config

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "bart_test",
+  "_name_or_path": "ccdv/lsg-bart-base-16384",
   "activation_dropout": 0.1,
   "activation_function": "gelu",
   "adaptive": true,
@@ -18,7 +18,7 @@
     "AutoModelForSequenceClassification": "modeling_lsg_bart.LSGBartForSequenceClassification"
   },
   "base_model_prefix": "lsg",
-  "block_size": 128,
+  "block_size": 256,
   "bos_token_id": 0,
   "classif_dropout": 0.1,
   "classifier_dropout": 0.0,
@@ -54,19 +54,19 @@
   "mask_first_token": false,
   "max_position_embeddings": 16384,
   "model_type": "bart",
-  "no_repeat_ngram_size": 3,
+  "no_repeat_ngram_size": null,
   "normalize_before": false,
   "normalize_embedding": true,
-  "num_beams": 4,
+  "num_beams": 5,
   "num_global_tokens": 1,
   "num_hidden_layers": 6,
   "pad_token_id": 1,
   "pass_global_tokens_to_decoder": true,
   "pool_with_global": true,
   "scale_embedding": false,
-  "sparse_block_size": 128,
-  "sparsity_factor": 2,
-  "sparsity_type": "norm",
+  "sparse_block_size": 0,
+  "sparsity_factor": 4,
+  "sparsity_type": "none",
   "task_specific_params": {
     "summarization": {
       "length_penalty": 1.0,