splitter base

config.json

@@ -2,9 +2,9 @@
   "architectures": [
     "T5ForConditionalGeneration"
   ],
-  "d_ff": 2048,
+  "d_ff": 3072,
   "d_kv": 64,
-  "d_model": 512,
+  "d_model": 768,
   "decoder_start_token_id": 0,
   "dropout_rate": 0.1,
   "eos_token_id": 1,
@@ -13,8 +13,8 @@
   "layer_norm_epsilon": 1e-06,
   "model_type": "t5",
   "n_positions": 512,
-  "num_heads": 8,
-  "num_layers": 6,
+  "num_heads": 12,
+  "num_layers": 12,
   "output_past": true,
   "pad_token_id": 0,
   "relative_attention_num_buckets": 32,

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6f3c61af0924b9b83cf7ea1c8f52c913c0bce5eb98922ef68d52af1330c31c66
-size 242090835
+oid sha256:df465178398f8e4a13cdb1ece74b2ff0c3e7760cc86244fd376d18a1971c049b
+size 891739457