LLM Reference

DeciLM 7B

About

DeciLM-7B is a cutting-edge large language model developed by Deci AI, featuring 7.04 billion parameters. This model incorporates an advanced transformer decoder architecture, utilizing variable Grouped-Query Attention (GQA) to achieve high accuracy and efficiency. The architecture is fine-tuned using Deci's proprietary Neural Architecture Search technology, AutoNAC, for optimal performance. Capable of handling sequences of up to 8192 tokens, DeciLM-7B outperforms similar or larger models across various benchmarks. An instruction-tuned version, DeciLM-7B-instruct, further enhances its capabilities, especially for instruction-following tasks. It is released under the Apache 2.0 license, making it suitable for both commercial and research use.

Capabilities

MultimodalFunction CallingTool UseJSON Mode

Providers(1)

ProviderInput (per 1M)Output (per 1M)Type
Azure OpenAI
Provisioned

Specifications

FamilyDeciLM
Parameters7B
ArchitectureDecoder Only
Specializationgeneral