README.md

metadata

base_model: microsoft/Phi-3.5-mini-instruct
library_name: mlx-lm
tags:
  - mlx
  - lora
  - phi-3.5
license: mit

Phi 3.5 Mini LoRA Fine-tuned with MLX

This is a LoRA (Low-Rank Adaptation) fine-tuned version of Phi 3.5 Mini using MLX.

Usage with MLX

from mlx_lm import load, generate

model, tokenizer = load("kacperbb/phi-3.5-mlx-finetuned")
response = generate(model, tokenizer, prompt="Hello, how are you?", max_tokens=100)
print(response)