Downloads · 30 days
0
qwen2.5-0.5b-sft-trial4
AR2479/qwen2.5-0.5b-sft-trial4
qwen2.5-0.5b-sft-trial4 is a machine learning model from AR2479. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for peft. The card lists the license as apache-2.0.
This model is a Supervised Fine-Tuned (SFT) derivative of the Qwen/Qwen2.5-0.5B-Instruct base model. It represents the best-performing trial out of five distinct SFT configurations, chosen for demonstrating the strong…
Access
Public
Updated Jun 5, 2026
Repo size
1000 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.json1000 MB · 100%
At a glance
- Library
- peft
- License
- apache-2.0
- Access
- Public
- Created
- Jun 5, 2026
- Updated
- Jun 5, 2026
- SHA
- 1bd2e015
Base models
Model card
From the Hugging Face model README
Qwen2.5-0.5B SFT Trial 4
Model Overview
This model is a Supervised Fine-Tuned (SFT) derivative of the Qwen/Qwen2.5-0.5B-Instruct base model. It represents the best-performing trial out of five distinct SFT configurations, chosen for demonstrating the strongest overall improvement over the base model while maintaining stable training behavior.
Training Hyperparameters
The model was fine-tuned using Low-Rank Adaptation (LoRA) with the following configuration:
- LoRA Rank (r): 16
- Learning Rate: 2e-4
- Epochs: 2
- Batch Size: 2
Evaluation Metrics
The model was evaluated against the base model using BLEU and BERTScore to measure structural and semantic generation quality.
| Model | BLEU | BERTScore |
|---|---|---|
| Base Model | 0.0496 | 0.8871 |
| SFT Trial 4 | 0.0739 | 0.8943 |
Usage
This model is intended to be used as a causal language model for natural language processing tasks. Due to the LoRA fine-tuning, it exhibits improved adherence to specific instruction formats compared to the base model.
- Library
- peft
- License
- apache-2.0
- Created
- Jun 5, 2026
- Updated
- Jun 5, 2026