Downloads · 30 days
20
49% of all-time downloads
phi3-mini-qlora-guanaco
parama123/phi3-mini-qlora-guanaco
phi3-mini-qlora-guanaco is a text generation model from parama123. Use it when you need the model to write or continue text. It is set up for peft. The card lists the license as mit.
QLoRA fine-tuned on Google Colab (T4 GPU) using free tools. Fits in 16GB VRAM.
All-time downloads
41
Public
Repo size
412 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors412 MB · 99%
At a glance
- Task
- Text Generation
- Library
- peft
- License
- mit
- Access
- Public
- Created
- Oct 30, 2025
- Updated
- Oct 30, 2025
- SHA
- 5b58ca3a
Try a prompt
Base models
Datasets
Model card
From the Hugging Face model README
Phi-3-mini QLoRA Fine-Tuned (3.8B)
QLoRA fine-tuned on Google Colab (T4 GPU) using free tools. Fits in 16GB VRAM.
Quick Stats
- Base:
microsoft/Phi-3-mini-4k-instruct(3.8B) - Method: 4-bit NF4 + LoRA (r=64) via PEFT/bitsandbytes
- Dataset: guanaco-llama2-1k
- Memory: ~14GB (from ~30GB)
- GitHub: paramashiva123/phi3-mini-qlora-finetune
Usage
from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
import torch
base = AutoModelForCausalLM.from_pretrained(
"microsoft/Phi-3-mini-4k-instruct",
torch_dtype=torch.bfloat16,
device_map="auto",
trust_remote_code=True
)
model = PeftModel.from_pretrained(base, "parama123/phi3-mini-qlora-guanaco")
tokenizer = AutoTokenizer.from_pretrained("parama123/phi3-mini-qlora-guanaco")
pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)
print(pipe("### Human: Explain QLoRA.\n### Assistant:")[0]["generated_text"])
## Training Notebook
[](https://colab.research.google.com/drive/1mLBrbGIoGg0HLqQYgADO-_Rr20bN32ry?usp=sharing)
- Task
- Text Generation
- Library
- peft
- License
- mit
- Created
- Oct 30, 2025
- Updated
- Oct 30, 2025