orkungedik/ege-v1.3

VISIONConcurrent Unit Cost:2Model Size:27BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jun 26, 2026Architecture:Transformer Featherless Exclusive Cold

orkungedik/ege-v1.3 is a 27 billion parameter language model fine-tuned by orkungedik using the TRL framework. This model is a fine-tuned version of an unspecified base model, optimized through supervised fine-tuning (SFT). It features a 32768 token context length, making it suitable for tasks requiring extensive contextual understanding. Its primary strength lies in its fine-tuned nature, suggesting specialized performance for specific applications, though the exact domain is not specified.

Loading preview...

Model Overview

orkungedik/ege-v1.3 is a 27 billion parameter language model developed by orkungedik. This model has been fine-tuned using the TRL (Transformer Reinforcement Learning) framework, specifically employing Supervised Fine-Tuning (SFT) as its training procedure. While the base model for this fine-tuning is not specified, the process indicates an optimization for particular tasks or performance characteristics.

Key Characteristics

  • Parameter Count: 27 billion parameters.
  • Context Length: Supports a substantial context window of 32768 tokens, enabling it to process and generate longer sequences of text.
  • Training Framework: Fine-tuned with TRL, a library for transformer reinforcement learning.
  • Training Method: Utilizes Supervised Fine-Tuning (SFT).

Usage

The model can be used for text generation tasks. A quick start example demonstrates how to use the transformers pipeline for generating responses to prompts, with a maximum new token generation of 128 tokens.

Technical Details

The model was trained with specific versions of key frameworks:

  • TRL: 0.24.0
  • Transformers: 5.12.1
  • Pytorch: 2.9.0
  • Datasets: 4.3.0
  • Tokenizers: 0.22.2

This fine-tuned model is designed for applications that can leverage its large parameter count and extended context window, particularly where SFT-based optimizations are beneficial.