Transformers
Safetensors
fc91's picture
Update README.md
812e61b verified
|
raw
history blame
6.25 kB
metadata
library_name: transformers
license: cc-by-4.0
datasets:
  - hendrycks/ethics

Model Card for Model ID

Fine-tuned version of Phi-3-mini-4k-instruct on a subset of the hendrycks/ethics dataset

How to Get Started with the Model

Use the code below to get started with the model.

from transformers import AutoModelForCausalLM
from peft import PeftModel

base_model = AutoModelForCausalLM.from_pretrained("microsoft/Phi-3-mini-4k-instruct")
peft_model_id = "fc91/phi3-mini-instruct-full_ethics-lora"
model = PeftModel.from_pretrained(base_model, peft_model_id)

Training Details

Training Data

"hendrycks/ethics"

The following subsets of the above dataset were leveraged:
-commonsense (10k random samples)
-deontology (10k random samples)
-justice (10k random samples)
-utilitarianism (10k random samples)

Training Procedure

Training Hyperparameters

per_device_train_batch_size=16  
per_device_eval_batch_size=32  
gradient_accumulation_steps=2
gradient_checkpointing=True
warmup_steps=100
num_train_epochs=1
learning_rate=0.00005
weight_decay=0.01
optim="adamw_hf"
fp16=True

Speeds, Sizes, Times

The overall training took 3 hours and 23 minutes.

Evaluation

Training Loss = 0.181700

Validation Loss = 0.119734

Testing Data, Factors & Metrics

Testing Data

"hendrycks/ethics"

The following subsets of the above dataset were leveraged:
-commonsense (2.5k random samples)
-deontology (2.5k random samples)
-justice (2.5k random samples)
-utilitarianism (2.5k random samples)

Hardware

NVIDIA A100-SXM4-40GB