Beta

Explore
Marketplace
Neural Labs
Chat
Wallet
Docs

Beta

Explore Marketplace Neural Labs Wallet Docs

Build

Manage

My assets Manage wallet Organization

Community

Discord X (Twitter)Telegram

Docs Privacy Policy Terms and Conditions

Proud member of

Proud member of

Docs Cookie Policy

Privacy Policy Terms and Conditions

© 2026 AlphaNeural AI. All rights reserved.

VinUni-lab21-Finetuning-w-LoRA-QLoRA-all-r16 – AI Model by DangMinh21 | AlphaNeural AI

You can deploy this model and start earning money today!

DangMinh21/
VinUni-lab21-Finetuning-w-LoRA-QLoRA-all-r16

like

0

unsloth/Llama-3.2-3B-Instruct-bnb-4bit

Views

No views yet

DangMinh21/VinUni-lab21-Finetuning-w-LoRA-QLoRA-all-r16

LoRA/QLoRA adapter trained for Lab 21 rank experiment.

Training Setup

Base model: unsloth/Llama-3.2-3B-Instruct-bnb-4bit
Dataset: 5CD-AI/Vietnamese-alpaca-gpt4-gg-translated
Sample size after cleaning: 273
Train/eval split: 245/28 with seed 42
Max sequence length: 1024
Epochs: 3
Learning rate: 0.0002
Scheduler: cosine
Batch size: 1
Gradient accumulation: 8
Effective batch size: 8
Optimizer: adamw_8bit
Packing: False

Adapter

Run ID: all-r16
Target modules: all
Rank: 16.0
Alpha: 32.0
Trainable parameters: 24313856.0
Trainable percent: 1.330241537665993

Evaluation

Eval loss: 1.7116294098751883
Eval perplexity: 5.537977765674739
Peak VRAM GB: 9.55007232
Train time minutes: 5.187989604473114

See the GitHub/Colab report for qualitative comparisons and rank trade-off analysis.