100x less compute with GPT-level LLM performance: How a little known open source project could help solve the GPU power conundrum — RWKV looks promising but challenges remain
CAT
1 min read
Open source project RWKV can train LLMs with far fewer GPUs than current models.
Originally appeared here:
100x less compute with GPT-level LLM performance: How a little known open source project could help solve the GPU power conundrum — RWKV looks promising but challenges remain
