OpenVINO
gemma2
alignment-handbook
Generated from Trainer
fakezeta commited on
Commit
512bce3
1 Parent(s): c2962ce

Update README.md

Browse files

added stopwords to localai configuration

Files changed (1) hide show
  1. README.md +5 -2
README.md CHANGED
@@ -14,7 +14,7 @@ license: mit
14
  # OpenVINO IR model with int8 quantization
15
 
16
  Model definition for LocalAI:
17
- ```
18
  name: gemma-2-9b-it-SimPO
19
  backend: transformers
20
  parameters:
@@ -23,10 +23,13 @@ context_size: 8192
23
  type: OVModelForCausalLM
24
  template:
25
  use_tokenizer_template: true
 
 
 
26
  ```
27
 
28
  To run the model directly with LocalAI:
29
- ```
30
  local-ai run huggingface://fakezeta/gemma-2-9b-it-SimPO-ov-int8/model.yaml
31
  ```
32
 
 
14
  # OpenVINO IR model with int8 quantization
15
 
16
  Model definition for LocalAI:
17
+ ```yaml
18
  name: gemma-2-9b-it-SimPO
19
  backend: transformers
20
  parameters:
 
23
  type: OVModelForCausalLM
24
  template:
25
  use_tokenizer_template: true
26
+ stopwords:
27
+ - "<eos>"
28
+ - "<end_of_turn>"
29
  ```
30
 
31
  To run the model directly with LocalAI:
32
+ ```bash
33
  local-ai run huggingface://fakezeta/gemma-2-9b-it-SimPO-ov-int8/model.yaml
34
  ```
35