Skip to content

Issues: pytorch/torchtune

v0.5.0 tracker
#2008 opened Nov 14, 2024 by joecummings
Open
Testing tracker
#1890 opened Oct 23, 2024 by felipemello1
Open
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

214 Open 521 Closed
214 Open 521 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

OpenCoder, Request for adding a model
#2058 opened Nov 23, 2024 by insop
1
add dataset.packed support for llama 3.2 vision
#2057 opened Nov 23, 2024 by felipemello1
@pbontrager
improve error msg for packed being incompatible
#2056 opened Nov 23, 2024 by felipemello1
error fast when tokenizer=null and dataset.packed=True
#2055 opened Nov 23, 2024 by felipemello1
1
Clean up PyTorch profiler usage in recipes + configs
#2053 opened Nov 22, 2024 by joecummings
improve _optim_ckpt_wrapper so it is a drop in replacement of optimizer best practice Things we should be doing but aren't community help wanted We would love the community's help completing this issue
#2052 opened Nov 22, 2024 by felipemello1
2
Recipe documentation
#2050 opened Nov 22, 2024 by joecummings
1
FullModelHFCheckpointer saved checkpoint isn't compatible with Huggingface transformers model loading high-priority
#2048 opened Nov 21, 2024 by vancoykendall
@joecummings @felipemello1
7
Change llama 3 and 2 configs to use hf ckpt instead of meta ckpt community help wanted We would love the community's help completing this issue
#2045 opened Nov 21, 2024 by felipemello1
PPO slow start
#2044 opened Nov 21, 2024 by felipemello1
@SalmanMohammadi
[BUG] Linter is not capturing issues, like unused arguments bug Something isn't working
#2039 opened Nov 20, 2024 by felipemello1
[BUG] decoder_lora and fusion_lora is a no-op in llama 3.2 vision bug Something isn't working
#2038 opened Nov 20, 2024 by felipemello1
@pbontrager
1
Apply gradient accumulation fix to DPO/PPO recipes
#2037 opened Nov 20, 2024 by SalmanMohammadi
@SalmanMohammadi
update tests and readme cu121 --> cu124
#2036 opened Nov 20, 2024 by felipemello1
1
Add NLL and DPOP weighting to DPO losses
#2032 opened Nov 20, 2024 by RdoubleA
Error: Are you sure that 'metric_logging' is importable from module 'torchtune.utils'?
#2031 opened Nov 20, 2024 by zhoufy20
3
Add config for distilling Llama3.1 70B into 1B
#2029 opened Nov 19, 2024 by joecummings
Integrate Lora fine-tuned model with HF
#2025 opened Nov 19, 2024 by GLorenzo679
6
Hown to config generation length on Android device
#2024 opened Nov 19, 2024 by HaomingSong
1
Allow recipes to work without a LR scheduler community help wanted We would love the community's help completing this issue
#2021 opened Nov 18, 2024 by felipemello1
1
Does torchtune support multi-node training? question Further information is requested
#2018 opened Nov 18, 2024 by tginart
3
Add comments explaining why 128011 is skipped for Llama3 Tokenizer documentation Improvements or additions to documentation
#2014 opened Nov 15, 2024 by RdoubleA
Add custom dataset section in Basics in live docs community help wanted We would love the community's help completing this issue documentation Improvements or additions to documentation
#2012 opened Nov 15, 2024 by RdoubleA
Warmup starts from lr=0
#2010 opened Nov 14, 2024 by cfhammill
9
v0.5.0 tracker
#2008 opened Nov 14, 2024 by joecummings
1 of 2 tasks
ProTip! Type g i on any issue or pull request to go back to the issue listing page.