migtissera/Tess-M-v1.2

TEXT GENERATIONConcurrency Cost:2Model Size:34BQuant:FP8Ctx Length:32kPublished:Nov 23, 2023License:yi-34bArchitecture:Transformer Cold

Tess-M-v1.2 by migtissera is an experimental 34 billion parameter general-purpose large language model, built upon the Yi-34B-200K base architecture. This model is designed for broad applications, leveraging its substantial parameter count and a 32768 token context length for comprehensive language understanding and generation. It serves as an earlier iteration in the Tess series, intended for general language tasks.

Loading preview...

migtissera/Tess-M-v1.2: An Experimental General-Purpose LLM

Tess-M-v1.2 is an experimental 34 billion parameter large language model developed by migtissera, part of the Tess series. It is built upon the Yi-34B-200K base model, indicating a foundation designed for extensive context handling with its 32768 token context window.

Key Characteristics:

  • Base Model: Utilizes the Yi-34B-200K as its foundational architecture.
  • Parameter Count: Features 34 billion parameters, suitable for a wide range of complex language tasks.
  • Context Length: Supports a substantial context window of 32768 tokens, allowing for processing and generating longer texts.
  • Purpose: Designed as a general-purpose language model, aiming for broad applicability across various natural language processing tasks.

Important Note:

This version (v1.2) is explicitly marked as experimental and deprecated. Users are advised to use the stable release, Tess-M-v1.3, for production or more reliable applications.

Prompt Format:

The model expects a specific prompt structure:

SYSTEM: <ANY SYSTEM CONTEXT>
USER: 
ASSISTANT: