-
Notifications
You must be signed in to change notification settings - Fork 313
Pull requests: ml-explore/mlx-swift-lm
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Qwen3 embedder: honor attentionMask instead of silently ignoring it
#418
opened Jul 12, 2026 by
spokvulcan
Contributor
Loading…
Enable MTP speculative decoding for Gemma 4 12B (gemma4_unified target + gemma4_unified_assistant drafter)
#415
opened Jul 10, 2026 by
fdagostino
Contributor
Loading…
ChatSession: keep the cached transcript canonical on continuation turns
#414
opened Jul 9, 2026 by
fdagostino
Contributor
Loading…
Generation: honor suppress_tokens and mask multimodal placeholder tokens
#412
opened Jul 9, 2026 by
fdagostino
Contributor
Loading…
Qwen3VL: apply the sRGB tone curve in the image preprocess path
#411
opened Jul 9, 2026 by
spokvulcan
Contributor
Loading…
Fix tool round-trip: record assistant tool_calls message before tool results
#409
opened Jul 9, 2026 by
jacobwillemsma
Loading…
4 tasks done
fix: Honor safetensors index when loading weights
#408
opened Jul 8, 2026 by
Zrxrxrx
Loading…
7 of 8 tasks
Gemma 4: aspect-preserving image resize with per-image dynamic token counts; fix multi-image requests
#405
opened Jul 7, 2026 by
prbme
Loading…
Fix Qwen3.5 VLM garbage output: gate the RMSNorm weight shift in sanitize
#403
opened Jul 6, 2026 by
GoodOlClint
Contributor
Loading…
3 of 4 tasks
Add custom LogitProcessor injection to GenerateParameters / ChatSession
#401
opened Jul 5, 2026 by
xocialize
Loading…
Gemma4Unified (12B): audio and video input in the processor
#400
opened Jul 5, 2026 by
fdagostino
Contributor
Loading…
Qwen3.5/3.6: windowed prefill (finishes the #344 deferred item) + warm windowed continuation
#399
opened Jul 5, 2026 by
spokvulcan
Contributor
Loading…
Qwen3VL: default per-image resolution to a 2,560 vision-token budget
#398
opened Jul 5, 2026 by
spokvulcan
Contributor
Loading…
Add regression coverage for mixed-precision (per-layer) quantized checkpoint loading
#395
opened Jul 5, 2026 by
fdagostino
Contributor
Loading…
Gemma 4: native audio input (port the gemma4_audio USM-Conformer encoder)
#392
opened Jul 4, 2026 by
fdagostino
Contributor
Loading…
Add end-to-end video input to the base Gemma 4 (gemma4) VLM
#391
opened Jul 3, 2026 by
fdagostino
Contributor
Loading…
Convert Gemma tool parameters by schema type
#388
opened Jul 3, 2026 by
atdrendel
Contributor
Loading…
4 tasks done
Add Gemma3 allHiddenStates for frozen-text-encoder use
#387
opened Jul 2, 2026 by
xocialize
Loading…
Fix MLXVLM Gemma4 loader to honor num_kv_shared_layers (E-series targets)
#384
opened Jun 30, 2026 by
GoodOlClint
Contributor
Loading…
3 of 4 tasks
Implement Gemma 4 E-series MTP centroid embedder (use_ordered_embeddings)
#383
opened Jun 30, 2026 by
GoodOlClint
Contributor
Loading…
3 of 4 tasks
Fix silent state truncation for non-multiple-of-32 GDN key dim
#381
opened Jun 29, 2026 by
GoodOlClint
Contributor
Loading…
4 tasks done
Decode PoolingConfiguration with optional pooling_mode flags
#376
opened Jun 29, 2026 by
GoodOlClint
Contributor
Loading…
4 tasks done
Previous Next
ProTip!
What’s not been updated in a month: updated:<2026-06-13.