CONTENT WARNING: Readers should be aware this section contains content that is disturbing, offensive, and can propagate historical and current stereotypes.

Significant research has explored bias and fairness issues with language models (see, e.g., Sheng et al. (2021) and Bender et al. (2021)).

Training

Training Data

Hebrew Wikipedia Dump (hewiki abstract) from May 2020

Training Procedure

This model was fined tuned upon hebrew-gpt_neo-tiny which was previously trained using EleutherAI's gpt-neo.

Fine-tuning on the wiki-absract text was done using @minimaxir's aitextgen.

Evaluation

Configs

Model configs for the hebrew-gpt_neo-tiny is available on the hebrew-gpt_neo model github

Activation Function: gelu
Number_Head: 12
Number_Vocab: 50257
Train batch size: 250
Eval batch size: 64
Predict batch size: 1

Environmental Impact

Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). We present the hardware type based on the associated paper.

Hardware Type: [More information needed]
Hours used: Unknown
Cloud Provider: GCP tpu-v8s
Compute Region: europe-west4
Carbon Emitted: [More information needed]

How to Get Started With the Model

A Google Colab Notebook is also available here

from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Norod78/hebrew-bad_wiki-gpt_neo-tiny")

model = AutoModelForCausalLM.from_pretrained("Norod78/hebrew-bad_wiki-gpt_neo-tiny")

Norod78/hebrew-bad_wiki-gpt_neo-tiny

hebrew-bad_wiki-gpt_neo-tiny

Table of Contents

Model Details

Uses

Direct Use

Misuse and Out-of-scope Use

Risks, Limitations and Biases