Sign inSign in
ConceptAI & MLPhysics

Fine-tuning

One word, taught differently in 2 fields. Each use is given apart.

In AI & ML

The process of taking a pre-trained model and further training it on a smaller, task-specific dataset to improve its performance on that task.
as Leslie Kaelbling defines itMITThe wording closest to all 8 definitions

In Physics

The requirement that parameters of a physical theory be set to extremely precise values to match observed phenomena.

Know first

Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.

Opens up

Concepts that name this one as a prerequisite.

In AI & ML

8

Each in their own terms, from the easiest level.

  • Fei-Fei Li, Justin Johnson & Serena Yeung

    Stanford University · Undergraduate

    Finetuning involves updating larger parts of a pre-trained network using a new dataset with a lower learning rate.
  • Andrej Karpathy

    Undergraduate

    The second stage of training where the base model is trained on high-quality Q&A conversations to become an assistant model.
    • 14:30Introduction to Large Language Models
  • Jeremy Howard

    Undergraduate

    Changing a model to behave based on specific documents that we have available.
    • 1:20:51A Hacker's Guide to Language Models
  • Sara Beery

    MIT · Undergraduate

    Taking a pre-trained model and adapting it to a new task by initializing a second network with some or all of the original parameters and training it on new data.
    • 13:54Transfer Learning: Models
  • Phillip Isola

    MIT · Undergraduate

    Fine-tuning initializes a second network with pre-trained weights and continues training on a new target data set.
    • 32:37Representation Learning: Reconstruction-Based
  • Melanie Mitchell

    Santa Fe Institute · Undergraduate

    The process of further training a pre-existing machine learning model on specific data using consumer-grade hardware or advanced tools like PEFT and LoRA.
    • 1:21:06The Future of Artificial Intelligence
  • Albert Jiang

    Stanford University · Graduate

    The process of further training a pre-trained model on a specific dataset or domain to adapt it to particular tasks.
    • 34:00Demystifying Mixtral of Experts

In Physics

1
  • Nima Arkani-Hamed

    Perimeter Institute · Undergraduate

    • 4:20:51The Future of Fundamental Physics: Spacetime and Quantum Amplitudes

Lectures that assume it

17

The first 12 are shown.

Connections

37

The lectures that define it or assume it, and the concepts next to it.

Concepts10

Show 4 more

Lectures27

Show 21 more

9 lecturers. Definitions are written by AI from each lecture’s transcript. A time with a play mark starts the lecture at the moment the lecturer says it.