fuser big

README.md

@@ -1,2 +0,0 @@
-hello
-hello

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 sha256:51fcbcfa8f61c4058f667fdc372f98898a4e2e51a48e8b5923823279adb1c0bf
-size 242089535
+oid sha256:ba25c3cb975d128d20278590a8c676b8caa9cccb184aac50ae3ef48a5b448cc7
+size 891739457