radi-cho/Meta-Llama-3.1-8B-FLUTE

Model

| | Wiki | C4 | PIQA | ARC-E | ARC-C | HellaSwag | Wino | Avg. |

0

5 commits

2 linked in READMEs

updated Aug 4, 2024

See the code

README

WikiC4PIQAARC-EARC-CHellaSwagWinoAvg.
Unquantized6.319.6079.1682.2052.6560.7174.0369.75
W4G646.2410.0679.3881.6151.5459.5773.5669.13
W3G647.2311.8377.9176.9846.3356.7470.3265.66

Revisions available in this repository:

  • main (W4G64, scales learned);
  • nfl_w3g64 (W3G64, scales learned);

Evaluations are provided for models with learned scales.
Benchmark scores (zero-shot) are computed with lm-evaluation-harness.

llama
safetensors

radi-cho/Meta-Llama-3.1-8B-FLUTE

Model

| | Wiki | C4 | PIQA | ARC-E | ARC-C | HellaSwag | Wino | Avg. |

0

5 commits

2 linked in READMEs

updated Aug 4, 2024

See the code

README

WikiC4PIQAARC-EARC-CHellaSwagWinoAvg.
Unquantized6.319.6079.1682.2052.6560.7174.0369.75
W4G646.2410.0679.3881.6151.5459.5773.5669.13
W3G647.2311.8377.9176.9846.3356.7470.3265.66

Revisions available in this repository:

  • main (W4G64, scales learned);
  • nfl_w3g64 (W3G64, scales learned);

Evaluations are provided for models with learned scales.
Benchmark scores (zero-shot) are computed with lm-evaluation-harness.

llama
safetensors