Skip to content

Issues: NVIDIA/NeMo

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

Clear current search query, filters, and sorts
56 Open 2,399 Closed
56 Open 2,399 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

Error in saving nemo checkpoint with Llama-3.1-70B SFT. /opt/NeMo/nemo/utils/callbacks/nemo_model_checkpoint.py bug Something isn't working stale
#12157 opened Feb 12, 2025 by songwang41
1
Possible bug in ASRDecoderTimeStamps - math.ceil on fractional tokens_per_chunk leads to timestamps displacements on long files bug Something isn't working
#11604 opened Dec 15, 2024 by bene-ges
5
when i use container to do sft for any model, it has context not found error bug Something isn't working
#11825 opened Jan 11, 2025 by munger1985
4
Annoyed NeMo User Rant
#11835 opened Jan 13, 2025 by yonas-g
@ko3n1g
1
Broken offline mode of NeMo bug Something isn't working
#11899 opened Jan 20, 2025 by maxstrobel
1
2
Hybrid Sharding support with FSDP stale
#11946 opened Jan 23, 2025 by Teng-xu
@okuchaiev
2
Pickling error when trying to save checkpoints with custom checkpointIO bug Something isn't working
#11955 opened Jan 24, 2025 by jdnurme
7
Add option for prefetch factor of data loader to config stale
#11977 opened Jan 28, 2025 by shengshiqi-google
@okuchaiev
2
ASR: How to convert .ckpt to nemo correctly? ASR bug Something isn't working
#12003 opened Jan 31, 2025 by ican24
3
llava-like dataset implementation "LazySupervisedDataset" likely fails to handle large dataset
#12034 opened Feb 3, 2025 by bernardhan33
@okuchaiev
3
AttributeError: 'HFDatasetDataModule' object has no attribute 'tokenizer' bug Something isn't working
#12080 opened Feb 6, 2025 by j40903272
@akoumpa
2
[QST] How to set MoE-specific TP size in recipe? stale
#12103 opened Feb 8, 2025 by umiswing
8
Bug when generating confidence scores with timestamps for a buffered rnnt model ASR bug Something isn't working
#11456 opened Dec 3, 2024 by aanchan
5
Fail to convert trained checkpoint to HF format bug Something isn't working stale
#12124 opened Feb 10, 2025 by Zhihan1996
2
NeMo is not friendly to HF compatibility.
#12166 opened Feb 13, 2025 by dyang67
@akoumpa
4
I am trying to train the FastConformer 120M model from scratch, but it is not converging? ASR help wanted Extra attention is needed
#12167 opened Feb 13, 2025 by PhamDangNguyen
@nithinraok
9
Update TE version for support of pad_between_seqs=True bug Something isn't working
#12174 opened Feb 13, 2025 by cyanguwa
@ko3n1g
HiFiGAN Finetune "Cannot re-initialize CUDA in forked subprocess." bug Something isn't working
#12178 opened Feb 13, 2025 by Fournogo
@blisc
1
Bloated pre-requirements
#12188 opened Feb 14, 2025 by ceyxasm
@nithinraok
1
Support configuration of num_workers and max_samples_per_sequence in llava_next_pretrain
#12195 opened Feb 14, 2025 by bernardhan33
@abhinavg4 @yaoyu-33
1
Checkpointing randomly fails bug Something isn't working
#12203 opened Feb 15, 2025 by aflah02
@hemildesai
2
Pre-Training Neva under pipeline parallel set to 2. bug Something isn't working
#12205 opened Feb 16, 2025 by takuya576
@yaoyu-33
1
loss divergence when CP>1 and MBS>1 bug Something isn't working
#12210 opened Feb 17, 2025 by hawkoli1987
@xrennvidia
2
There exists a 2.4B training parameter during fine-tuned training of a 70B model, where did this parameter come from?
#12213 opened Feb 17, 2025 by echo-valor
@akoumpa
1
HUGE Inconsistency between logged tokens_per_second_per_GPU and actual wall time and Global Step is Not Monotonically Increasing bug Something isn't working
#12727 opened Mar 21, 2025 by aflah02
ProTip! Add no:assignee to see everything that’s not assigned.