abhiramivs/malayalam_asr_model

🔧 LoRA Fine-Tuned Adapter for [Your Task or Dataset Name]

This repository contains a LoRA (Low-Rank Adaptation) fine-tuned adapter for the base model Whisper-small(https://huggingface.co/openai/whisper-small). The adapter is trained on [Malayalam speech data].

---

🧠 Base Model

• Base Model: whisper-small
• Model Type: Seq2Seq (pick one)
• LoRA Library: PEFT (Hugging Face)

---

🏋️ Training Details

• Framework: PyTorch + 🤗 Transformers + 🤗 PEFT
• LoRA Config:
  • r = 32
  • lora_alpha = 64
  • lora_dropout = 0.05
  • bias = "none"
  • `task_type = None

---

🧪 How to Use

from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel

# Load base model and tokenizer
base_model = AutoModelForCausalLM.from_pretrained("openai/whisper-small")
tokenizer = AutoTokenizer.from_pretrained("openai/whisper-small")

# Load LoRA adapter
model = PeftModel.from_pretrained(base_model, "abhiramivs/malayalam_asr_model")