docs: working ollama for desktop (#125)

This commit is contained in:
Michael Neale
2024-10-09 08:04:20 +11:00
committed by GitHub
parent c3c5a4dc77
commit fc25247efd

View File

@@ -12,18 +12,18 @@ OLLAMA_MODEL = "mistral-nemo"
class OllamaProvider(OpenAiProvider):
"""Provides chat completions for models hosted by Ollama"""
__doc__ += f"""
__doc__ += """Here's an example profile configuration to try:
Here's an example profile configuration to try:
First run: ollama pull qwen2.5, then use this profile:
ollama:
provider: ollama
processor: {OLLAMA_MODEL}
accelerator: {OLLAMA_MODEL}
moderator: passive
toolkits:
- name: developer
requires: {{}}
ollama:
provider: ollama
processor: qwen2.5
accelerator: qwen2.5
moderator: truncate
toolkits:
- name: developer
requires: {}
"""
def __init__(self, client: httpx.Client) -> None: