-
Notifications
You must be signed in to change notification settings - Fork 585
Issues: allenai/OLMo
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.
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
Author
Label
Projects
Milestones
Assignee
Sort
Issues list
Stage 2 checkpoints on huggingface
type/feature
An issue or pull request that introduces a new feature
#816
opened Mar 26, 2025 by
loftusa
scripts/convert_olmo_to_hf_new.py example usage mistakes
type/bug
An issue about a bug
#815
opened Mar 26, 2025 by
tutao0123
Some stage2 checkpoints load but not others
type/bug
An issue about a bug
#813
opened Mar 24, 2025 by
loftusa
OLMoConfigurationError: Checkpoint for step 0 already exists, use --save_overwrite to overwrite it
type/bug
An issue about a bug
#812
opened Mar 21, 2025 by
cs-pikachuu
Document GPU hours for the ptrained models
type/documentation
An issue or pull request related to documentation
#807
opened Mar 6, 2025 by
nlgranger
Why QK-layer norm before ROPE ?
type/question
An issue that's a question
#806
opened Mar 4, 2025 by
dhia680
Activations Exploding Across Layers
type/question
An issue that's a question
#797
opened Feb 15, 2025 by
c3-utsavdutta98
Optimizer and trainer states for OLMo-7B (Feb. 2024)
type/question
An issue that's a question
#796
opened Feb 13, 2025 by
rahuln
Tokenizer to be used for generation of data to .npy files
type/question
An issue that's a question
#791
opened Jan 29, 2025 by
WenJett
High CrossEntropy and Z Loss variance after loading from checkpoint
type/bug
An issue about a bug
#776
opened Jan 6, 2025 by
abhijangda
Generating training mix of OLMo2 from dolmino-mix
type/question
An issue that's a question
#775
opened Jan 5, 2025 by
Cy-47
tokenizer.encode function`s param add_special_tokens=False not work.
type/bug
An issue about a bug
#765
opened Dec 12, 2024 by
xiaohan2909
About eos_token_id in config file (20M, 1B)
type/question
An issue that's a question
#757
opened Nov 29, 2024 by
lllabmaster
Fail to load tokenizer for checkpoints
type/bug
An issue about a bug
#741
opened Oct 24, 2024 by
tresiwald
Error Encountered During Multi-Node Pretraining with Torchrun
type/bug
An issue about a bug
#737
opened Oct 21, 2024 by
Zehui127
RuntimeError: Triton Error [CUDA]: invalid device context
type/bug
An issue about a bug
#700
opened Aug 13, 2024 by
andymvp2018
Gflops computation is faulty for FSDP due to bug in
OLMo.num_params()
#695
opened Aug 7, 2024 by
AkshitaB
Olmo 0724 An issue about a bug
-hf checkpoints don't load the proper config when instantiating with OLMoForCausalLM
type/bug
#689
opened Aug 5, 2024 by
sarahwie
Model ladder has no documentation
type/documentation
An issue or pull request related to documentation
#683
opened Jul 31, 2024 by
IanMagnusson
mlp_ratio not adjusted in config if mlp_hidden_size is set
type/bug
An issue about a bug
#673
opened Jul 21, 2024 by
Muennighoff
Does global_train_batch_size support gradient accumulation?
type/question
An issue that's a question
#672
opened Jul 21, 2024 by
jinzhuoran
Is there explicitly instruction-following data in the version of Dolma used to train v1?
type/question
An issue that's a question
#658
opened Jul 15, 2024 by
john-hewitt
Previous Next
ProTip!
Adding no:label will show everything without a label.