Skip to content

Issues: axolotl-ai-cloud/axolotl

Improve Adapter/LoRA handling
#1095 opened Jan 11, 2024 by winglian
Open 3
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

200 Open 509 Closed
200 Open 509 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
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

Trainer Builder Does not Use Custom Jinja Template bug Something isn't working
#2218 opened Dec 23, 2024 by NJordan72
6 of 8 tasks
1
deepspeed zero1 zero2 zero 3 out of memory when big model bug Something isn't working
#2217 opened Dec 23, 2024 by sankexin
6 of 8 tasks
1
Accelerate v1.2.1 Causes Consistent Errors bug Something isn't working
#2215 opened Dec 23, 2024 by williambarberjr
6 of 8 tasks
1
max_grad_norm doesn't appear to be clipping gradients bug Something isn't working
#2214 opened Dec 22, 2024 by DevonPeroutky
6 of 8 tasks
"RuntimeError: Invalid device argument : did you call init? "When setting CUDA_VISIBLE_DEVICES bug Something isn't working waiting for reporter
#2199 opened Dec 18, 2024 by zhanghanxing2022
6 of 8 tasks
2
load_from_disk for rl tpye training enhancement New feature or request
#2192 opened Dec 15, 2024 by leeparkuky
5 tasks done
1
'AdamW' object has no attribute 'optim_bits' bug Something isn't working waiting on upstream wip
#2191 opened Dec 15, 2024 by e-p-armstrong
1 task done
@bursteratom
8
APOLLO optimizer enhancement New feature or request
#2175 opened Dec 11, 2024 by fblgit
5 tasks done
1
When starting with DPO datasets, failed error with TypeError. bug Something isn't working waiting for reporter
#2174 opened Dec 11, 2024 by Yuto-24
6 of 8 tasks
1
Error During Model Saving QLORA + FSDP bug Something isn't working waiting on upstream
#2149 opened Dec 7, 2024 by ghsama
6 of 8 tasks
2
7
Show sample batch content enhancement New feature or request
#2145 opened Dec 7, 2024 by fzyzcjy
5 tasks done
2
Support ORPO/DPO Liger losses (and LigerORPOTrainer) enhancement New feature or request wip
#2141 opened Dec 6, 2024 by ccdv-ai
5 tasks done
@NanoCode012
Poential memory leak for axolotl v0.5.2 pretrain streaming datasets with liger kernel bug Something isn't working
#2108 opened Nov 30, 2024 by deter3
6 of 8 tasks
@bursteratom
4
Various bugs with ORPO bug Something isn't working
#2105 opened Nov 26, 2024 by ccdv-ai
6 of 8 tasks
2
Mistral Nemo LoRA training has super high grad_norm bug Something isn't working
#2095 opened Nov 21, 2024 by Nero10578
6 of 8 tasks
6
chat_template masking is broken with Mistral Small (possibly others) bug Something isn't working under review
#2089 opened Nov 19, 2024 by kubernetes-bad
6 of 8 tasks
@NanoCode012
2
Deepspeed zero3 + LoRA: RuntimeError: Only Tensors of floating point and complex dtype can require gradients bug Something isn't working waiting on upstream wip
#2068 opened Nov 16, 2024 by bursteratom
6 of 8 tasks
1
Support for Sequence / Context Parallelism enhancement New feature or request
#1972 opened Oct 15, 2024 by dwzhu-pku
5 tasks done
1
Should tokenizer_legacy be default as false? enhancement New feature or request under review
#1955 opened Oct 10, 2024 by tongyx361
5 tasks done
@bursteratom
fix_untrained_tokens doesn't work with zero-3 bug Something isn't working
#1944 opened Oct 4, 2024 by winglian
6 of 8 tasks
1
mistrall small support enhancement New feature or request
#1922 opened Sep 21, 2024 by win4r
5 tasks done
1
Different training losses when flash_attention is on/off bug Something isn't working
#1918 opened Sep 18, 2024 by zhangchen-xu
6 of 8 tasks
@bursteratom
6
pretrain doesn't work on json\jsonl bug Something isn't working
#1895 opened Sep 5, 2024 by SicariusSicariiStuff
6 of 8 tasks
1
Training with a large json dataset (>650K) throw error:pyarrow.lib.ArrowInvalid: offset overflow while concatenating arrays bug Something isn't working
#1888 opened Sep 3, 2024 by bofei5675
6 of 8 tasks
1
MixLoRA finetuning enhancement New feature or request
#1880 opened Aug 28, 2024 by winglian
5 tasks done
ProTip! Adding no:label will show everything without a label.