model : add tokenizer from LFM2.5-Audio-1.5B (#19687)

* model : Add tokenizer from LFM2.5-Audio-1.5B

[LFM2.5-Audio-1.5B](https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B) introduced lightweight audio tokenizer.

Tokenizer based on LFM2 architecture and acts as "embedding" model with
different input `n_embd` and output `n_embd_out`.

To be used in https://github.com/ggml-org/llama.cpp/pull/18641.

To convert use

```shell
python3 convert_hf_to_gguf.py /path/to/LFM2.5-Audio-1.5B/audio_detokenizer
```

* Update convert_hf_to_gguf.py

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>

* Formatting

* Rework check for attention layers

* Add LFM2 SWA model support

* Address PR feedback

* Set vocab to none

* Move helper function definitions to cpp file

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>

convert_hf_to_gguf.py

@@ -10726,7 +10726,7 @@ class LFM2Model(TextModel):
     def set_gguf_parameters(self):
         # set num_key_value_heads only for attention layers
         self.hparams["num_key_value_heads"] = [
-            self.hparams["num_key_value_heads"] if layer_type == "full_attention" else 0
+            self.hparams["num_key_value_heads"] if layer_type != "conv" else 0
             for layer_type in self.hparams["layer_types"]
         ]
 
@@ -10912,6 +10912,28 @@ class LFM2AudioModel(ConformerAudioModel):
         yield from super().modify_tensors(data_torch, name, bid)
 
 
+@ModelBase.register("Lfm25AudioTokenizer")
+class LFM25AudioTokenizer(LFM2Model):
+    model_arch = gguf.MODEL_ARCH.LFM2
+
+    def set_vocab(self):
+        self._set_vocab_none()
+
+    def set_gguf_parameters(self):
+        super().set_gguf_parameters()
+        self.gguf_writer.add_sliding_window(self.hparams["sliding_window"])
+        self.gguf_writer.add_embedding_length_out(self.hparams["output_size"])
+
+    def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+        if name == "istft.window" or name.startswith("emb.emb"):
+            return
+
+        if name.startswith("lin"):
+            name = name.replace("lin", "dense_2_out")
+
+        yield from super().modify_tensors(data_torch, name, bid)
+
+
 @ModelBase.register("SmallThinkerForCausalLM")
 class SmallThinkerModel(TextModel):
     model_arch = gguf.MODEL_ARCH.SMALLTHINKER