Support for slerp merging models (#455)

* support for slerp merging models

* docs

* update docs

* format'

llms/mlx_lm/models/llama.py

@@ -205,3 +205,7 @@ class Model(nn.Module):
         return {
             k: v for k, v in weights.items() if "self_attn.rotary_emb.inv_freq" not in k
         }
+
+    @property
+    def layers(self):
+        return self.model.layers

llms/mlx_lm/models/mixtral.py

@@ -11,7 +11,6 @@ from .base import BaseModelArgs
 @dataclass
 class ModelArgs(BaseModelArgs):
     model_type: str
-    vocab_size: int
     vocab_size: int = 32000
     max_position_embeddings: int = 4096 * 32
     hidden_size: int = 4096
@@ -260,3 +259,7 @@ class Model(nn.Module):
     ):
         out, cache = self.model(inputs, cache)
         return self.lm_head(out), cache
+
+    @property
+    def layers(self):
+        return self.model.layers

llms/mlx_lm/models/olmo.py

@@ -178,3 +178,7 @@ class Model(nn.Module):
         cache=None,
     ):
         return self.model(inputs, cache)
+
+    @property
+    def layers(self):
+        return self.model.transformer.blocks

llms/mlx_lm/models/phi.py

@@ -178,3 +178,7 @@ class Model(nn.Module):
 
         y, cache = self.model(x, mask, cache)
         return self.lm_head(y), cache
+
+    @property
+    def layers(self):
+        return self.model.layers

llms/mlx_lm/models/phixtral.py

@@ -216,3 +216,7 @@ class Model(nn.Module):
 
         y, cache = self.transformer(x, mask, cache)
         return self.lm_head(y), cache
+
+    @property
+    def layers(self):
+        return self.transformer.h

llms/mlx_lm/models/qwen2.py

@@ -205,3 +205,7 @@ class Model(nn.Module):
         return {
             k: v for k, v in weights.items() if "self_attn.rotary_emb.inv_freq" not in k
         }
+
+    @property
+    def layers(self):
+        return self.model.layers

llms/mlx_lm/models/stablelm_epoch.py

@@ -184,3 +184,7 @@ class Model(nn.Module):
 
         y, cache = self.model(x, mask, cache)
         return self.lm_head(y), cache
+
+    @property
+    def layers(self):
+        return self.model.layers