github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-16 21:25:46 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #4206] Llama3 generating Incorrect and repeated response for my custom modelfile (finetuned llama3 8b) #64657

New Issue
Closed
opened 2026-05-03 18:27:08 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-05-03 18:27:08 -05:00
Owner
Copy Link

Originally created by @balaji-2k1 on GitHub (May 6, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/4206

What is the issue?

Fine-tuned the Llama3 8b model for my custom dataset following the llama3 template, used the unsloth library for fine-tuning, and generated the GGUF file.
Got relevant results when using this model for inference, but when I created the Modelfile and ran it as an Ollama process, the responses were irrelevant and repetitive.

Below is the prompt I used during the inference in the jupyter notebook (It generated the expected response):
prompt = """<|start_header_id|>system<|end_header_id|>You are Dbot, Respond to user questions naturally. <|eot_id|><|start_header_id|>user<|end_header_id|> {}<|eot_id|> <|start_header_id|>Dbot<|end_header_id|> {}<|eot_id|>"""

Below is my Modelfile
`FROM llama3_merged.gguf

TEMPLATE """<|start_header_id|>system<|end_header_id|>You are Dbot,
Respond to user questions naturally. >
<|eot_id|><|start_header_id|>user<|end_header_id|>
{{ .Prompt }}<|eot_id|>
<|start_header_id|>Dbot<|end_header_id|>
{{ .Response }}<|eot_id|>"""

PARAMETER stop "<|start_header_id|>"
PARAMETER stop "<|end_header_id|>"
PARAMETER stop "<|eot_id|>"
PARAMETER stop "<|reserved_special_token"
`
I used below command to run the Modelfile and infer the model which gave me irrelevant responses and did not stop.

`ollama run llama3:v1 "What is EoC"

EoC: Ecosystems and EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems.
The EoC: Ecosystems and EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems.
Ecosystems have a significant impact on the EoC: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems.
The EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems.
Ecosystems have a significant impact on the EoC: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems.
The EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems.
Ecosystems have a significant impact on the EoC: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems.
The EoP: Ecosystems, EoT: Ecosystems, and EoL: Ecosystems.
Ecosystems have a significant impact on the EoC: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems.
The EoT: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoE: Ecosystems.
Ecosystems have a significant impact on the EoC: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems.
The EoE: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems.
Ecosystems have a significant impact on the EoC: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems.
The EoT: Ecosystems, EoE: Ecosystems, and EoL: Ecosystems.
Ecosystems have a significant impact on the EoC: Ecos`

OS

Linux

GPU

Nvidia

CPU

Intel

Ollama version

0.1.32

Originally created by @balaji-2k1 on GitHub (May 6, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/4206 ### What is the issue? Fine-tuned the Llama3 8b model for my custom dataset following the llama3 template, used the unsloth library for fine-tuning, and generated the GGUF file. Got relevant results when using this model for inference, but when I created the Modelfile and ran it as an Ollama process, the responses were irrelevant and repetitive. Below is the prompt I used during the inference in the jupyter notebook (It generated the expected response): `prompt = """<|start_header_id|>system<|end_header_id|>You are `Dbot,` Respond to user questions naturally. <|eot_id|><|start_header_id|>user<|end_header_id|> {}<|eot_id|> <|start_header_id|>Dbot<|end_header_id|> {}<|eot_id|>"""` Below is my Modelfile `FROM llama3_merged.gguf TEMPLATE """<|start_header_id|>system<|end_header_id|>You are Dbot, Respond to user questions naturally. > <|eot_id|><|start_header_id|>user<|end_header_id|> {{ .Prompt }}<|eot_id|> <|start_header_id|>Dbot<|end_header_id|> {{ .Response }}<|eot_id|>""" PARAMETER stop "<|start_header_id|>" PARAMETER stop "<|end_header_id|>" PARAMETER stop "<|eot_id|>" PARAMETER stop "<|reserved_special_token" ` I used below command to run the Modelfile and infer the model which gave me irrelevant responses and did not stop. `ollama run llama3:v1 "What is EoC" EoC: Ecosystems and EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems. The EoC: Ecosystems and EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems. Ecosystems have a significant impact on the EoC: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems. The EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems. Ecosystems have a significant impact on the EoC: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems. The EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems. Ecosystems have a significant impact on the EoC: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems. The EoP: Ecosystems, EoT: Ecosystems, and EoL: Ecosystems. Ecosystems have a significant impact on the EoC: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems. The EoT: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoE: Ecosystems. Ecosystems have a significant impact on the EoC: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems. The EoE: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems. Ecosystems have a significant impact on the EoC: Ecosystems, EoL: Ecosystems, EoP: Ecosystems, and EoT: Ecosystems. The EoT: Ecosystems, EoE: Ecosystems, and EoL: Ecosystems. Ecosystems have a significant impact on the EoC: Ecos` ### OS Linux ### GPU Nvidia ### CPU Intel ### Ollama version 0.1.32
GiteaMirror added the bug label 2026-05-03 18:27:08 -05:00
GiteaMirror commented 2026-05-03 18:27:11 -05:00
Author
Owner
Copy Link

@jkongWPJK commented on GitHub (Jul 12, 2024):

I'm having the same problem, still couldn't figure out why

<!-- gh-comment-id:2225650619 --> @jkongWPJK commented on GitHub (Jul 12, 2024): I'm having the same problem, still couldn't figure out why
GiteaMirror commented 2026-05-03 18:27:12 -05:00
Author
Owner
Copy Link

@qq1273834091 commented on GitHub (Dec 16, 2024):

我有同样的问题，仍然无法弄清楚为什么

<!-- gh-comment-id:2545011510 --> @qq1273834091 commented on GitHub (Dec 16, 2024): 我有同样的问题，仍然无法弄清楚为什么
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 bug
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#64657
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?