Shivik-Code-1B / README.md
theaicompany02's picture
Initial release: Shivik-Code-1B
ee7c543 verified
metadata
language:
  - en
license: apache-2.0
library_name: transformers
tags:
  - code
  - causal-lm
  - shivik

Shivik-Code-1B

Shivik-Code-1B is a 1B-parameter code-focused causal language model from the Shivik family.

Model Overview

  • Architecture: Decoder-only Transformer
  • Parameters: ~1B
  • Domain: Code generation and understanding
  • Training: Custom-trained

Intended Use

  • Code generation
  • Code reasoning
  • Lightweight local inference
  • Research

Limitations

  • Not instruction-aligned by default
  • May hallucinate or produce incorrect code

License

Apache 2.0