Add Intern-S2-Mobius (interns2_mobius) model - #1771
Draft
nightscape wants to merge 1 commit into
Draft
Conversation
nightscape
force-pushed
the
add-interns2-mobius
branch
from
August 21, 2026 12:39
7aa901c to
1647b7a
Compare
nightscape
marked this pull request as ready for review
August 21, 2026 13:32
zcbenz
reviewed
Aug 22, 2026
zcbenz
left a comment
Member
There was a problem hiding this comment.
Can you please use the new pull request template:
https://github.com/ml-explore/mlx-lm/blob/main/.github/PULL_REQUEST_TEMPLATE/new_model.md
Member
|
@nightscape You did not fill the template. |
nightscape
force-pushed
the
add-interns2-mobius
branch
from
August 25, 2026 10:31
8f98aab to
2616525
Compare
nightscape
marked this pull request as draft
August 25, 2026 10:34
nightscape
force-pushed
the
add-interns2-mobius
branch
4 times, most recently
from
August 25, 2026 11:43
4f343d3 to
cc625a8
Compare
Hybrid Gated-DeltaNet / full-attention MoE with globally shared expert banks, fused gate/up SwitchGLU, and a guarded Metal router-topk kernel with a pure-MLX fallback.
nightscape
force-pushed
the
add-interns2-mobius
branch
from
August 25, 2026 13:34
cc625a8 to
64f1121
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Requirements
Model information
Supported checkpoints
Verification
Run mlx-lm command:
mlx_lm.generate --model nightscape/Intern-S2-Mobius-4bit-mlx -p "The secret to baking a good cake is" -m 1024 --trust-remote-code--trust-remote-codeis required for the customtokenization_interns1.pytokenizer.Output (from the 4-bit conversion; upstream bf16 would use the same code path):
(measured on a MacBook Max M4 128GB RAM)