Download logs/kld-nvidia-code.log from AtomicChat/DeepSeek-V4.1-Flash-NVFP4-metrics: direct link, hf CLI and curl.
- Browser
- Download file 509 Bytes
-
https://huggingface.co/datasets/AtomicChat/DeepSeek-V4.1-Flash-NVFP4-metrics/resolve/8f498e6f7153994aea5f07d04c4f0f5bbdec5dff/logs/kld-nvidia-code.log
- Command line
-
hf download hf://datasets/AtomicChat/DeepSeek-V4.1-Flash-NVFP4-metrics@8f498e6f7153994aea5f07d04c4f0f5bbdec5dff/logs/kld-nvidia-code.log
-
curl -L -o kld-nvidia-code.log https://huggingface.co/datasets/AtomicChat/DeepSeek-V4.1-Flash-NVFP4-metrics/resolve/8f498e6f7153994aea5f07d04c4f0f5bbdec5dff/logs/kld-nvidia-code.log
509 Bytes
| KL lower bound : 0.019801 95% CI over 24 windows [0.014174, 0.025716] | |
| median KLD : 0.000037 p90 0.046985 p95 0.094739 p99 0.314394 max 3.4148 | |
| same top-1 : 96.757 % [95.549, 97.894] | |
| ppl : reference 1.8919 candidate 1.8955 (+0.187 %) | |
| |d logprob| : 0.08350 on the actual tokens | |
| common set : mass median 1.00000 p1 0.99213 min 0.7041 (coarsened with 9 run(s)) | |
| normalization : stored top-K mass max 1.0000018, rows past 1+1e-4: 0 | |
| written to /logs/kld-nvidia-code.json | |