README.md

---
license: 
- apache-2.0
- other
tags:
- generated_from_trainer
- text-generation
- opt
- non-commercial
- dialogue
- chatbot

inference: false
---

# pszemraj/opt-peter-2.7B

This model is a fine-tuned version of [facebook/opt-2.7b](https://huggingface.co./facebook/opt-2.7b) on about 80k whatsapp/text messages (mine). Please use responsibly :)

Test it out on Google Colab [here](https://colab.research.google.com/gist/pszemraj/26a69775c9d012051396ab5ae980f5c1/example-text-gen-pszemraj-opt-peter-2-7b.ipynb)!

![chatdemo](https://i.imgur.com/1EgQYat.png)

## Model description

- Exploring to see how OPT does in terms of dialogue/conversational applications
- Seems to do a lot better than GPT-Neo with similar training parameters 
- you can create your own digital clone and deploy it leveraging [this repository I am working on](https://github.com/pszemraj/ai-msgbot).

## Intended uses & limitations

> The base model has a custom license which propogates to this one. Most importantly, it cannot be used commercially. Read more here: [facebook/opt-2.7b](https://huggingface.co./facebook/opt-2.7b) 

- the model is probably too large to use via API here. Use in Python with GPU RAM / CPU RAM > 12 gb, Colab notebook linked above.
  - alternatively, you can message [a bot on telegram](http://t.me/GPTPeter_bot) where I test LLMs for dialogue generation
- **any statements or claims made by this model do not reflect actual claims/statements by me.** Keep in mind it is a _fine-tuned_ version of the model on my data, so things from pre-training are also present in outputs.

## Training and evaluation data

WhatsApp & iMessage parsed using [ai-msgbot](https://github.com/pszemraj/ai-msgbot) and then fed as a text dataset to the HF trainer.

## Training procedure

### Training hyperparameters

**SESSION ONE**

The following hyperparameters were used during training:
- learning_rate: 4e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- gradient_accumulation_steps: 16
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.01
- num_epochs: 3

**SESSION TWO**

The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- distributed_type: multi-GPU
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.05
- num_epochs: 4


### Framework versions

- Transformers 4.19.2
- Pytorch 1.10.0+cu113
- Datasets 2.2.2
- Tokenizers 0.12.1