Continue LLM VS Code Plugin
@CurrentFile
@Files
https://docs.continue.dev/customize/custom-providers
Ollama remote access: https://4sysops.com/archives/remote-http-access-to-self-hosted-ollama-ai-models/
sudo EDITOR=vi systemctl edit ollama.service
sudo systemctl daemon-reload
sudo systemctl restart ollama
models:
- name: "Remote Llama 3"
provider: "ollama"
model: "llama3"
apiBase: "http://<REMOTE_IP_ADDRESS>:11434" # Replace with your server's IP
GPU Power Limit
Make Power Limit Persistent Across Reboots [1]
The best way to make this setting stick on Linux is using a systemd service. [1]
Create a script (e.g., /usr/local/bin/nvidia-set-power.sh):
bash
sudo nano /usr/local/bin/nvidia-set-power.sh
Add the following content (adjust wattage):
bash
#!/bin/bash
/usr/bin/nvidia-smi -pl 200
Make the script executable:
bash
sudo chmod +x /usr/local/bin/nvidia-set-power.sh
Create a systemd service file (/etc/systemd/system/nvidia-power.service):
bash
sudo nano /etc/systemd/system/nvidia-power.service
Add this content to the service file:
ini
[Unit]
Description=Set NVIDIA Power Limit
After=multi-user.target
[Service]
Type=oneshot
ExecStart=/usr/local/bin/nvidia-set-power.sh
RemainAfterExit=yes
[Install]
WantedBy=multi-user.target
Enable the service to run on boot:
bash
sudo systemctl enable nvidia-power.service
sudo systemctl start nvidia-power.service
3. Verify
After a reboot, verify the setting: [1]
bash
nvidia-smi -q -d POWER
Continue VS Plugin With Ollama Support
name: Local Config
version: 1.0.0
schema: v1
models:
- name: Autodetect
provider: ollama
model: AUTODETECT
defaultCompletionOptions:
contextLength: 30000
roles:
- chat
- edit
- apply
- rerank
- autocomplete
context:
- provider: codebase
- provider: folder
- provider: clipboard
- provider: url
- provider: tree
- provider: problems