[GH-ISSUE #1314] Wrong Prompt on Library Starling-lm Modelfile #47193

Closed
opened 2026-04-28 03:25:23 -05:00 by GiteaMirror · 1 comment
Owner

Originally created by @fakezeta on GitHub (Nov 29, 2023).
Original GitHub issue: https://github.com/ollama/ollama/issues/1314

Hi,

prompt in library model starling-lm is wrong.
Later has been fixed (in this thread) that it must be:

GPT4 User: {prompt}<|end_of_turn|>GPT4 Assistant:

I've tested locally with the following Modelfile:

FROM starling.gguf
TEMPLATE """{{ .System }} GPT4 User: {{ .Prompt }}<|end_of_turn|>GPT4 Assistant:"""
PARAMETER num_ctx 4096
PARAMETER stop "<|endoftext|>"
PARAMETER stop "<|end_of_turn|>"

Naturally, num_ctx can be omitted.

Originally created by @fakezeta on GitHub (Nov 29, 2023). Original GitHub issue: https://github.com/ollama/ollama/issues/1314 Hi, prompt in library model [starling-lm](https://ollama.ai/library/starling-lm:latest) is wrong. Later has been fixed ([in this thread](https://huggingface.co/TheBloke/Starling-LM-7B-alpha-GGUF/discussions/1#6566093f8556065b5238eeb2)) that it must be: `GPT4 User: {prompt}<|end_of_turn|>GPT4 Assistant:` I've tested locally with the following Modelfile: ``` FROM starling.gguf TEMPLATE """{{ .System }} GPT4 User: {{ .Prompt }}<|end_of_turn|>GPT4 Assistant:""" PARAMETER num_ctx 4096 PARAMETER stop "<|endoftext|>" PARAMETER stop "<|end_of_turn|>" ``` Naturally, `num_ctx` can be omitted.
GiteaMirror added the model label 2026-04-28 03:25:23 -05:00
Author
Owner

@jmorganca commented on GitHub (Nov 29, 2023):

Thanks! This is fixed now

<!-- gh-comment-id:1832087689 --> @jmorganca commented on GitHub (Nov 29, 2023): Thanks! This is fixed now
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#47193