Downloads · 30 days
0
preln-llama-400M-L24-lr5e-4
aspect-ratio-scaling/preln-llama-400M-L24-lr5e-4
preln-llama-400M-L24-lr5e-4 is a machine learning model from aspect-ratio-scaling. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for olmo-core.
This repository stores the raw OLMo-core checkpoint files uploaded from the original training output directory.
Access
Public
Updated Aug 9, 2026
Repo size
4.9 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.distcp4.9 GB · 100%
At a glance
- Library
- olmo-core
- Access
- Public
- Created
- Aug 9, 2026
- Updated
- Aug 9, 2026
- SHA
- e85a01f4
Model card
From the Hugging Face model README
preln-llama-400M-L24-lr5e-4
This repository stores the raw OLMo-core checkpoint files uploaded from the original training output directory.
Contents
- Local source directory:
/lustre/fast/fast/wangk/ckpt/depthbench/pretrain-llama-350M-standard-lr5e-4 - Included step directories:
step0, step7600 - Latest step:
step7600 - Included tokenizer directory:
tokenizer/ - Selected file count:
218 - Selected upload size:
4.54 GiB
Important
This is a raw distributed checkpoint, not a Hugging Face from_pretrained() export.
To use it, download the repository and load it with OLMo-core utilities such as load_model_and_optim_state().
Download
from huggingface_hub import snapshot_download
local_dir = snapshot_download(repo_id="aspect-ratio-scaling/preln-llama-400M-L24-lr5e-4")
print(local_dir)
Notes
wandb/is not included in this upload by default.- The original checkpoint layout is preserved for reproducibility and later conversion if needed.
- Library
- olmo-core
- Created
- Aug 9, 2026
- Updated
- Aug 9, 2026