metadata
library_name: transformers
tags:
- unsloth
license: mit
datasets:
- efederici/alpaca-vs-alpaca-orpo-dpo
Model Card for Model ID
This is phi-3-mini-4k-instruct ORPO finetuning for the italian language over the Alpaca vs. Alpaca italian dataset: efederici/alpaca-vs-alpaca-orpo-dpo
Model Details
Model Description
- Developed by: Diego Giorgini
- Funded by: AI Technologies SRL - www.aitechnologies.it
- Language(s) (NLP): Italian
- License: llama3
- Finetuned from model: unsloth/Phi-3-mini-4k-instruct
Training Details
Environment
unsloth: 2024.5
torch: 2.2
Training Data
efederici/alpaca-vs-alpaca-orpo-dpo: The Alpaca vs. Alpaca dataset is a curated blend of the Alpaca dataset and the Alpaca GPT-4 dataset, both available on HuggingFace Datasets. It uses the standard GPT dataset as the 'rejected' answer, steering the model towards the GPT-4 answer, which is considered as the 'chosen' one.
Training Procedure
Preprocessing [optional]
No preprocessing has been performed, except for formatting with the phi-3 chat_template from unsloth:
tokenizer = get_chat_template(tokenizer, chat_template = "phi-3")
Training Hyperparameters
Training regime: bf16
Model loading parameters:
max_seq_length = 8192
dtype = None
load_in_4bit = False
- PEFT parameters:
r = 64
lora_alpha = 64
lora_dropout = 0
bias = "none"
random_state = 3407
use_rslora = False
loftq_config = None
- ORPOConfig parameters:
max_length = 8192
max_prompt_length = max_seq_length//2
max_completion_length = max_seq_length//2
warmup_ratio = 0.1
weight_decay = 0.01
per_device_train_batch_size = 1
gradient_accumulation_steps = 16
learning_rate=8e-6
beta = 0.1
optim = "paged_adamw_8bit"
lr_scheduler_type = "linear"
num_train_epochs = 1
Speeds, Sizes, Times
7h on an A100-40GB