gurubot/smollm-1.7b-q4m

gurubot/ smollm-1.7b-q4m:latest

163 Downloads Updated 1 year ago

Instruct model. important - read description

ollama run gurubot/smollm-1.7b-q4m

curl http://localhost:11434/api/chat \
  -d '{
    "model": "gurubot/smollm-1.7b-q4m",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='gurubot/smollm-1.7b-q4m',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'gurubot/smollm-1.7b-q4m',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Details

Updated 1 year ago

1 year ago

81812aeebd19 · 1.1GB ·

model

archllama

parameters1.71B

quantizationQ4_K_M

1.1GB

params

{ "stop": [ "user:", "assistant:" ] }

32B

template

user: {{ .Prompt }} assistant:

32B

Readme

SmolLM-Instruct 1.7b

Important: nobody seems to have hit on the correct template for this but i’m uploading it with the best performing template I’ve tested. It responds well to most questions but you can’t chat with it like normal, for example if you start by writing “hello” it will probably just respond with maths formula, but you can ask it why the sky is blue, how to bake a cake etc.. It’s very much an “instruct” type interaction.

source https://huggingface.co/MaziyarPanahi/SmolLM-1.7B-Instruct-GGUF