Downloads · 30 days
8
17% of all-time downloads
qwen3-4B-base-luna-kv-tags-explained-prepend-step-epoch-2
JackHsieh/qwen3-4B-base-luna-kv-tags-explained-prepend-step-epoch-2
qwen3-4B-base-luna-kv-tags-explained-prepend-step-epoch-2 is a machine learning model from JackHsieh. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
The best Qwen3-4B-base checkpoint CPT'd against a prepended key+explanation+thought (so far).
All-time downloads
47
Public
Parameters
4B
8.1 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors8 GB · 100%
At a glance
- License
- apache-2.0
- Model type
- qwen3
- Access
- Public
- Created
- Aug 13, 2026
- Updated
- Aug 13, 2026
- SHA
- 2b5e5181
Base models
Model card
From the Hugging Face model README
qwen3-4B-base-luna-kv-tags-explained-prepend-step-epoch-2
The best Qwen3-4B-base checkpoint CPT'd against a prepended key+explanation+thought (so far).
Run: wandb · epoch 2 (step 676), lr 3e-6, replay 0.1, batch 32.
Thoughts: JackHsieh/luna-reason-only.k-8.statml-arxiv-qwen3.qwen3-ids.kv-tags-explained,
prepended to the document so prefix and suffix stay adjacent.
<|note|> / <|/note|> (ids 151669 / 151670) were added to the vocab and trained.
- Type
- qwen3
- License
- apache-2.0
- Created
- Aug 13, 2026
- Updated
- Aug 13, 2026