scisharp
/
LLamaSharp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge pull request #698 from martindevans/slightly_safer_quantize_params 

Slightly Safer Quantize Params
This commit is contained in:
Martin Evans 2024-04-26 13:53:55 +01:00 committed by GitHub
parent ecb359c9e7 58ec798bff
commit a2f8573831
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 6 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
LLama/LLamaQuantizer.cs
View File

@ -34,12 +34,12 @@ namespace LLama
quantizeParams.nthread = nthread;
quantizeParams.allow_requantize = allowRequantize;
quantizeParams.quantize_output_tensor = quantizeOutputTensor;
//todo: fill in other quantize params fields.
unsafe
{
return NativeApi.llama_model_quantize(srcFileName, dstFilename, &quantizeParams) == 0;
}
// todo: fill in other quantize params fields.
// This method could probably do with a redesign - passing in a config object (maybe directly
// expose `LLamaModelQuantizeParams`) instead of an ever growing list of method parameters!
return NativeApi.llama_model_quantize(srcFileName, dstFilename, ref quantizeParams) == 0;
}
/// <summary>

2
LLama/Native/NativeApi.Quantize.cs
View File

@ -12,6 +12,6 @@ namespace LLama.Native
/// <param name="param"></param>
/// <returns>Returns 0 on success</returns>
[DllImport(libraryName, CallingConvention = CallingConvention.Cdecl)]
public static extern unsafe uint llama_model_quantize(string fname_inp, string fname_out, LLamaModelQuantizeParams* param);
public static extern uint llama_model_quantize(string fname_inp, string fname_out, ref LLamaModelQuantizeParams param);
}
}