Edit model card

Model Card for Model ID

This is Jamba quantized with bitsandbytes to 4-bit. It is based on ai21labs/Jamba-v0.1.

Model Details

The model was created using a recipe detailed in this article: Jamba: The New Hybrid Transformer/Mamba

No modules have been "skipped" for the quantization.

Model Description

  • Developed by: The Salt
  • Model type: Causal
  • Language(s) (NLP): English
  • License: Apache 2.0
Downloads last month
24
Safetensors
Model size
26.9B params
Tensor type
F32
BF16
U8