TRAC-FLVN/stratagem-instruct-nemo-non-adapated

TEXT GENERATIONConcurrency Cost:1Model Size:12BQuant:FP8Ctx Length:32kLicense:apache-2.0Architecture:Transformer Open Weights Cold

The TRAC-FLVN/stratagem-instruct-nemo-non-adapated is a 12 billion parameter instruction-tuned Mistral-based causal language model developed by TRAC-FLVN. It features a 32768 token context length and was fine-tuned using Unsloth and Huggingface's TRL library, enabling faster training. This model is designed for general instruction-following tasks, leveraging its Mistral architecture for robust performance.

Loading preview...

Model Overview

The TRAC-FLVN/stratagem-instruct-nemo-non-adapated is a 12 billion parameter language model developed by TRAC-FLVN. It is built upon the Mistral architecture and has been instruction-tuned to enhance its ability to follow diverse prompts. A notable aspect of its development is the use of Unsloth and Huggingface's TRL library, which facilitated a 2x faster training process compared to conventional methods.

Key Characteristics

  • Architecture: Based on the Mistral model family.
  • Parameter Count: 12 billion parameters.
  • Context Length: Supports a substantial context window of 32768 tokens.
  • Training Efficiency: Fine-tuned with Unsloth for accelerated training.

Intended Use Cases

This model is suitable for a broad range of instruction-following applications where a balance of performance and efficiency is desired. Its large context window makes it effective for tasks requiring extensive input or generating longer, coherent responses.