Training in progress epoch 0

config.json

@@ -1,11 +1,11 @@
 {
-  "_name_or_path": "google/mt5-small",
+  "_name_or_path": "google/mt5-base",
   "architectures": [
     "MT5ForConditionalGeneration"
   ],
-  "d_ff": 1024,
+  "d_ff": 2048,
   "d_kv": 64,
-  "d_model": 512,
+  "d_model": 768,
   "decoder_start_token_id": 0,
   "dense_act_fn": "gelu_new",
   "dropout_rate": 0.1,
@@ -16,9 +16,10 @@
   "is_gated_act": true,
   "layer_norm_epsilon": 1e-06,
   "model_type": "mt5",
-  "num_decoder_layers": 8,
-  "num_heads": 6,
-  "num_layers": 8,
+  "num_decoder_layers": 12,
+  "num_heads": 12,
+  "num_layers": 12,
+  "output_past": true,
   "pad_token_id": 0,
   "relative_attention_max_distance": 128,
   "relative_attention_num_buckets": 32,

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7d165e2c4379faea6aea0be2e84a577b0db94b87d57ad5090e112362152388ba
-size 1200772485
+oid sha256:81c9b7274f957080b3564692ab9421e47252bf3b99c3b568962ecba53edd3a64
+size 2329702453