github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #14319] Request to add SARVAM-30b and 105b #71377

New Issue
Open
opened 2026-05-05 01:25:20 -05:00 by GiteaMirror · 17 comments
GiteaMirror commented 2026-05-05 01:25:20 -05:00
Owner
Copy Link

Originally created by @Swagatade on GitHub (Feb 19, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/14319

Originally created by @Swagatade on GitHub (Feb 19, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/14319
GiteaMirror added the model label 2026-05-05 01:25:20 -05:00
GiteaMirror commented 2026-05-05 01:25:21 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 19, 2026):

Image
<!-- gh-comment-id:3926398905 --> @rick-github commented on GitHub (Feb 19, 2026): <img width="547" height="101" alt="Image" src="https://github.com/user-attachments/assets/ab2ced25-071e-44bf-b72c-c9519d877ed5" />
GiteaMirror commented 2026-05-05 01:25:21 -05:00
Author
Owner
Copy Link

@Swagatade commented on GitHub (Feb 19, 2026):

This is the link
sarvamai/sarvam-m · Hugging Face https://share.google/QiP332OJTP0Wtcw6Y

😅

<!-- gh-comment-id:3926948090 --> @Swagatade commented on GitHub (Feb 19, 2026): This is the link sarvamai/sarvam-m · Hugging Face https://share.google/QiP332OJTP0Wtcw6Y 😅
GiteaMirror commented 2026-05-05 01:25:21 -05:00
Author
Owner
Copy Link

@Swagatade commented on GitHub (Feb 19, 2026):

Can I take this job ?? And try it....
I am very excited to add this model and contribute ollama

<!-- gh-comment-id:3926956156 --> @Swagatade commented on GitHub (Feb 19, 2026): Can I take this job ?? And try it.... I am very excited to add this model and contribute ollama
GiteaMirror commented 2026-05-05 01:25:21 -05:00
Author
Owner
Copy Link

@Swagatade commented on GitHub (Feb 19, 2026):

This model launched yesterday in india ai summit.

<!-- gh-comment-id:3927178928 --> @Swagatade commented on GitHub (Feb 19, 2026): This model launched yesterday in india ai summit.
GiteaMirror commented 2026-05-05 01:25:22 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 19, 2026):

Not clicking on a random share link, but https://huggingface.co/sarvamai/sarvam-m is a finetune of mistralai/Mistral-Small-3.1-24B-Base-2503 and so is a supported architecture.

To import the model weights, you can either pull the BF16 weights or the Q8 weights, the command for this is shown by the "Use this model" button on the right hand side of the model card.

Unfortunately the ollama template that comes with these pulls is broken, so it will have be replaced by a version of the original template converted to Go template.

These models are 46G and 25G in size respectively, if you want to make the model more accessible it will have to be quantized to q4_K_M. To do that you will have to download the safetensors form of the model and use the llama.cpp tool converty_hf_to_gguf.py to convert to GGUF. Then you can either use ollama or llama.cpp to quantize.

When you have a working model, you can push it to your ollama namespace to make it available to other ollama users.

<!-- gh-comment-id:3927484201 --> @rick-github commented on GitHub (Feb 19, 2026): Not clicking on a random share link, but https://huggingface.co/sarvamai/sarvam-m is a finetune of [mistralai/Mistral-Small-3.1-24B-Base-2503](https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Base-2503) and so is a supported architecture. To import the model weights, you can either pull the [BF16](https://huggingface.co/sarvamai/sarvam-m-gguf) weights or the [Q8](https://huggingface.co/sarvamai/sarvam-m-q8-gguf) weights, the command for this is shown by the "Use this model" button on the right hand side of the model card. Unfortunately the ollama template that comes with these pulls is broken, so it will have be replaced by a version of the [original template](https://huggingface.co/sarvamai/sarvam-m/blob/main/tokenizer_config.json#L9010) converted to [Go template](https://github.com/ollama/ollama/blob/main/docs/template.mdx). These models are 46G and 25G in size respectively, if you want to make the model more accessible it will have to be quantized to q4_K_M. To do that you will have to download the [safetensors](https://huggingface.co/sarvamai/sarvam-m) form of the model and use the [llama.cpp](https://github.com/ggml-org/llama.cpp) tool [converty_hf_to_gguf.py](https://github.com/ggml-org/llama.cpp/blob/master/convert_hf_to_gguf.py) to convert to GGUF. Then you can either use [ollama](https://github.com/ollama/ollama/blob/main/docs/import.mdx#quantizing-a-model) or [llama.cpp](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize) to quantize. When you have a working model, you can [push](https://github.com/ollama/ollama/blob/main/docs/import.mdx#sharing-your-model-on-ollamacom) it to your ollama namespace to make it available to other ollama users.
GiteaMirror commented 2026-05-05 01:25:23 -05:00
Author
Owner
Copy Link

@Swagatade commented on GitHub (Feb 19, 2026):

Ok next time I will be careful.

<!-- gh-comment-id:3928104037 --> @Swagatade commented on GitHub (Feb 19, 2026): Ok next time I will be careful.
GiteaMirror commented 2026-05-05 01:25:24 -05:00
Author
Owner
Copy Link

@Swagatade commented on GitHub (Feb 20, 2026):

IMG20260220083921_20260220084101.jpg

These two models are coming soon

<!-- gh-comment-id:3931355532 --> @Swagatade commented on GitHub (Feb 20, 2026): ![IMG20260220083921_20260220084101.jpg](https://github.com/user-attachments/assets/c3d7532d-511a-4ccb-94d5-66a71ba08185) These two models are coming soon
GiteaMirror commented 2026-05-05 01:25:25 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 20, 2026):

If the models are "built from scratch" there's a good chance they will not be supported in ollama at release.

<!-- gh-comment-id:3932792504 --> @rick-github commented on GitHub (Feb 20, 2026): If the models are "built from scratch" there's a good chance they will not be supported in ollama at release.
GiteaMirror commented 2026-05-05 01:25:26 -05:00
Author
Owner
Copy Link

@Swagatade commented on GitHub (Feb 20, 2026):

No those models are not built for scratch

<!-- gh-comment-id:3934291036 --> @Swagatade commented on GitHub (Feb 20, 2026): No those models are not built for scratch
GiteaMirror commented 2026-05-05 01:25:27 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 20, 2026):

Image
<!-- gh-comment-id:3934315120 --> @rick-github commented on GitHub (Feb 20, 2026): <img width="597" height="207" alt="Image" src="https://github.com/user-attachments/assets/803de1dd-1cb0-4b62-b075-bdc4750d4d18" />
GiteaMirror commented 2026-05-05 01:25:27 -05:00
Author
Owner
Copy Link

@irzathe-coder commented on GitHub (Feb 26, 2026):

A kind request to add sarvam 30B and sarvam 105B to ollama

<!-- gh-comment-id:3967769069 --> @irzathe-coder commented on GitHub (Feb 26, 2026): A kind request to add sarvam 30B and sarvam 105B to ollama
GiteaMirror commented 2026-05-05 01:25:30 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 26, 2026):

Sarvam have to release the models first.

<!-- gh-comment-id:3967781581 --> @rick-github commented on GitHub (Feb 26, 2026): Sarvam have to release the models first.
GiteaMirror commented 2026-05-05 01:25:33 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Mar 7, 2026):

https://github.com/ggml-org/llama.cpp/issues/20175

<!-- gh-comment-id:4017014665 --> @rick-github commented on GitHub (Mar 7, 2026): https://github.com/ggml-org/llama.cpp/issues/20175
GiteaMirror commented 2026-05-05 01:25:36 -05:00
Author
Owner
Copy Link

@Swagatade commented on GitHub (Mar 10, 2026):

@rick-github https://github.com/ollama/ollama/pull/14703 check this

<!-- gh-comment-id:4028364805 --> @Swagatade commented on GitHub (Mar 10, 2026): @rick-github https://github.com/ollama/ollama/pull/14703 check this
GiteaMirror commented 2026-05-05 01:25:37 -05:00
Author
Owner
Copy Link

@rohan-pandeyy commented on GitHub (Apr 4, 2026):

Hey @jmorganca, can I take this issue up?
Both sarvam-30b and sarvam-105b have been added to the HF library
Image

<!-- gh-comment-id:4186665800 --> @rohan-pandeyy commented on GitHub (Apr 4, 2026): Hey @jmorganca, can I take this issue up? Both [sarvam-30b](https://huggingface.co/sarvamai/sarvam-30b) and [sarvam-105b](https://huggingface.co/sarvamai/sarvam-30b) have been added to the HF library <img width="625" height="159" alt="Image" src="https://github.com/user-attachments/assets/213d6b3f-c946-49a2-a64f-87dca19e8dc2" />
GiteaMirror commented 2026-05-05 01:25:38 -05:00
Author
Owner
Copy Link

@Swagatade commented on GitHub (Apr 4, 2026):

@jmorganca yes

<!-- gh-comment-id:4186675238 --> @Swagatade commented on GitHub (Apr 4, 2026): @jmorganca yes
GiteaMirror commented 2026-05-05 01:25:40 -05:00
Author
Owner
Copy Link

@utsavhanda commented on GitHub (Apr 23, 2026):

Open PR: https://github.com/ollama/ollama/pull/14703

Is there a timeline for the Sarvam models to be available on the Ollama?

<!-- gh-comment-id:4305441779 --> @utsavhanda commented on GitHub (Apr 23, 2026): Open PR: https://github.com/ollama/ollama/pull/14703 Is there a timeline for the Sarvam models to be available on the 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 model
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#71377
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?