Skip to content

update mistral common version for PR 47507 - #47677

Merged
itazap merged 4 commits into
mainfrom
update_mistral_common
Aug 4, 2026
Merged

update mistral common version for PR 47507#47677
itazap merged 4 commits into
mainfrom
update_mistral_common

Conversation

@itazap

@itazap itazap commented Jul 31, 2026

Copy link
Copy Markdown
Collaborator

CI

update mistral common version for PR 47507

@HuggingFaceDocBuilderDev

Copy link
Copy Markdown

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

juliendenize added a commit to juliendenize/transformers that referenced this pull request Jul 31, 2026
…vergence error, and drop the mistral-common bump landing in huggingface#47677
@itazap
itazap requested a review from Cyrilvallez August 3, 2026 09:29

@ArthurZucker ArthurZucker left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

lgtm . failing seems unrelated.
I would maybe mark with 🔴 if this does break for previous version (we don't want to break transformers if possible, if we do let's make it explicit in release notes)

@github-actions

github-actions Bot commented Aug 4, 2026

Copy link
Copy Markdown
Contributor

CI recap

Dashboard: View test results in Grafana
Latest run: 30799491698:2
Result: success | Jobs: 16 | Tests: 159,083 | Failures: 0 | Duration: 15h 1m

@itazap
itazap added this pull request to the merge queue Aug 4, 2026
Merged via the queue into main with commit 4806a9b Aug 4, 2026
112 checks passed
@itazap
itazap deleted the update_mistral_common branch August 4, 2026 21:34
pull Bot pushed a commit to ishandutta2007/transformers that referenced this pull request Aug 4, 2026
…ngface#47507)

* [Mistral] Add tekken tokenizer detection, conversion, and save utilities

Add native Mistral "tekken" tokenizer support:
- MistralConverter + convert_tekken_tokenizer: build a PreTrainedTokenizerFast from tekken.json
- resolve_mistral_format: tekken-first backend detection (prefer MistralCommonBackend when
  tekken.json is present and mistral-common is installed; mistral_format=False/True override)
- save_as_tekken + MistralCommonBackend save_format for round-tripping tekken.json
- AutoTokenizer runtime detection wired to prefer the native backend
- auto-generate a chat template from tekken.json during text conversion
  (requires mistral-common>=1.11.7); keep any explicit or on-disk template

* Address review: wire mistral_format in AutoTokenizer and move save_format to TokenizersBackend

* Isolate tekken save_format into TekkenBackend instead of TokenizersBackend

* Fix stray whitespace in tekken tokenizer test docstring

* Address review: drop TekkenBackend, guard lossy mistral-format saves, and align tekken file detection with mistral-common

* Fix push_to_hub dropping the copied tekken.json, clarify the vocab divergence error, and drop the mistral-common bump landing in huggingface#47677
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants