Merge pull request #430 from austinbv/patch-1

Add 32b Qwen 2.5

exo/models.py

@@ -43,6 +43,7 @@ model_base_shards = {
   ### qwen
   "qwen-2.5-coder-1.5b": {"MLXDynamicShardInferenceEngine": Shard(model_id="mlx-community/Qwen2.5-Coder-1.5B-Instruct-4bit", start_layer=0, end_layer=0, n_layers=28),},
   "qwen-2.5-coder-7b": {"MLXDynamicShardInferenceEngine": Shard(model_id="mlx-community/Qwen2.5-Coder-7B-Instruct-4bit", start_layer=0, end_layer=0, n_layers=28),},
+  "qwen-2.5-coder-32b": {"MLXDynamicShardInferenceEngine": Shard(model_id="mlx-community/Qwen2.5-Coder-32B-Instruct-4bit", start_layer=0, end_layer=0, n_layers=28),},
   "qwen-2.5-7b": {"MLXDynamicShardInferenceEngine": Shard(model_id="mlx-community/Qwen2.5-7B-Instruct-4bit", start_layer=0, end_layer=0, n_layers=28),},
   "qwen-2.5-math-7b": {"MLXDynamicShardInferenceEngine": Shard(model_id="mlx-community/Qwen2.5-Math-7B-Instruct-4bit", start_layer=0, end_layer=0, n_layers=28),},
   "qwen-2.5-14b": {"MLXDynamicShardInferenceEngine": Shard(model_id="mlx-community/Qwen2.5-14B-Instruct-4bit", start_layer=0, end_layer=0, n_layers=48),},