chintanshrinath/chintan-Llama-3.2-1B-Instruct

Model Details

Model Description

• Developed by: Chintan Shah

• Model type: meta-llama/Llama-3.2-1B-Instruct

• Finetuned from model [optional]: meta-llama/Llama-3.2-1B-Instruct

Training Details

Training Data

mlabonne/orpo-dpo-mix-40k

Training Procedure

ORPO

Training Parameters

Training Arguments:

• Learning Rate: 1e-5
• Batch Size: 1
• max_steps: 1
• Block Size: 512
• Warmup Ratio: 0.1
• Weight Decay: 0.01
• Gradient Accumulation: 4
• Mixed Precision: bf16

Training Hyperparameters

• Training regime: fp16 mixed precision

LoRA Configuration:

• R: 16
• Alpha: 32
• Dropout: 0.05

Evaluation

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
hellaswag	1	none	0	acc	↑	0.4408	±	0.0050
		none	0	acc_norm	↑	0.5922	±	0.0049

Testing Data, Factors & Metrics

Testing Data

https://github.com/EleutherAI/lm-evaluation-harnes