AzureBlack
/

Umbra-v3-MoE-4x11b-exl2-8bpw-8h-rpcal

Text Generation

Mixture of Experts

Himitsui/Kaiju-11B

Sao10K/Fimbulvetr-11B-v2

decapoda-research/Antares-11b-v2

beberik/Nyxene-v3-11B

text-generation-inference

Inference Endpoints

Model card Files Files and versions Community

AzureBlack commited on Mar 5

Commit

19eebd2

•

1 Parent(s): cbba7d3

Update README.md

Files changed (1) hide show

README.md +1 -1

README.md CHANGED Viewed

@@ -128,7 +128,7 @@ calibration dataset [here.](https://huggingface.co/datasets/royallab/PIPPA-clean
 Requires ExllamaV2, which is being developed by turboderp https://github.com/turboderp/exllamav2 under an MIT license.
-DONT USE MAIN BRANCH - Test using 8192 measurement length and rp dataset. Perplexity came out too high. Will update with a normal length rp later.
 Branch is 8b8h using wikitext at 4096 length

 Requires ExllamaV2, which is being developed by turboderp https://github.com/turboderp/exllamav2 under an MIT license.
+Test using 4096 measurement length and rp dataset. Perplexity came out to an 8 vs the Wiki which was at a 6. Haven't tested enough to tell if there is much difference in practice between the two.
 Branch is 8b8h using wikitext at 4096 length