github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-21 13:11:39 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #5264] How to use the .mf model configuration file to register a customize vision-language model in Ollama #3297

New Issue
Open
opened 2026-04-12 13:51:35 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-12 13:51:35 -05:00
Owner
Copy Link

Originally created by @LJY16114 on GitHub (Jun 25, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/5264

This is my .mf model configuration file:

FROM /root/llava-llama-3-8b-v1_1-gguf/llava-llama-3-8b-v1_1-f16.gguf
TEMPLATE """
{{ if .System }}<|start_header_id|>system<|end_header_id|>

{{ .System }}<|eot_id|>{{ end }}{{ if .Prompt }}<|start_header_id|>user<|end_header_id|>

{{ .Prompt }}<|eot_id|>{{ end }}<|start_header_id|>assistant<|end_header_id|>

{{ .Response }}<|eot_id|>
"""
PARAMETER stop "<|system|>"
PARAMETER stop "<|user|>"
PARAMETER stop "<|assistant|>

Ollama can recognize our customized .gguf model and display a model option in open-webui.
However, the loaded model cannot understand the content of the input image and randomly says things that are completely unrelated to the image.

What could be the cause of this problem?

Thank you in advance!

@jmorganca

Originally created by @LJY16114 on GitHub (Jun 25, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/5264 This is my .mf model configuration file: FROM /root/llava-llama-3-8b-v1_1-gguf/llava-llama-3-8b-v1_1-f16.gguf TEMPLATE """ {{ if .System }}<|start_header_id|>system<|end_header_id|> {{ .System }}<|eot_id|>{{ end }}{{ if .Prompt }}<|start_header_id|>user<|end_header_id|> {{ .Prompt }}<|eot_id|>{{ end }}<|start_header_id|>assistant<|end_header_id|> {{ .Response }}<|eot_id|> """ PARAMETER stop "<|system|>" PARAMETER stop "<|user|>" PARAMETER stop "<|assistant|> Ollama can recognize our customized .gguf model and display a model option in open-webui. However, the loaded model cannot understand the content of the input image and randomly says things that are completely unrelated to the image. What could be the cause of this problem? Thank you in advance! @jmorganca
GiteaMirror commented 2026-04-12 13:51:37 -05:00
Author
Owner
Copy Link

@ljluestc commented on GitHub (Mar 11, 2025):


# Specify the base language model file
FROM /root/llava-llama-3-8b-v1_1-gguf/llava-llama-3-8b-v1_1-f16.gguf

# Specify the multimodal projection file (adjust the path as needed)
# This is critical for image processing
PARAMETER mmproj /root/llava-llama-3-8b-v1_1-gguf/llava-llama-3-8b-v1_1-mmproj-f16.gguf

# Define the template for multimodal input
# The <image> tag is a placeholder for the image data
TEMPLATE """
{{ if .System }}<|start_header_id|>system<|end_header_id>

{{ .System }}<|eot_id>{{ end }}{{ if .Prompt }}<|start_header_id|>user<|end_header_id>

<image>
{{ .Prompt }}<|eot_id>{{ end }}<|start_header_id|>assistant<|end_header_id>

{{ .Response }}<|eot_id>
"""

# Stop tokens to ensure proper conversation flow
PARAMETER stop "<|system|>"
PARAMETER stop "<|user|>"
PARAMETER stop "<|assistant|>"
PARAMETER stop "<|eot_id>"

# Optional: Set additional parameters for better performance
PARAMETER temperature 0.7
PARAMETER top_p 0.9
PARAMETER num_ctx 4096
​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​
<!-- gh-comment-id:2715469114 --> @ljluestc commented on GitHub (Mar 11, 2025): ``` # Specify the base language model file FROM /root/llava-llama-3-8b-v1_1-gguf/llava-llama-3-8b-v1_1-f16.gguf # Specify the multimodal projection file (adjust the path as needed) # This is critical for image processing PARAMETER mmproj /root/llava-llama-3-8b-v1_1-gguf/llava-llama-3-8b-v1_1-mmproj-f16.gguf # Define the template for multimodal input # The <image> tag is a placeholder for the image data TEMPLATE """ {{ if .System }}<|start_header_id|>system<|end_header_id> {{ .System }}<|eot_id>{{ end }}{{ if .Prompt }}<|start_header_id|>user<|end_header_id> <image> {{ .Prompt }}<|eot_id>{{ end }}<|start_header_id|>assistant<|end_header_id> {{ .Response }}<|eot_id> """ # Stop tokens to ensure proper conversation flow PARAMETER stop "<|system|>" PARAMETER stop "<|user|>" PARAMETER stop "<|assistant|>" PARAMETER stop "<|eot_id>" # Optional: Set additional parameters for better performance PARAMETER temperature 0.7 PARAMETER top_p 0.9 PARAMETER num_ctx 4096 ​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​ ```
GiteaMirror commented 2026-04-12 13:51:38 -05:00
Author
Owner
Copy Link

@rjmalagon commented on GitHub (Mar 17, 2025):


# Specify the base language model file
FROM /root/llava-llama-3-8b-v1_1-gguf/llava-llama-3-8b-v1_1-f16.gguf

# Specify the multimodal projection file (adjust the path as needed)
# This is critical for image processing
PARAMETER mmproj /root/llava-llama-3-8b-v1_1-gguf/llava-llama-3-8b-v1_1-mmproj-f16.gguf

# Define the template for multimodal input
# The <image> tag is a placeholder for the image data
TEMPLATE """
{{ if .System }}<|start_header_id|>system<|end_header_id>

{{ .System }}<|eot_id>{{ end }}{{ if .Prompt }}<|start_header_id|>user<|end_header_id>

<image>
{{ .Prompt }}<|eot_id>{{ end }}<|start_header_id|>assistant<|end_header_id>

{{ .Response }}<|eot_id>
"""

# Stop tokens to ensure proper conversation flow
PARAMETER stop "<|system|>"
PARAMETER stop "<|user|>"
PARAMETER stop "<|assistant|>"
PARAMETER stop "<|eot_id>"

# Optional: Set additional parameters for better performance
PARAMETER temperature 0.7
PARAMETER top_p 0.9
PARAMETER num_ctx 4096
​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​

This is not correct, the projection file is not loaded like a PARAMETER.
Both gguf files are loaded with FROM

FROM /path/main_model.gguf
FROM /path/projection_mmproj.gguf
<!-- gh-comment-id:2728280403 --> @rjmalagon commented on GitHub (Mar 17, 2025): > ``` > > # Specify the base language model file > FROM /root/llava-llama-3-8b-v1_1-gguf/llava-llama-3-8b-v1_1-f16.gguf > > # Specify the multimodal projection file (adjust the path as needed) > # This is critical for image processing > PARAMETER mmproj /root/llava-llama-3-8b-v1_1-gguf/llava-llama-3-8b-v1_1-mmproj-f16.gguf > > # Define the template for multimodal input > # The <image> tag is a placeholder for the image data > TEMPLATE """ > {{ if .System }}<|start_header_id|>system<|end_header_id> > > {{ .System }}<|eot_id>{{ end }}{{ if .Prompt }}<|start_header_id|>user<|end_header_id> > > <image> > {{ .Prompt }}<|eot_id>{{ end }}<|start_header_id|>assistant<|end_header_id> > > {{ .Response }}<|eot_id> > """ > > # Stop tokens to ensure proper conversation flow > PARAMETER stop "<|system|>" > PARAMETER stop "<|user|>" > PARAMETER stop "<|assistant|>" > PARAMETER stop "<|eot_id>" > > # Optional: Set additional parameters for better performance > PARAMETER temperature 0.7 > PARAMETER top_p 0.9 > PARAMETER num_ctx 4096 > ​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​ > ``` This is not correct, the projection file is not loaded like a `PARAMETER`. Both gguf files are loaded with `FROM` ``` FROM /path/main_model.gguf FROM /path/projection_mmproj.gguf ```
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
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#3297
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?