Skip to content

Pull requests: axolotl-ai-cloud/axolotl

New pull request New
95 Open 2,316 Closed
95 Open 2,316 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

train only on last assistant message
#3556 opened Mar 28, 2026 by VarunGumma Loading…
5
Add precompute_ref_log_probs to config schema
#3555 opened Mar 27, 2026 by joaquinhuigomez Loading…
1
6
qwen3.5 configs
#3554 opened Mar 27, 2026 by ved1beta Loading…
3
MX QAT patch
#3553 opened Mar 27, 2026 by ved1beta Loading…
4
feat: move to uv first
#3545 opened Mar 25, 2026 by NanoCode012 Loading…
2 tasks done
13
fix: DPO tool role KeyError (#3217), dataset hash output_dir (#3303), config validators
#3538 opened Mar 23, 2026 by Edward-Zion-Saji Loading…
7
feature: Mixlora addition
#3535 opened Mar 22, 2026 by OnePunchMonk Loading…
25
MoE aux loss free
#3533 opened Mar 22, 2026 by winglian Loading…
16
fix: make prepare_context_parallel_inputs no-op
#3520 opened Mar 20, 2026 by NanoCode012 • Draft
2
feat: add sonicmoe fused lora support
#3519 opened Mar 20, 2026 by NanoCode012 • Draft
1
super nemo support scheduled_release This PR is slated for the upcoming release
#3508 opened Mar 18, 2026 by ved1beta Loading…
52
add textui scheduled_release This PR is slated for the upcoming release
#3505 opened Mar 17, 2026 by winglian • Draft
17
feat: add dataset weight support for weighted data mixing scheduled_release This PR is slated for the upcoming release
#3501 opened Mar 16, 2026 by madScientist10 Loading…
8 tasks done
6
Datasets dynamic work queue scheduled_release This PR is slated for the upcoming release
#3460 opened Mar 5, 2026 by Nero10578 Loading…
13
feat: add torchao int4, nf4, int8 LoRA QLoRA support
#3417 opened Feb 16, 2026 by NanoCode012 • Draft
Compatibility: torch 2.10 baseline and transformers>=5 interoperability
#3414 opened Feb 16, 2026 by cccat6 Loading…
7
feat: add auto-chunking support for streaming pretraining datasets
#3389 opened Feb 3, 2026 by madScientist10 Loading…
3 tasks done
8
Dynamic Fine-tuning
#3360 opened Jan 15, 2026 by salmanmohammadi Loading…
1
6
New "muonclip" Muon implementation (FSDP/DSZ3 compatible, faster, slow Muon clip support) waiting on upstream
#3265 opened Nov 14, 2025 by lhl Loading…
23
MoE Grouped MM support (5X+ MoE training perf gains) waiting on upstream
#3260 opened Nov 12, 2025 by lhl Loading…
17
Moe aux loss free waiting on upstream
#3259 opened Nov 11, 2025 by lhl Loading…
11
Allow muon optimizer with DeepSpeed Zero 1-2 waiting on upstream
#3258 opened Nov 11, 2025 by lhl Loading…
6
include tool in default message_property_mappings hold don't merge this yet
#3228 opened Oct 23, 2025 by winglian Loading…
3
Fix: Resolve torch_dtype correctly for AMP
#3189 opened Sep 27, 2025 by onesnep Loading…
10
fix: change docker miniconda install to workspace
#3172 opened Sep 23, 2025 by NanoCode012 • Draft
1
ProTip! Mix and match filters to narrow down what you’re looking for.