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

220 Open 464 Closed
220 Open 464 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

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
3
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
1
RuntimeError: CUDA error: unknown error When attempting to fine-tune llama-3 models bug Something isn't working waiting for reporter
#2071 opened Nov 17, 2024 by cjfreeze
7 of 8 tasks
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
Mlflow duplicate logging bug Something isn't working
#2063 opened Nov 15, 2024 by jsh2581
6 of 8 tasks
1
Axolotl hanging on bench evals with fsdp bug Something isn't working under review
#2058 opened Nov 14, 2024 by bsc001
6 of 8 tasks
@NanoCode012
8
Logging behavior since GA fix bug Something isn't working under review waiting for reporter
#2004 opened Oct 30, 2024 by ccdv-ai
6 of 8 tasks
@NanoCode012
10
Support for Sequence / Context Parallelism enhancement New feature or request
#1972 opened Oct 15, 2024 by dwzhu-pku
5 tasks done
1
Flash attention and multipack failing for qwen and mistral bug Something isn't working waiting for reporter
#1966 opened Oct 12, 2024 by tiger241
6 of 8 tasks
@bursteratom
12
Add resize_token_embeddings feature enhancement New feature or request waiting for reporter
#1965 opened Oct 11, 2024 by ccdv-ai
5 tasks done
5
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
Llama will not save properly bug Something isn't working waiting for reporter
#1947 opened Oct 6, 2024 by mfirth-truffle
6 of 8 tasks
@bursteratom
3
Feature Request: Adding dataset deduplication process enhancement New feature or request
#1946 opened Oct 5, 2024 by Weyaxi
5 tasks done
1
5
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
Cannot install on Google Colab bug Something isn't working wip
#1933 opened Sep 27, 2024 by benjamin-marie
5 of 8 tasks
1
7
Using two 8xH100 nodes to train. encounter error bf16 requested, but AMP is not supported on this GPU. Requires Ampere series or above. bug Something isn't working waiting for reporter
#1924 opened Sep 23, 2024 by michaellin99999
6 of 8 tasks
8
mistrall small support enhancement New feature or request
#1922 opened Sep 21, 2024 by win4r
5 tasks done
1
Gemma 2 chat template inserts eos_token after every chat turn bug Something isn't working waiting for reporter
#1921 opened Sep 20, 2024 by Nero10578
6 of 8 tasks
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
Unable to load ORPO dataset in a *.json file bug Something isn't working
#1868 opened Aug 26, 2024 by SicariusSicariiStuff
6 of 8 tasks
@bursteratom
3
ORPO results in Cannot flatten integer dtype tensors bug Something isn't working
#1838 opened Aug 20, 2024 by maziyarpanahi
6 of 8 tasks
4
inst chat jinja template does not match prompt format used while training with conversation: mistral bug Something isn't working
#1832 opened Aug 18, 2024 by nyxkrage
6 of 8 tasks
ProTip! no:milestone will show everything without a milestone.