ITransformer trainedModel;
            using (var stream = new FileStream(modelFileFullPath, FileMode.Open, FileAccess.Read, FileShare.Read))
            {
               var trainedModel = mlContext.Model.Load(stream, out var inputSchema);
               var = mlContext.Model.CreatePredictionEngine<TInput, TOutput>(trainedModel, inputSchemaDefinition: inputSchema); 

//Throws an exception features<vector<single>> does not match features<vector<single, 432>> because CreatePredictionEngine returns transformer.CreatePredictionEngine<TSrc, TDst>(_env, false, DataViewConstructionUtils.GetSchemaDefinition<TSrc>(_env, inputSchema)); and `DataViewConstructionUtils.GetSchemaDefinition` overrides inputSchema.

            }

//Below workaround that overrides the override!

using (var stream = new FileStream(modelFileFullPath", FileMode.Open, FileAccess.Read, FileShare.Read))
            {

                var trainedModel = mlContext.Model.Load(stream, out var inputSchema);
                var outputSchemaDefinition = SchemaDefinition.Create(typeof(TInput));
                outputSchemaDefinition["Features"].ColumnType = new VectorDataViewType(NumberDataViewType.Single, Convert.ToInt32(432));
                var p = mlContext.Model.CreatePredictionEngine<TInput, TOutput>(trainedModel, inputSchemaDefinition: outputSchemaDefinition);
            }