172: Transformers and Large Language Models
Programming Throwdown
00:00
Exploring Model Parameters, Memory Usage, and Fine-Tuning in Large Language Models
Exploring the impact of parameter count on memory usage in large language models and the efficacy of techniques like four bit quantization. A comparison between Gemma and llama models highlights the importance of the embedding layer in Gemma for text comprehension and generation.
Transcript
Play full episode
