🦙 LoRA Implementation of LLaMA
This model is a Low-Rank Adaptation (LoRA) of the LLaMA 3.2-1B model.
Original size: ~1 Billion parameters
LoRA fine-tuned parameters: ~86,000
Result: Achieves similar performance on many prompts with dramatically reduced trainable weights.
About
This model was fine-tuned using the PEFT (Parameter-Efficient Fine-Tuning) library and Hugging Face's transformers.
- Base Model:
meta-llama/Llama-3.2-1B
Usage
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained("your-username/Llama-3.2-1B_LoRA")
tokenizer = AutoTokenizer.from_pretrained("your-username/Llama-3.2-1B_LoRA")
inputs = tokenizer("Tell me something interesting about space.", return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=50)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support
Model tree for atharvashembade/Llama-3.2-1B_LoRA
Base model
meta-llama/Llama-3.2-1B