better

nanovllm/config.py

@@ -1,10 +1,11 @@
+import os
 from dataclasses import dataclass
 from transformers import AutoConfig
 
 
 @dataclass
 class Config:
-    model: str = ''
+    model: str
     max_num_batched_tokens: int = 32768
     max_num_seqs: int = 512
     max_model_len: int = 4096
@@ -17,5 +18,8 @@ class Config:
     num_kvcache_blocks: int = -1
 
     def __post_init__(self):
-        assert self.model
+        assert os.path.isdir(self.model)
         assert self.kvcache_block_size % 256 == 0
+        assert 1 <= self.tensor_parallel_size <= 8
+        self.hf_config = AutoConfig.from_pretrained(self.model)
+        self.max_model_len = min(self.max_model_len, self.hf_config.max_position_embeddings)