github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #13707] Support request for https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 #71048

New Issue
Closed
opened 2026-05-04 23:50:56 -05:00 by GiteaMirror · 6 comments
GiteaMirror commented 2026-05-04 23:50:56 -05:00
Owner
Copy Link

Originally created by @eslowney on GitHub (Jan 13, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/13707

When attempting to run ollama create from the HF /safetensors*, I get:

invalid character 'I' looking for beginning of value

v0.13.5

Originally created by @eslowney on GitHub (Jan 13, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/13707 When attempting to run `ollama create` from the HF /safetensors*, I get: `invalid character 'I' looking for beginning of value` v0.13.5
GiteaMirror commented 2026-05-04 23:51:00 -05:00
Author
Owner
Copy Link

@illusdolphin commented on GitHub (Jan 14, 2026):

It's already supported https://ollama.com/library/nemotron-3-nano so just
ollama run nemotron-3-nano

<!-- gh-comment-id:3748098373 --> @illusdolphin commented on GitHub (Jan 14, 2026): It's already supported https://ollama.com/library/nemotron-3-nano so just `ollama run nemotron-3-nano`
GiteaMirror commented 2026-05-04 23:51:01 -05:00
Author
Owner
Copy Link

@eslowney commented on GitHub (Jan 14, 2026):

It's already supported https://ollama.com/library/nemotron-3-nano so just ollama run nemotron-3-nano

I'm trying to run it off-line, so the run commands don't work for me (or any network connection to ollama.com or anywhere else)

<!-- gh-comment-id:3751012499 --> @eslowney commented on GitHub (Jan 14, 2026): > It's already supported https://ollama.com/library/nemotron-3-nano so just `ollama run nemotron-3-nano` I'm trying to run it off-line, so the `run` commands don't work for me (or any network connection to ollama.com or anywhere else)
GiteaMirror commented 2026-05-04 23:51:03 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 15, 2026):

Download the model when online, copy it to your offline machine.

$ cd /tmp
$ OLLAMA_MODELS=/tmp/models ollama serve 2>&- >&- &
$ ollama pull nemotron-3-nano
$ kill $!
$ zip -r models.zip models
$ rm -rf models

Copy models.zip to your offline machine and extract into ollama directory.

$ sudo -u ollama -s
$ cd ~/.ollama
$ unzip /tmp/models.zip

Alternatively, show what command you are using to convert the HF safetensors into an ollama model.

<!-- gh-comment-id:3753175162 --> @rick-github commented on GitHub (Jan 15, 2026): Download the model when online, copy it to your offline machine. ```console $ cd /tmp $ OLLAMA_MODELS=/tmp/models ollama serve 2>&- >&- & $ ollama pull nemotron-3-nano $ kill $! $ zip -r models.zip models $ rm -rf models ``` Copy models.zip to your offline machine and extract into ollama directory. ```console $ sudo -u ollama -s $ cd ~/.ollama $ unzip /tmp/models.zip ``` Alternatively, show what command you are using to convert the HF safetensors into an ollama model.
GiteaMirror commented 2026-05-04 23:51:03 -05:00
Author
Owner
Copy Link

@eslowney commented on GitHub (Jan 15, 2026):

i just have a Modelfile with
FROM /models/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8
and then
ollama create Nemotron-3-Nano-30B from within that Modelfile directory.

However, I also tried to convert the model to a gguf first with convert_hf_to_gguf.py from llama.cpp, and that is also throwing a different error, so it could just be the model itself isn't supported by anything yet.

<!-- gh-comment-id:3756291785 --> @eslowney commented on GitHub (Jan 15, 2026): i just have a Modelfile with `FROM /models/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8` and then `ollama create Nemotron-3-Nano-30B` from within that Modelfile directory. However, I also tried to convert the model to a gguf first with `convert_hf_to_gguf.py` from llama.cpp, and that is also throwing a different error, so it could just be the model itself isn't supported by anything yet.
GiteaMirror commented 2026-05-04 23:51:05 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 15, 2026):

that is also throwing a different error

Difficult to diagnose the issue without the actual error message. Nemotron is supported by ollama, safetensors import is not, llama.cpp conversion should work with the BF16 tensors.

<!-- gh-comment-id:3757128153 --> @rick-github commented on GitHub (Jan 15, 2026): > that is also throwing a different error Difficult to diagnose the issue without the actual error message. Nemotron is supported by ollama, safetensors import is not, llama.cpp conversion should work with the BF16 tensors.
GiteaMirror commented 2026-05-04 23:51:06 -05:00
Author
Owner
Copy Link

@eslowney commented on GitHub (Jan 16, 2026):

that is also throwing a different error

Difficult to diagnose the issue without the actual error message. Nemotron is supported by ollama, safetensors import is not, llama.cpp conversion should work with the BF16 tensors.

Thanks, I downloaded and imported the -BF16 version (instead of -FP8) and it quantized fine and I was able to import it into ollama.

<!-- gh-comment-id:3761314030 --> @eslowney commented on GitHub (Jan 16, 2026): > > that is also throwing a different error > > Difficult to diagnose the issue without the actual error message. Nemotron is supported by ollama, safetensors import is not, llama.cpp conversion should work with the BF16 tensors. Thanks, I downloaded and imported the -BF16 version (instead of -FP8) and it quantized fine and I was able to import it into ollama.
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#71048
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?