Skip to content

Issues: huggingface/nanotron

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

48 Open 45 Closed
48 Open 45 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

Invalid Usage of NCCL Library
#265 opened Dec 17, 2024 by awsankur
4
Support custom logging API for wandb alternatives (ex. ClearML)
#264 opened Dec 12, 2024 by arcyleung
AllReduce taking extra long for ShardedCrossEntropy help wanted Extra attention is needed
#262 opened Dec 10, 2024 by NouamaneTazi
Fix DDP with Zero1 help wanted Extra attention is needed
#261 opened Dec 10, 2024 by NouamaneTazi
Optimize memory in loss consumption help wanted Extra attention is needed
#259 opened Dec 4, 2024 by NouamaneTazi
FileNotFoundError: No files matching "datasets/fineweb-edu-dedup/*.ds" found in /home/nanotron/datasets/fineweb-edu-dedup
#257 opened Dec 2, 2024 by sankexin
2
the generations not right
#254 opened Nov 28, 2024 by sankexin
2
ValueError: ('The number of process requires to run all replicas (8)', 'must be equal to the world size (4).')
#250 opened Nov 27, 2024 by sankexin
2
Sync weights across dp dimension during consistency validation
#242 opened Nov 15, 2024 by gritukan
convert_nt_to_hf is broken for non-interleaved RoPE
#241 opened Nov 15, 2024 by gritukan
1
Support for AMD GPUs
#240 opened Nov 9, 2024 by YuTian8328
Nanoset reuses same shuffling for repeated data (epochs)
#237 opened Oct 31, 2024 by Lauler
1
CUDA_DEVICE_MAX_CONNECTIONS
#236 opened Oct 30, 2024 by jeromeku
1
ImportError: cannot import name 'LoggingArgs' from partially initialized module 'nanotron.config.config error (circular imports)
#235 opened Oct 24, 2024 by Akhilvallala2023
Learning rate restart broken with Nanoset? bug Something isn't working
#233 opened Sep 25, 2024 by Pclanglais
13
[feature] Use Unified Sequence Parallel (USP) instead of Ring attention
#226 opened Sep 1, 2024 by feifeibear
Use different positional embeddings
#225 opened Aug 28, 2024 by aflah02
2
Will audio input training be supported in the future? enhancement New feature or request
#210 opened Jul 26, 2024 by zira-wang
multi-node pp hang when enable gradient accumulation
#209 opened Jul 24, 2024 by yuuxiaooqingg
4
[Bug] Missing _is_using_mup when resume checkpoint bug Something isn't working good first issue Good for newcomers help wanted Extra attention is needed
#198 opened Jun 14, 2024 by xrsrke
1
Add Debug utility to be able to preview first samples used for training good first issue Good for newcomers help wanted Extra attention is needed
#184 opened May 27, 2024 by NouamaneTazi
PyTorch profiler is unable to serialize numpy datatypes sometimes inserted as process group ranks
#177 opened May 22, 2024 by hatanp
1
We don't save checkpoint after training ends good first issue Good for newcomers help wanted Extra attention is needed Low Priority
#163 opened May 6, 2024 by NouamaneTazi
out of memory for continuing pretraining llama3-8B
#161 opened May 3, 2024 by ckzbullbullet
5
moe in src and load balancing losses
#159 opened May 2, 2024 by haeggee
2
ProTip! Type g i on any issue or pull request to go back to the issue listing page.