Efficient-Large-Model
/

qwen2-7b-longvila-1M

Model card Files Files and versions

Yukang commited on Jan 14, 2025

Commit

be6a428

·

verified ·

1 Parent(s): 9188af7

Upload config.json

Files changed (1) hide show

config.json +3 -3

config.json CHANGED Viewed

@@ -47,10 +47,10 @@
     },
     "length_penalty": 1.0,
     "max_length": 20,
-    "max_position_embeddings": 1048576,
     "max_window_layers": 28,
     "min_length": 0,
-    "model_max_length": 1048576,
     "model_type": "qwen2",
     "no_repeat_ngram_size": 0,
     "num_attention_heads": 28,
@@ -85,7 +85,7 @@
     "tie_encoder_decoder": false,
     "tie_word_embeddings": false,
     "tokenizer_class": null,
-    "tokenizer_model_max_length": 1048576,
     "tokenizer_padding_side": "right",
     "top_k": 50,
     "top_p": 1.0,

     },
     "length_penalty": 1.0,
     "max_length": 20,
+    "max_position_embeddings": 524288,
     "max_window_layers": 28,
     "min_length": 0,
+    "model_max_length": 524288,
     "model_type": "qwen2",
     "no_repeat_ngram_size": 0,
     "num_attention_heads": 28,
     "tie_encoder_decoder": false,
     "tie_word_embeddings": false,
     "tokenizer_class": null,
+    "tokenizer_model_max_length": 524288,
     "tokenizer_padding_side": "right",
     "top_k": 50,
     "top_p": 1.0,