-
Notifications
You must be signed in to change notification settings - Fork 50
Pull requests: erfanzar/EasyDeL
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix(generation): fp4 byte-size and stale v1→v2 MLA warning
#264
opened Apr 24, 2026 by
Gpgabriel25
Loading…
feat(qwen3_next): add slow reference path and CPU dispatch guard
#263
opened Apr 24, 2026 by
Gpgabriel25
Loading…
fix(caching): use dtype.itemsize instead of finfo().bits for byte size
#262
opened Apr 24, 2026 by
Gpgabriel25
Loading…
fix(modules): guard HF modeling patches behind torch availability
#261
opened Apr 24, 2026 by
Gpgabriel25
Loading…
feat(grpo): whitelist input_features for audio-aware multimodal GRPO
#260
opened Apr 23, 2026 by
cliangyu
Loading…
6 tasks done
Default finish_reason to "stop" instead of unsupported "finished"
#232
opened Oct 17, 2025 by
AlienKevin
Loading…
fix: auto-detect platform and exclude Triton on non-Linux systems
#229
opened Oct 4, 2025 by
opooladz
Loading…
feat: extend GRPO with ProRL/DAPO controls and document RL/inference stack
#228
opened Oct 4, 2025 by
opooladz
Loading…
feat: Add attention transfer and feature matching to DistillationTrainer
#227
opened Oct 4, 2025 by
opooladz
Loading…
(feat) add attention logits to model output, add attention soft_cap to vanilla attention; (fix) DP sharding of batch, update dtype of memory tracking interval
#209
opened Aug 2, 2025 by
dvruette
Contributor
Loading…
ProTip!
Find all pull requests that aren't related to any open issues with -linked:issue.