Gemma 3 Lithuanian Weather Caption LoRA — Run B
This is the second LoRA adapter fine-tuned for Lithuanian weather-focused image captioning.
This version was trained with fewer epochs than the first run to reduce the risk of overfitting on the small dataset.
The model was trained to describe the weather in an image using short, simple Lithuanian sentences.
Task
Given an image, the model generates a short Lithuanian caption focused on:
- cloudiness
- sunlight
- precipitation
- visibility
- time of day
- general weather conditions
Base model
unsloth/gemma-3-4b-it
Dataset
Dataset: Matas5/GMM_team_task
Training examples used: 103
The dataset contains images with Lithuanian weather captions. Captions were standardized to focus mainly on weather conditions rather than unrelated objects.
Training setup
Training method: LoRA fine-tuning with Unsloth
Model loading: 4-bit quantized base model
Fine-tuning type: PEFT / LoRA adapter
Run version: B
Purpose of this run: less training, lower overfitting risk
Number of epochs: 2
Total training steps: approximately 52
Per-device batch size: 1
Gradient accumulation steps: 2
Number of GPUs: 2 Tesla T4 GPUs
Effective total batch size: 4
Learning rate: 2e-4
Optimizer: adamw_8bit
Gradient checkpointing: enabled
Save strategy: save every epoch
LoRA rank: 4
LoRA alpha: 8
Target modules: q_proj, v_proj
Vision layers fine-tuned: yes
Prompt used during training
Trumpai apibūdink orą šioje nuotraukoje lietuviškai. Atsakyk vienu paprastu sakiniu.
Example expected output style
Dangus giedras ir ryškiai mėlynas, debesų beveik nėra. Oras saulėtas, sausas, matomumas labai geras.
Training result summary
This second fine-tuning run used fewer epochs than the first version.
The first version was trained for 3 epochs and 78 total steps.
This Run B version was trained for 2 epochs and approximately 52 total steps.
The goal of Run B was to check whether less training gives better generalization on unseen images and reduces overfitting risk.
Because the dataset contains only 103 training examples, fewer epochs may help the model avoid memorizing the training captions too strongly.
Intended use
This adapter is intended for a university project demonstrating fine-tuning of a vision-language model for Lithuanian weather captioning.
It is not intended for professional meteorological forecasting.
- Downloads last month
- 11