didula-wso2/exp_24_0_juliasft_16bit_vllm

TEXT GENERATIONConcurrency Cost:1Model Size:7.6BQuant:FP8Ctx Length:32kPublished:Jan 6, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

The didula-wso2/exp_24_0_juliasft_16bit_vllm is a 7.6 billion parameter Qwen2.5-Coder-7B-Instruct model, fine-tuned by didula-wso2. This model was optimized for faster training using Unsloth and Huggingface's TRL library. It is designed for tasks requiring a capable language model with efficient training characteristics.

Loading preview...

Model Overview

The didula-wso2/exp_24_0_juliasft_16bit_vllm is a 7.6 billion parameter language model, fine-tuned by didula-wso2. It is based on the Qwen2.5-Coder-7B-Instruct architecture and utilizes a 131072 token context length. This model was specifically fine-tuned using Unsloth and Huggingface's TRL library, which enabled a 2x faster training process compared to standard methods.

Key Characteristics

  • Base Model: Qwen2.5-Coder-7B-Instruct
  • Parameter Count: 7.6 billion parameters
  • Context Length: 131072 tokens
  • Training Optimization: Fine-tuned with Unsloth and Huggingface's TRL library for accelerated training.
  • License: Apache-2.0

Use Cases

This model is suitable for applications that benefit from a Qwen2.5-based architecture with a large context window, particularly where efficient fine-tuning was a development priority. Its origins from a 'Coder' base suggest potential strengths in code-related tasks, though specific performance metrics are not detailed in the provided information.