Skip to content

Issues: OpenAccess-AI-Collective/axolotl

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

189 Open 388 Closed
189 Open 388 Closed
Author
Filter by author
Label
Filter by label
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Milestones
Filter by milestone
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

Train FAILED. Crashed while training with SIGTERM bug Something isn't working
#1670 opened May 29, 2024 by RodriMora
6 of 8 tasks
Llama Reserved Tokens Initialization enhancement New feature or request
#1666 opened May 28, 2024 by cinjon
5 tasks done
Not enough information on pre-tokenized dataset.
#1661 opened May 27, 2024 by rumbleFTW
3 tasks done
Pulling the image from Docker retuns with ''unauthorized: authentication required'' bug Something isn't working
#1658 opened May 25, 2024 by Fischherboot
6 of 8 tasks
Llama3 Lora training fails to output and save bug Something isn't working
#1650 opened May 23, 2024 by austinm1120
6 of 8 tasks
dataset type sharegpt no longer works bug Something isn't working
#1649 opened May 22, 2024 by thepowerfuldeez
6 of 8 tasks
8
DPO Prompt Strategies only support single-turn and will fail silently on multi-turn datasets bug Something isn't working
#1645 opened May 21, 2024 by bjoernpl
6 of 8 tasks
1
Llama 3 & Mistral LoRA Examples Error (needs eval_sample_packing: False) bug Something isn't working
#1644 opened May 21, 2024 by VelocityRa
6 of 8 tasks
Llama 3 8b OOM with GaLore on 2x A100s (Mistral 7b is fine?) bug Something isn't working possibly_solved
#1641 opened May 19, 2024 by e-p-armstrong
6 of 8 tasks
4
Support RecurrentGemma enhancement New feature or request
#1637 opened May 17, 2024 by julien-blanchon
5 tasks done
1
Error Finetuning CodeQwen 1.5 7B: Column 1 named token_type_ids expected length 113 but got length 112 bug Something isn't working waiting for reporter
#1632 opened May 16, 2024 by artemdinaburg
6 of 8 tasks
4
Auto resume from checkpoint looks for "trainer_state.json", but no file is generated bug Something isn't working waiting for reporter
#1631 opened May 16, 2024 by l3utterfly
5 of 8 tasks
1
Llama-3 DPO is broken in main bug Something isn't working possibly_solved waiting for reporter
#1624 opened May 15, 2024 by maziyarpanahi
6 of 8 tasks
13
DPO Trainer Incorrectly Inserts BoS Before Chosen and Rejected Prompts for Llama3 bug Something isn't working
#1616 opened May 14, 2024 by Catgat
6 of 8 tasks
Preprocess with debug gives error. bug Something isn't working possibly_solved
#1599 opened May 7, 2024 by amitagh
6 of 8 tasks
2
GaLore_AdamW_8bit_Layerwise issue: AttributeError: 'Parameter' object has no attribute 'register_post_accumulate_grad_hook' bug Something isn't working
#1598 opened May 6, 2024 by bhavnicksm
6 of 8 tasks
Recent RunPod Axolotl error bug Something isn't working possibly_solved
#1596 opened May 6, 2024 by drummerv
6 of 8 tasks
7
8-Bit DoRA training with FSDP doesn't work, but 4-bit QDoRA does / peft_use_dora is ignored? bug Something isn't working
#1589 opened May 4, 2024 by kalomaze
1 task done
3
import flash_attn_2_cuda as flash_attn_cuda fails bug Something isn't working
#1588 opened May 3, 2024 by Dyke-F
6 of 8 tasks
2
Challenges in Integrating New ML Tools into Axolotl Due to Manual Code Modifications enhancement New feature or request
#1586 opened May 3, 2024 by seungduk-yanolja
5 tasks done
Axolotl install and train is broken. bug Something isn't working
#1571 opened Apr 26, 2024 by amitagh
6 of 8 tasks
3
Phi 3 support enhancement New feature or request
#1562 opened Apr 23, 2024 by fakerybakery
5 tasks done
4
Can't finetune Mixtral model bug Something isn't working
#1554 opened Apr 20, 2024 by hahmad2008
6 of 8 tasks
3
OPRO, DPO don't work with Mixtral-8x22B. FSDP + QLORA & bigstral-ds-zero3 bug Something isn't working
#1534 opened Apr 18, 2024 by 0-hero
6 of 8 tasks
1
OLMo Support enhancement New feature or request
#1533 opened Apr 18, 2024 by fakerybakery
1
ProTip! Find all open issues with in progress development work with linked:pr.