf
+ model: Meta-Llama-3.1-70B-Instruct.Q4_K_M.gguf
  context_size: 8192
  f16: true
  template:
@@ -74,5 +74,5 @@ config: |
  stopwords:
  - <|im_end|>
  - <dummy32000>
- - \"<|eot_id|>\"
+ - <|eot_id|>
  - <|end_of_text|>