click gear for settings, then + button

Choose ^{LM Studio} for provider

Often models from huggingface like,

won't show, choose any, and a config.yaml file will open, where, you setup you model from ^{LM Studio}

Create any name for -name, in our case we're naming it Autocomplete

provider, in our case is LM Studio

4

The first image, in this step , shows the model name that should be copied and inserted into model

model, in our case is ^{qwen2.5-coder-1.5b-instruct}

get the apiBase from same widow in lm studio as name to copy, Add the GET v1 base api endpoint to the end of server address

Finally , add roles of model in program we connect, which in vs code is 'autocomplete'

Now autocomplete shows up in the models tab of 'Continue' settings/configurations,

**Important Note, .yaml file won't work if spacing is wrong

          - name: Autocomplete
            provider: lmstudio
            roles:
              - autocomplete
            model: qwen2.5-coder-1.5b-instruct
            apiBase: http://127.0.0.1:1234/v1

You can see model working in LM Studio while autocommplete

**Important Notes if time by model is too slow you won't see autocommplete , you can check total time taken by model in logs in LM studio. In our case it's 1000ms