Model: Change cache_size/max_seq_len behavior

- Cache size is now given only by the cache_size config option. Default is 4096 (user should always override to max out VRAM)
- max_seq_len, if not overridden in the config, will default to the model's config.json
- max_seq_len is reduced to be no larger than the cache

common/model.py

@@ -157,10 +157,8 @@ async def load_model_gen(model_path: pathlib.Path, **kwargs):
 
     # Override the max sequence length based on user
     max_seq_len = kwargs.get("max_seq_len")
-    if max_seq_len == -1:
+    if max_seq_len == -1 or max_seq_len is None:
         kwargs["max_seq_len"] = hf_model.hf_config.max_position_embeddings
-    elif max_seq_len is None:
-        kwargs["max_seq_len"] = 4096
 
     # Create a new container and check if the right dependencies are installed
     backend = unwrap(kwargs.get("backend"), detect_backend(hf_model))