2023-02-09_the-inference-cost-of-search-disruption
Canonical Target Counts
| Target | Occurrences |
|---|
| 02_companies/GOOG | 45 |
| 04_knowledge_base/Large language model | 32 |
| 02_companies/MSFT | 18 |
| 04_knowledge_base/ChatGPT | 18 |
| 02_companies/OpenAI | 11 |
| 04_knowledge_base/H100 | 9 |
| 04_knowledge_base/Sequence length | 9 |
| 04_knowledge_base/A100 | 7 |
| 02_companies/NVDA | 6 |
| 04_knowledge_base/TPU | 5 |
| 04_knowledge_base/GPU | 3 |
| 04_knowledge_base/Model Distillation | 3 |
| 04_knowledge_base/TPUv4 | 3 |
| 04_knowledge_base/GPT-3 | 3 |
| 04_knowledge_base/Mixture of Experts | 2 |
| 04_knowledge_base/HGX | 2 |
| 04_knowledge_base/TPUv5 | 2 |
| 04_knowledge_base/Pruning | 2 |
| 02_companies/AAPL | 1 |
| 04_knowledge_base/FP8 | 1 |
| 04_knowledge_base/Quantization | 1 |
| 04_knowledge_base/Sparsity | 1 |
| 04_knowledge_base/Early Exit | 1 |
| 04_knowledge_base/Model FLOPS Utilization | 1 |
| 04_knowledge_base/Batch size | 1 |