Japanese Stable LM 2 Base 1.6B
Japanese Stable LM 2 Base 1.6B continues training the multilingual Stable LM 2 base on Japanese data. It provides a compact starting point for Japanese-language adaptation.
Research release
Model design and access
The model keeps the 24-layer architecture and Arcade100k tokenizer of its source model. The card distinguishes it from Japanese Stable LM 2 Instruct. Continued pretraining improves language coverage, while instruction behavior requires a separate adaptation stage.Japanese Stable LM 2 Base 1.6B official model card
Practical use
Practical guidance. The base is useful for Japanese domain experiments that need local control over training. Measure how its tokenizer handles the actual documents, including mixed Japanese and English text. Do not carry over the Instruct checkpoint's license merely because the two releases share a family name.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Access | Downloadable |
| Model identifiers | japanese-stablelm-2-base-1_6b |
| Context window | 4,096 tokens Published checkpoint sequence length; input and generated text share the window. Japanese Stable LM 2 Base 1.6B official model card |
| Parameters | 1,644,417,024.Japanese Stable LM 2 Base 1.6B official model card |
| Continued training | Japanese and English Wikipedia plus CulturaX.Japanese Stable LM 2 Base 1.6B official model card |
| License | Non-Commercial Research Community License, dated 8 May 2024.Japanese Stable LM 2 Base 1.6B license |
Related models
- Japanese Stable LM 2 Instruct 1.6B
Stability AI · Stable LM
Japanese Stable LM 2 Instruct 1.6B adapts the small Stable LM 2 family for Japanese instructions.
Launch date not confirmedDownloadable - Stable LM 2 1.6B
Stability AI · Stable LM
A 1.6B base language model pretrained on multilingual text and code.
Launch date not confirmedDownloadable