Basemodel: roBERTa

Configs: Vocab size: 10,000 Hidden size: 512 Max position embeddings: 512 Number of layers: 2 Number of heads: 4 Window size: 256 Intermediate-size: 1024

Results:

Task: glue Score: 57.44 Confidence Interval: [56.52, 58.73]
Task: blimp Score: 58.48 Confidence Interval: [56.68, 59.5]

Downloads last month: 165

Inference Providers NEW

Fill-Mask

This model is not currently available via any of the supported Inference Providers.

Dataset used to train AISE-TUDelft/Custom-Activations-BERT-GELU

Collection including AISE-TUDelft/Custom-Activations-BERT-GELU

BRP Tiny-Transformers

Collection

Models for the 2024-Q4 BSc. Research Project: "Architectural Decisions for Language Modelling with Small Transformers". • 14 items • Updated Jun 25, 2024 • 1