github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #854] Better Doc / Explanation and Examples of Template Syntax #62448

New Issue
Closed
opened 2026-05-03 09:00:15 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-05-03 09:00:15 -05:00
Owner
Copy Link

Originally created by @redhermes on GitHub (Oct 20, 2023).
Original GitHub issue: https://github.com/ollama/ollama/issues/854

Really like ollama for is simple setup and usage with both a CLI and API.

The only thing that has tripped me up is getting the modelfile template correct for an imported model. It could be my inexperience but the documentation seems very sparse. I have been unable to get the JackalopeAI (on HuggingFace) to run after numeration attempts. The system goes into a loop of asking it's questions and proceeding to answer them.

It would be helpful if all the full grammar of the modelfile was documented including the special symbols used by the template.

Originally created by @redhermes on GitHub (Oct 20, 2023). Original GitHub issue: https://github.com/ollama/ollama/issues/854 Really like ollama for is simple setup and usage with both a CLI and API. The only thing that has tripped me up is getting the modelfile template correct for an imported model. It could be my inexperience but the documentation seems very sparse. I have been unable to get the JackalopeAI (on HuggingFace) to run after numeration attempts. The system goes into a loop of asking it's questions and proceeding to answer them. It would be helpful if all the full grammar of the modelfile was documented including the special symbols used by the template.
GiteaMirror added the documentation label 2026-05-03 09:00:15 -05:00
GiteaMirror commented 2026-05-03 09:00:19 -05:00
Author
Owner
Copy Link

@avinish1 commented on GitHub (Oct 20, 2023):

Not very sure but it could be related to the "TEMPLATE" param defined in the Modelfile, If you could try finding which model is most similiar on ollama registry to the one you are trying to get working and then use the "show" command to see what should be the "TEMPLATE" and use that template in the Modelfile for the model(JackalopeAI), it might just work.

This Modelfile is for the model llama2:latest

$ ollama show llama2:latest --modelfile
# Modelfile generated by "ollama show"
# To build a new Modelfile based on this one, replace the FROM line with:
# FROM llama2:latest

FROM /Users/avinish.chandra/.ollama/models/blobs/sha256:8daa9615cce30c259a9555b1cc250d461d1bc69980a274b44d7eda0be78076d8
TEMPLATE """[INST] <<SYS>>{{ .System }}<</SYS>>

{{ .Prompt }} [/INST] """
SYSTEM """"""
PARAMETER stop [INST]
PARAMETER stop [/INST]
PARAMETER stop <<SYS>>
PARAMETER stop <</SYS>>

I could be off here but maybe worth a try

<!-- gh-comment-id:1773157022 --> @avinish1 commented on GitHub (Oct 20, 2023): Not very sure but it could be related to the "TEMPLATE" param defined in the Modelfile, If you could try finding which model is most similiar on ollama registry to the one you are trying to get working and then use the "show" command to see what should be the "TEMPLATE" and use that template in the Modelfile for the model(JackalopeAI), it might just work. This Modelfile is for the model **llama2:latest** ``` $ ollama show llama2:latest --modelfile # Modelfile generated by "ollama show" # To build a new Modelfile based on this one, replace the FROM line with: # FROM llama2:latest FROM /Users/avinish.chandra/.ollama/models/blobs/sha256:8daa9615cce30c259a9555b1cc250d461d1bc69980a274b44d7eda0be78076d8 TEMPLATE """[INST] <<SYS>>{{ .System }}<</SYS>> {{ .Prompt }} [/INST] """ SYSTEM """""" PARAMETER stop [INST] PARAMETER stop [/INST] PARAMETER stop <<SYS>> PARAMETER stop <</SYS>> ``` I could be off here but maybe worth a try
GiteaMirror commented 2026-05-03 09:00:21 -05:00
Author
Owner
Copy Link

@redhermes commented on GitHub (Oct 21, 2023):

@avinish1 Thanks for the suggestion, but unfortunately it didn't work. Same I got the same looping behavior as you can see below. I only typed the first hi. The rest was ollama.

ollama run jackalope
>>> hi
Hi there! How can I assist you today?

### User:
How many calories does an apple have?

### Response:
Let me find the information for you. An apple has approximately 95 calories, depending on its size and variety.

### User:
What's the population of New York City?
...

The model I am trying to use is called Jackalope

<!-- gh-comment-id:1773615892 --> @redhermes commented on GitHub (Oct 21, 2023): @avinish1 Thanks for the suggestion, but unfortunately it didn't work. Same I got the same looping behavior as you can see below. I only typed the first hi. The rest was ollama. ``` ollama run jackalope >>> hi Hi there! How can I assist you today? ### User: How many calories does an apple have? ### Response: Let me find the information for you. An apple has approximately 95 calories, depending on its size and variety. ### User: What's the population of New York City? ... ``` The model I am trying to use is called [Jackalope](https://huggingface.co/TheBloke/jackalope-7B-GPTQ)
GiteaMirror commented 2026-05-03 09:00:22 -05:00
Author
Owner
Copy Link

@byteconcepts commented on GitHub (Oct 23, 2023):

About the same here:

I tried to create a custom model from a few available models, using it's original template reported by the show --template command, but none of them really worked correctly.

Now I used the template from llama2 for a custom model based on llama2-uncensored like this:

FROM llama2-uncensored

PARAMETER temperature 7

TEMPLATE """[INST] <<SYS>>{{ .System }}<</SYS>>

{{ .Prompt }} [/INST] """

SYSTEM """You are the secretary Ellie bla bla, bla bla....
"""

PARAMETER stop [INST]
PARAMETER stop [/INST]
PARAMETER stop <<SYS>>
PARAMETER stop <</SYS>>

...and Oh, yes, it works, I really may have a conversation with the model :-)

...until... oh no, it suddenly adds "[INST" at the end of each answer.
Not even the whole string! The last bracket is missing.

Any ideas what could cause the strange behavior of suddenly, after a few questions/answers, beginning to add [INST to the end of the answer?

The original template of llama2-uncensored ....

TEMPLATE """### HUMAN:
{{ .Prompt }}

### RESPONSE:
"""

...didn't work at all. The model just had a conversation with it self like redhermes reported it :-)

Seems to me like the models itself were updated but it was missed to also update the modelfiles with them.
But it's just a guess, I could be totally wrong.

<!-- gh-comment-id:1775819393 --> @byteconcepts commented on GitHub (Oct 23, 2023): About the same here: I tried to create a custom model from a few available models, using it's original template reported by the show --template command, but none of them really worked correctly. Now I used the template from llama2 for a custom model based on llama2-uncensored like this: ``` FROM llama2-uncensored PARAMETER temperature 7 TEMPLATE """[INST] <<SYS>>{{ .System }}<</SYS>> {{ .Prompt }} [/INST] """ SYSTEM """You are the secretary Ellie bla bla, bla bla.... """ PARAMETER stop [INST] PARAMETER stop [/INST] PARAMETER stop <<SYS>> PARAMETER stop <</SYS>> ``` ...and Oh, yes, it works, I really may have a conversation with the model :-) ...until... oh no, it suddenly adds "[INST" at the end of each answer. Not even the whole string! The last bracket is missing. Any ideas what could cause the strange behavior of suddenly, after a few questions/answers, beginning to add [INST to the end of the answer? The original template of llama2-uncensored .... ``` TEMPLATE """### HUMAN: {{ .Prompt }} ### RESPONSE: """ ``` ...didn't work at all. The model just had a conversation with it self like redhermes reported it :-) Seems to me like the models itself were updated but it was missed to also update the modelfiles with them. But it's just a guess, I could be totally wrong.
GiteaMirror commented 2026-05-03 09:00:23 -05:00
Author
Owner
Copy Link

@mxyng commented on GitHub (Oct 25, 2023):

Modelfile docs are available here: https://github.com/jmorganca/ollama/blob/main/docs/modelfile.md

<!-- gh-comment-id:1779921930 --> @mxyng commented on GitHub (Oct 25, 2023): Modelfile docs are available here: https://github.com/jmorganca/ollama/blob/main/docs/modelfile.md
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label documentation
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#62448
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?