You can deploy this model and start earning money today!

arbib/
darija-qwen2.5-unsloth-lora_v2

Name: darija-qwen2.5-unsloth-lora_v2
Availability: InStock

Qwen/Qwen2.5-1.5B-Instruct

adapter

Views

No views yet

Darija Qwen2.5-1.5B — LoRA Adapter (Unsloth)

Modele fine-tune sur Qwen/Qwen2.5-1.5B-Instruct pour le dialecte marocain (Darija). Entrainement via Unsloth (2x plus rapide, 40% moins de VRAM).

Dataset

MBZUAI-Paris/Darija-SFT-Mixture
15 000 samples selectionnes et nettoyes

Configuration LoRA

lora_rank: 8
lora_alpha: 16
lora_dropout: 0.05
lora_target: q, k, v, o, gate, up, down

Utilisation

python

from unsloth import FastLanguageModel
import torch

model, tokenizer = FastLanguageModel.from_pretrained(
    model_name   = "/kaggle/working/darija_qwen_lora_adapter",
    max_seq_length = 512,
    dtype        = torch.float16,
    load_in_4bit = True,
)
FastLanguageModel.for_inference(model)

arbib/darija-qwen2.5-unsloth-lora_v2

Darija Qwen2.5-1.5B — LoRA Adapter (Unsloth)

Dataset

Configuration LoRA

Utilisation

arbib/
darija-qwen2.5-unsloth-lora_v2