JungZoona/T3Q-qwen2.5-14b-v1.2-e2

Hugging Face
TEXT GENERATIONConcurrent Unit Cost:1Model Size:14.8BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Mar 14, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Warm

JungZoona/T3Q-qwen2.5-14b-v1.2-e2 is a 14 billion parameter causal language model, post-trained by JungZoona based on Qwen/Qwen2.5-14B-Instruct-1M. This model utilizes LoRA fine-tuning (8-4-0.0001-cosine-32-16) with a specific dataset (train_data_v1.2) to enhance its instruction-following capabilities. It is designed for general-purpose conversational AI and instruction-based tasks, building upon the robust Qwen2.5 architecture.

Loading preview...

Model Overview

JungZoona/T3Q-qwen2.5-14b-v1.2-e2 is a 14 billion parameter language model developed by JungZoona. It is a post-trained version of the established Qwen/Qwen2.5-14B-Instruct-1M model, indicating a focus on refining its performance for specific applications or instruction adherence.

Training Details

This model has undergone further fine-tuning using a LoRA (Low-Rank Adaptation) technique. The specific LoRA configuration used is 8-4-0.0001-cosine-32-16, applied with a dataset referred to as train_data_v1.2. This post-training process aims to adapt the base Qwen2.5-14B-Instruct-1M model to potentially improve its responses or align it more closely with certain task requirements.

Key Characteristics

  • Base Model: Qwen/Qwen2.5-14B-Instruct-1M
  • Parameter Count: 14 billion parameters
  • Fine-tuning Method: LoRA (8-4-0.0001-cosine-32-16)
  • Training Data: Utilized train_data_v1.2 for post-training

Intended Use

Given its foundation in an instruction-tuned model and subsequent fine-tuning, T3Q-qwen2.5-14b-v1.2-e2 is suitable for a range of applications requiring instruction-following and conversational abilities. Developers can leverage this model for tasks where a refined version of the Qwen2.5-14B-Instruct-1M model is beneficial, particularly if their use case aligns with the train_data_v1.2 used for its post-training.