InferenceIllusionist
/

Hyperion-1.5-Mistral-7B-iMat-GGUF

Inference Endpoints

Model card Files Files and versions Community

InferenceIllusionist commited on Mar 3

Commit

ece29ce

•

1 Parent(s): 401bc6b

Update README.md

Files changed (1) hide show

README.md +1 -1

README.md CHANGED Viewed

@@ -7,7 +7,7 @@ license: apache-2.0
 # Hyperion-1.5-Mistral-7B-iMat-GGUF
 New importance matrix quantizations for Hyperion-1.5-Mistral-7B.
-These i-quants have a better size to perplexity ratio as they were creating using an Importance Matrix file calcualted from the fp16 (unquantized) gguf.
 <b>All files created using latest (3/2) llama.cpp build, including IQ3_S improvements covered [here](https://github.com/ggerganov/llama.cpp/pull/5829)</b>

 # Hyperion-1.5-Mistral-7B-iMat-GGUF
 New importance matrix quantizations for Hyperion-1.5-Mistral-7B.
+These i-quants have a better size to perplexity ratio as they were creating using an Importance Matrix file calculated from the fp16 (unquantized) gguf.
 <b>All files created using latest (3/2) llama.cpp build, including IQ3_S improvements covered [here](https://github.com/ggerganov/llama.cpp/pull/5829)</b>