bart-large-mnli: instruction tuned - v1

<a href="https://colab.research.google.com/gist/pszemraj/43431a164e2e3ab0640182f3419c584d/bart-large-instruct-example.ipynb"> <img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/> </a>

This model is a fine-tuned version of facebook/bart-large-mnli on the pszemraj/dolly_hhrlhf-text2text dataset.

Model description

text2text models fine-tuned on a modified dataset for text2text generation based on the relatively more permissive mosaicml/dolly_hhrlhf dataset.

Basic usage in Python:

# pip install -q transformers accelerate
import torch
from transformers import pipeline, GenerationConfig

model_name = "pszemraj/bart-large-mnli-instruct-dolly_hhrlhf-v1"
assistant = pipeline(
    "text2text-generation",
    model_name,
    device_map="auto",
)
cfg = GenerationConfig.from_pretrained(model_name)

# pass an 'instruction' as the prompt to the pipeline
prompt = "Write a guide on how to become a ninja while working a 9-5 job."
result = assistant(prompt, generation_config=cfg)[0]["generated_text"]
print(result)

The use of the generation config is optional, it can be replaced by other generation params.

Intended Uses & Limitations

• This is not tuned with RLHF, etc, and may produce offensive results.
• While larger than BART-base, this model is relatively small compared to recent autoregressive models (MPT-7b, LLaMA, etc.), and therefore it's "cognition" capabilities may be practically limited for some tasks.

Training

Training hyperparameters

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: 8
• 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: 0.03
• num_epochs: 3.0