diff --git a/chapters/en/chapter1/4.mdx b/chapters/en/chapter1/4.mdx index 3870b541f..71c8b727d 100644 --- a/chapters/en/chapter1/4.mdx +++ b/chapters/en/chapter1/4.mdx @@ -121,8 +121,10 @@ This pretraining is usually done on very large amounts of data. Therefore, it re For example, one could leverage a pretrained model trained on the English language and then fine-tune it on an arXiv corpus, resulting in a science/research-based model. The fine-tuning will only require a limited amount of data: the knowledge the pretrained model has acquired is "transferred," hence the term *transfer learning*.
-The fine-tuning of a language model is cheaper than pretraining in both time and money. - + +The fine-tuning of a language model is cheaper than pretraining in both time and money. + +
Fine-tuning a model therefore has lower time, data, financial, and environmental costs. It is also quicker and easier to iterate over different fine-tuning schemes, as the training is less constraining than a full pretraining. diff --git a/chapters/en/images/finetuning_darkupdated.svg b/chapters/en/images/finetuning_darkupdated.svg new file mode 100644 index 000000000..d82109c76 --- /dev/null +++ b/chapters/en/images/finetuning_darkupdated.svg @@ -0,0 +1,178 @@ + + + + + + + + + + + + + + + + + + + + + + + + + Easily Reproducible + diff --git a/chapters/en/images/finetuning_updated.svg b/chapters/en/images/finetuning_updated.svg new file mode 100644 index 000000000..8df4d3adb --- /dev/null +++ b/chapters/en/images/finetuning_updated.svg @@ -0,0 +1,156 @@ + + + + + + + + + + + + + + + + + + + + + + + Easily Reproducible +