github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 08:02:14 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #12753] 如何将多模态模型从safetensor格式转到ollama可用的格式？ #70516

New Issue
Closed
opened 2026-05-04 21:49:42 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-05-04 21:49:42 -05:00
Owner
Copy Link

Originally created by @DoubleRedX on GitHub (Oct 23, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/12753

  • 我尝试了llama.cpp的转换脚本将safetensor转成gguf格式，只有text的部分被处理。
  • 我尝试了直接加载safetensor格式的权重，但是出现了bug。

请问是否有类似于llama.cpp中的转换脚本类似的工具，能够将safetensor转换成符合ollama要求的格式的权重？包括这里涉及的量化过程。

Originally created by @DoubleRedX on GitHub (Oct 23, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/12753 - 我尝试了llama.cpp的转换脚本将safetensor转成gguf格式，只有text的部分被处理。 - 我尝试了直接加载safetensor格式的权重，但是出现了bug。 请问是否有类似于llama.cpp中的转换脚本类似的工具，能够将safetensor转换成符合ollama要求的格式的权重？包括这里涉及的量化过程。
GiteaMirror added the needs more info label 2026-05-04 21:49:42 -05:00
GiteaMirror commented 2026-05-04 21:49:44 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Oct 23, 2025):

Which model?

<!-- gh-comment-id:3436482426 --> @rick-github commented on GitHub (Oct 23, 2025): Which model?
GiteaMirror commented 2026-05-04 21:49:44 -05:00
Author
Owner
Copy Link

@DoubleRedX commented on GitHub (Oct 24, 2025):

Which model?

Qwen2.5-vl. 我的最终目的是在特定平台上适配qwen3-vl等多模态模型。这里边涉及到权重格式转换和量化，因此我想用qwen2.5-vl首先验证一下模型转换步骤的正确性。然后发现1、多模态的模型用llama.cpp提供的脚本转成gguf格式的时候，vision的部分是一个单独的文件。2、如果直接从safetensor使用，因为平台的原因无法使用这种方式。3、目前我在尝试从ollama解析权重文件的逻辑倒推编码的逻辑。
所以如果能提供一个权重转换的脚本会方便一些。

<!-- gh-comment-id:3440562685 --> @DoubleRedX commented on GitHub (Oct 24, 2025): > Which model? Qwen2.5-vl. 我的最终目的是在特定平台上适配qwen3-vl等多模态模型。这里边涉及到权重格式转换和量化，因此我想用qwen2.5-vl首先验证一下模型转换步骤的正确性。然后发现1、多模态的模型用llama.cpp提供的脚本转成gguf格式的时候，vision的部分是一个单独的文件。2、如果直接从safetensor使用，因为平台的原因无法使用这种方式。3、目前我在尝试从ollama解析权重文件的逻辑倒推编码的逻辑。 所以如果能提供一个权重转换的脚本会方便一些。
GiteaMirror commented 2026-05-04 21:49:45 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Oct 24, 2025):

What platform issues? Importing directly into ollama from the safetensors format is the recommended way for models using the new ollama engine.

<!-- gh-comment-id:3441296321 --> @rick-github commented on GitHub (Oct 24, 2025): What platform issues? Importing directly into ollama from the safetensors format is the recommended way for models using the new ollama engine.
GiteaMirror commented 2026-05-04 21:49:46 -05:00
Author
Owner
Copy Link

@DoubleRedX commented on GitHub (Nov 6, 2025):

convert_hf_to_gguf.py

I attempted to implement a script for converting a multimodal model from safetensor to gguf format. Adapted from the script in llama.cpp. This script takes Qwen 2.5L as an example. Usage: python convert_hf_to_gguf.py --ollama --outfile /path/to/output.gguf --outtype f16 /root/.cache/modelscope/hub/models/Qwen/Qwen2___5-VL-3B-Instruct/
It might be useful.🙂

<!-- gh-comment-id:3496817889 --> @DoubleRedX commented on GitHub (Nov 6, 2025): [convert_hf_to_gguf.py](https://github.com/user-attachments/files/23390230/convert_hf_to_gguf.py) I attempted to implement a script for converting a multimodal model from safetensor to gguf format. Adapted from the script in llama.cpp. This script takes Qwen 2.5L as an example. Usage: python convert_hf_to_gguf.py --ollama --outfile /path/to/output.gguf --outtype f16 /root/.cache/modelscope/hub/models/Qwen/Qwen2___5-VL-3B-Instruct/ It might be useful.🙂
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 needs more info
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#70516
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?