llama-quant: use LLM_KV constants instead of hardcoded strings

Remove TODO comments by using the proper LLM_KV_GENERAL_QUANTIZATION_VERSION
and LLM_KV_GENERAL_FILE_TYPE constants in llama-quant.cpp.

Also remove an obsolete TODO comment in unicode.cpp.

Author: lainon1

src/llama-quant.cpp

@@ -926,8 +926,8 @@ static void llama_model_quantize_impl(const std::string & fname_inp, const std::
 
     // copy the KV pairs from the input file
     gguf_set_kv     (ctx_out.get(), ml.metadata);
-    gguf_set_val_u32(ctx_out.get(), "general.quantization_version", GGML_QNT_VERSION); // TODO: use LLM_KV
-    gguf_set_val_u32(ctx_out.get(), "general.file_type", ftype); // TODO: use LLM_KV
+    gguf_set_val_u32(ctx_out.get(), ml.llm_kv(LLM_KV_GENERAL_QUANTIZATION_VERSION).c_str(), GGML_QNT_VERSION);
+    gguf_set_val_u32(ctx_out.get(), ml.llm_kv(LLM_KV_GENERAL_FILE_TYPE).c_str(), ftype);
 
     // Remove split metadata
     gguf_remove_key(ctx_out.get(), ml.llm_kv(LLM_KV_SPLIT_NO).c_str());

src/unicode.cpp

@@ -804,7 +804,6 @@ static std::vector<size_t> unicode_regex_split_custom(const std::string & text,
         // tiny_aya digit grouping pattern from tokenizer.json:
         //   {"type": "Split", "pattern": {"Regex": "\\d{1,3}(?=(?:\\d{3})*\\b)"}, "behavior": "Isolated"}
         // Splits digits into groups of 3 from the right (e.g., 1234567 -> 1, 234, 567)
-        // TODO: Revisit this regex, in case there are any subtle tokenization differences with the original regex.
         bpe_offsets = unicode_regex_split_custom_afmoe(text, offsets);
     }