codybum commited on
Commit
635f1cf
·
verified ·
1 Parent(s): 7688fad

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -41,7 +41,7 @@ instruction ability comes from **GRPO reinforcement learning** on a from-scratch
41
  — the exact metric the Hugging Face Open LLM Leaderboard publishes as "IFEval." We do **not** use the looser
42
  four-metric average anywhere.
43
 
44
- **David and Goliath.** KOS-V4 was trained on **180 billion tokens using 24 GPUs**. OpenAI trained the original
45
  ChatGPT (**GPT-3.5**) on a ~10,000-GPU supercomputer, and today's open models train on **trillions** of tokens
46
  (Qwen2.5 ~18T, Llama-3 ~15T). KOS-V4's 180B tokens is **~100× less data**. From that budget:
47
 
 
41
  — the exact metric the Hugging Face Open LLM Leaderboard publishes as "IFEval." We do **not** use the looser
42
  four-metric average anywhere.
43
 
44
+ KOS-V4 was trained on **180 billion tokens using 24 GPUs**. OpenAI trained the original
45
  ChatGPT (**GPT-3.5**) on a ~10,000-GPU supercomputer, and today's open models train on **trillions** of tokens
46
  (Qwen2.5 ~18T, Llama-3 ~15T). KOS-V4's 180B tokens is **~100× less data**. From that budget:
47