update model checkpoint

config.json

@@ -7,12 +7,14 @@
   "d_kv": 64,
   "d_model": 1472,
   "decoder_start_token_id": 0,
+  "dense_act_fn": "gelu_new",
   "dropout_rate": 0.1,
   "eos_token_id": 1,
   "feed_forward_proj": "gated-gelu",
   "gradient_checkpointing": false,
   "initializer_factor": 1.0,
   "is_encoder_decoder": true,
+  "is_gated_act": true,
   "layer_norm_epsilon": 1e-06,
   "model_type": "t5",
   "num_decoder_layers": 4,
@@ -23,8 +25,8 @@
   "relative_attention_num_buckets": 32,
   "tie_word_embeddings": false,
   "tokenizer_class": "ByT5Tokenizer",
-  "torch_dtype": "bfloat16",
-  "transformers_version": "4.18.0",
+  "torch_dtype": "float32",
+  "transformers_version": "4.20.1",
   "use_cache": true,
   "vocab_size": 384
 }

detail.json

@@ -0,0 +1,6 @@
+{
+    "avg_loss": 0.15668370458297431,
+    "step": 801983,
+    "best_loss": 0.15668370458297431,
+    "lr": 0.0003
+}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:643d73436994fb62530ca417473791e35666b52492e64776d60950c7f50920df
-size 599299947
+oid sha256:4535f6480711aa85def46839173baefe2d9fb9a396e3fbc4dacee4ae065d28cc
+size 1198608045