Model Pretraining & Fine-Tuning
Workflow of training large models on broad data and adapting them to downstream tasks with smaller curated datasets.
Core metadata
- ID: model_pretraining_finetuning
- Era: Modern
- First known date: 2021 (exact)
- Region: Global / multiple regions
- Review status: source_checked
- Maturity: established
Prerequisites
- AI Training Clusters (ai_training_clusters)
- Self-Supervised Learning (self_supervised_learning)
- Supervised Learning Pipelines (supervised_learning_pipelines)
Dependents
Fields
Field lanes
- Artificial Intelligence & Machine Learning: Foundation Models
Node sources
- On the Opportunities and Risks of Foundation Models (arXiv / Stanford CRFM, 2021, primary_paper) • Supports: node, maturity
- Aligning Language Models to Follow Instructions (OpenAI, 2022, generic_overview) • Supports: node, maturity
Prerequisite edge evidence
Edge/source evidence summary:
- Prerequisite edges: 3
- Average edge confidence: 68%
- Prerequisite sources: 3
- expert_inference: 3
| Prerequisite | Type | Confidence | Evidence level | Note | Sources |
|---|---|---|---|---|---|
| Self-Supervised Learning (self_supervised_learning) | enabling | 68% | expert_inference | Self-Supervised Learning provides a capability that enables this technology without being the only possible path. |
|
| AI Training Clusters (ai_training_clusters) | enabling | 68% | expert_inference | AI Training Clusters provides a capability that enables this technology without being the only possible path. |
|
| Supervised Learning Pipelines (supervised_learning_pipelines) | enabling | 68% | expert_inference | Supervised Learning Pipelines provides a capability that enables this technology without being the only possible path. |
|
This page is generated from canonical era JSON and is indexable by URL.