Benchmarks/Seed-OSS-36B-Instruct-unsloth-Q5_K/bench_metrics.json

{
  "llamabench": {
    "backend": "CUDA",
    "log_path": "Seed-OSS-36B-Instruct-unsloth-Q5_K/llamabench.md",
    "ngl": "35",
    "raw_row": {
      "backend": "CUDA",
      "model": "seed_oss 36B Q5_K - Medium",
      "ngl": "35",
      "params": "36.15 B",
      "size": "23.83 GiB",
      "t/s": "22.04 \u00b1 0.37",
      "test": "pp8",
      "tps_value": 22.04
    },
    "test": "pp8",
    "tps": 22.04
  },
  "perplexity": {
    "code": {
      "log_path": "Seed-OSS-36B-Instruct-unsloth-Q5_K/perplexity_code.log",
      "ppl": 1.4169,
      "ppl_error": 0.00959
    },
    "general": {
      "log_path": "Seed-OSS-36B-Instruct-unsloth-Q5_K/perplexity_general.log",
      "ppl": 6.9056,
      "ppl_error": 0.16854
    },
    "math": {
      "log_path": "Seed-OSS-36B-Instruct-unsloth-Q5_K/perplexity_math.log",
      "ppl": 5.4616,
      "ppl_error": 0.12126
    }
  }
}