Skip to content

convert: respect Mistral format when loading MoE config - #30669

Merged
CISC merged 1 commit into
ggml-org:masterfrom
aiqubits:fix/mistral-moe-hparams
Oct 16, 2026
Merged

CISC merged 1 commit into
ggml-org:masterfrom
aiqubits:fix/mistral-moe-hparams

Conversation

@aiqubits

@aiqubits aiqubits commented Oct 16, 2026 •

Copy link
Copy Markdown
Contributor

Overview

Use the instance's is_mistral_format flag when DeepseekV2Model reloads configuration,
allowing Mistral MoE checkpoints to read params.json.

Fixes #29899

Additional information

None

Requirements

@github-actions github-actions Bot added testing Everything test related conversion labels Oct 16, 2026
@aiqubits

Copy link
Copy Markdown
Contributor Author

@CISC @ggerganov Can you take a look at this PR?

@CISC CISC left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Drop the test.

@ggml-gh-bot

ggml-gh-bot Bot commented Oct 16, 2026

Copy link
Copy Markdown

Hi @aiqubits, thanks for your contribution!

Per our contribution guidelines, the automated PR checker found the following issue(s) that need your attention:

  • PR Template not respected: Please respect the template when creating a new pull request. Make sure to fill out all required sections.

Please note that maintainers reserve the right to make final decisions on PRs. If you believe there is a mistake, please comment below.

@ggml-gh-bot ggml-gh-bot Bot added the draft PR will be changed to draft by github-actions bot label Oct 16, 2026
@github-actions
github-actions Bot marked this pull request as draft October 16, 2026 11:45
@github-actions github-actions Bot removed the draft PR will be changed to draft by github-actions bot label Oct 16, 2026
Use the instance's is_mistral_format flag when DeepseekV2Model reloads
configuration, allowing Mistral MoE checkpoints to read params.json.

Fixes ggml-org#29899

Signed-off-by: aiqubits <aiqubits@hotmail.com>
@aiqubits
aiqubits force-pushed the fix/mistral-moe-hparams branch from ac24879 to 421adbe Compare October 16, 2026 12:08
@aiqubits

Copy link
Copy Markdown
Contributor Author

@CISC Ok, I have completed the change.

@aiqubits
aiqubits marked this pull request as ready for review October 16, 2026 12:11
@CISC

CISC commented Oct 16, 2026

Copy link
Copy Markdown
Member

@CISC Ok, I have completed the change.

Pay your respects to the bot as well. :)

@aiqubits

Copy link
Copy Markdown
Contributor Author

@CISC Ok, I have completed the change.

Pay your respects to the bot as well. :)

It's just a parameter issue. My bot isn't worth worrying about.

@CISC

CISC commented Oct 16, 2026

Copy link
Copy Markdown
Member

@CISC Ok, I have completed the change.

Pay your respects to the bot as well. :)

It's just a parameter issue. My bot isn't worth worrying about.

Our bot is: #30669 (comment)

@aiqubits

Copy link
Copy Markdown
Contributor Author

@CISC Thanks, I have updated the messages according to the template.

@CISC
CISC merged commit 2a66703 into ggml-org:master Oct 16, 2026
5 checks passed
edwardyoon pushed a commit to edwardyoon/focus-llama that referenced this pull request Oct 24, 2026
Use the instance's is_mistral_format flag when DeepseekV2Model reloads
configuration, allowing Mistral MoE checkpoints to read params.json.

Fixes ggml-org#29899

Signed-off-by: aiqubits <aiqubits@hotmail.com>
auRose94 pushed a commit to auRose94/llama.cpp that referenced this pull request Oct 28, 2026
Use the instance's is_mistral_format flag when DeepseekV2Model reloads
configuration, allowing Mistral MoE checkpoints to read params.json.

Fixes ggml-org#29899

Signed-off-by: aiqubits <aiqubits@hotmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

conversion testing Everything test related

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Conversion to GGUF bug: --mistral-format failed to be applied

2 participants