---
license_name: microsoft-research-license
license_link: https://huggingface.co/microsoft/phi-2/resolve/main/LICENSE
language:
- en
pipeline_tag: text-generation
base_model: microsoft/phi-2
tags:
- generated_from_trainer
model-index:
- name: phi-2-instruct
  results: []
datasets:
- HuggingFaceH4/ultrachat_200k
library_name: adapter-transformers
license: mit
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# phi-2-instruct

This model is a fine-tuned version of [microsoft/phi-2](https://huggingface.co/microsoft/phi-2) on the filtered ultrachat200k dataset using the SFT technique.

## Model description

More information about the model architecture and specific modifications made during fine-tuning is needed.

## Intended uses & limitations

More information about the intended use cases and any limitations of the model is needed.

## Training and evaluation data

More information about the datasets used for training and evaluation is needed.

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9, 0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- training_steps: 51967

### Training results

Detailed training results and performance metrics are not provided. It's recommended to reach out to the model creator for more information.

### Framework versions

- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0

## Evaluation and Inference Example

- For an evaluation of the model and an inference example, refer to the [Inference Notebook](https://huggingface.co/venkycs/phi-2-instruct/blob/main/inference_phi_2_instruct.ipynb).

## Full Training Metrics on TensorBoard

View the full training metrics on TensorBoard [here](https://huggingface.co/venkycs/phi-2-instruct/tensorboard).

## Author's LinkedIn Profile

[venkycs](https://linkedin.com/in/venkycs)