define required max_tokens for turbo-instruct model

1 files changed, 3 insertions, 3 deletions

diff --git a/README.md b/README.md
index d0af65d..34fdf9d 100644
--- a/README.md
+++ b/README.md
@@ -257,7 +257,7 @@ generate a paragraph of lorem ipsum
 ```
 
 Below are listed all available configuration options, along with their default values.
-Please note that there isn't any token limit imposed, though 1000 is recommended.
+Please note that there isn't any token limit imposed on chat model.
 
 ```vim
 " :AI
@@ -272,7 +272,7 @@ let g:vim_ai_complete = {
 \  "options": {
 \    "model": "gpt-3.5-turbo-instruct",
 \    "endpoint_url": "https://api.openai.com/v1/completions",
-\    "max_tokens": 0,
+\    "max_tokens": 1000,
 \    "temperature": 0.1,
 \    "request_timeout": 20,
 \    "enable_auth": 1,
@@ -295,7 +295,7 @@ let g:vim_ai_edit = {
 \  "options": {
 \    "model": "gpt-3.5-turbo-instruct",
 \    "endpoint_url": "https://api.openai.com/v1/completions",
-\    "max_tokens": 0,
+\    "max_tokens": 1000,
 \    "temperature": 0.1,
 \    "request_timeout": 20,
 \    "enable_auth": 1,