Q2 and Q3 quantization

Author: sw

examples/quantize/quantize.cpp

@@ -12,6 +12,8 @@ int main(int argc, char ** argv) {
 
     if (argc < 4) {
         fprintf(stderr, "usage: %s model-f32.bin model-quant.bin type [nthread]\n", argv[0]);
+        fprintf(stderr, "  type = %d - q2_0\n", LLAMA_FTYPE_MOSTLY_Q2_0);
+        fprintf(stderr, "  type = %d - q3_0\n", LLAMA_FTYPE_MOSTLY_Q3_0);
         fprintf(stderr, "  type = %d - q4_0\n", LLAMA_FTYPE_MOSTLY_Q4_0);
         fprintf(stderr, "  type = %d - q4_1\n", LLAMA_FTYPE_MOSTLY_Q4_1);
         fprintf(stderr, "  type = %d - q4_2\n", LLAMA_FTYPE_MOSTLY_Q4_2);