llama: honor add_space_prefix from the model configuration

propagate the add_space_prefix configuration from the HF model
configuration to the gguf file and honor it with the gpt2 tokenizer.

Signed-off-by: Giuseppe Scrivano <[email protected]>

Author: giuseppe

convert-hf-to-gguf.py

@@ -1315,6 +1315,13 @@ def set_gguf_parameters(self):
                 self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.LINEAR)
                 self.gguf_writer.add_rope_scaling_factor(self.hparams["rope_scaling"]["factor"])
 
+        tokenizer_config_file = self.dir_model / 'tokenizer_config.json'
+        if tokenizer_config_file.is_file():
+            with open(tokenizer_config_file, "r", encoding="utf-8") as f:
+                tokenizer_config_json = json.load(f)
+                if "add_prefix_space" in tokenizer_config_json:
+                    self.gguf_writer.add_add_space_prefix(tokenizer_config_json["add_prefix_space"])
+
     @staticmethod
     def permute(weights: Tensor, n_head: int, n_head_kv: int | None):
         if n_head_kv is not None and n_head != n_head_kv:

llama.cpp

@@ -4494,6 +4494,11 @@ static void llm_load_vocab(
         } else {
             if (tokenizer_model == "gpt2") {
                 vocab.type = LLAMA_VOCAB_TYPE_BPE;
+
+                const int add_space_prefix_keyidx = gguf_find_key(ctx, kv(LLM_KV_TOKENIZER_ADD_PREFIX).c_str());
+                if (add_space_prefix_keyidx != -1) {
+                    vocab.add_space_prefix = gguf_get_val_bool(ctx, add_space_prefix_keyidx);
+                }
             } else {
                 LLAMA_LOG_WARN("%s: unknown tokenizer: '%s'", __func__, tokenizer_model.c_str());
                 LLAMA_LOG_WARN("%s: using default tokenizer: 'llama'", __func__);