feat: fill-in-middle support (#1386)

* Proper fill-in-middle support

Use prefix/middle/suffix tokens when metadata is present in GGUF, like f.ex. in [this](https://huggingface.co/CISCai/CodeQwen1.5-7B-Chat-SOTA-GGUF) one.

* fall back to internal prefix/middle/suffix id

In some cases llama.cpp will make a guess at fim tokens, use them if there's no metadata.

* typo--

* don't insert special tokens that are not there in suffix

Note: add_bos is misnamed, it's actually add_special and can cause several special tokens to be added to the token list (the special parameter is actually parse_special).

* don't add/parse any special tokens when using fim

I've left original behavior when no fim tokens are found, but this should perhaps be re-evaluated.

* don't append suffix to prompt_tokens unless fim tokens are detected

* make sure we only do this for fim

---------

Co-authored-by: Andrei <abetlen@gmail.com>

llama_cpp/llama.py

@@ -955,19 +955,54 @@ class Llama:
 
         completion_id: str = f"cmpl-{str(uuid.uuid4())}"
         created: int = int(time.time())
+        prefix_token_id: int = int(self.metadata.get("tokenizer.ggml.prefix_token_id", self._model.token_prefix()))
+        middle_token_id: int = int(self.metadata.get("tokenizer.ggml.middle_token_id", self._model.token_middle()))
+        suffix_token_id: int = int(self.metadata.get("tokenizer.ggml.suffix_token_id", self._model.token_suffix()))
         # If prompt is empty, initialize completion with BOS token to avoid
         # detokenization including a space at the beginning of the completion
         completion_tokens: List[int] = [] if len(prompt) > 0 else [self.token_bos()]
         # Add blank space to start of prompt to match OG llama tokenizer
         prompt_tokens: List[int] = (
             (
-                self.tokenize(prompt.encode("utf-8"), special=True)
+                [prefix_token_id]
+                if prefix_token_id >= 0 and suffix is not None
+                else []
+            )
+            +
+            (
+                (
+                    self.tokenize(prompt.encode("utf-8"), add_bos=(prefix_token_id < 0 or suffix is None), special=(prefix_token_id < 0 or suffix is None))
                     if prompt != ""
+                    else (
+                        []
+                        if prefix_token_id >= 0 and suffix is not None
                         else [self.token_bos()]
                     )
+                )
                 if isinstance(prompt, str)
                 else prompt
             )
+            +
+            (
+                (
+                    [suffix_token_id]
+                    +
+                    (
+                        self.tokenize(suffix.encode("utf-8"), add_bos=False, special=False)
+                        if suffix
+                        else []
+                    )
+                )
+                if suffix_token_id >= 0 and suffix is not None
+                else []
+            )
+            +
+            (
+                [middle_token_id]
+                if middle_token_id >= 0 and suffix is not None
+                else []
+            )
+        )
         text: bytes = b""
         returned_tokens: int = 0
         stop = (
@@ -1346,7 +1381,7 @@ class Llama:
         if echo:
             text_str = prompt + text_str
 
-        if suffix is not None:
+        if suffix_token_id < 0 and suffix is not None:
             text_str = text_str + suffix
 
         logprobs_or_none: Optional[CompletionLogprobs] = None