Training in progress, step 500

added_tokens.json

@@ -12,15 +12,5 @@
   ":ARGM-MNR": 32108,
   ":ARGM-NEG": 32113,
   ":ARGM-PRP": 32109,
-  ":ARGM-TMP": 32107,
-  ":graph1": 32114,
-  ":graph10": 32123,
-  ":graph2": 32115,
-  ":graph3": 32116,
-  ":graph4": 32117,
-  ":graph5": 32118,
-  ":graph6": 32119,
-  ":graph7": 32120,
-  ":graph8": 32121,
-  ":graph9": 32122
+  ":ARGM-TMP": 32107
 }

config.json

@@ -1,11 +1,11 @@
 {
-  "_name_or_path": "t5-small",
+  "_name_or_path": "t5-large",
   "architectures": [
     "T5ForConditionalGeneration"
   ],
-  "d_ff": 2048,
+  "d_ff": 4096,
   "d_kv": 64,
-  "d_model": 512,
+  "d_model": 1024,
   "decoder_start_token_id": 0,
   "dense_act_fn": "relu",
   "dropout_rate": 0.1,
@@ -17,9 +17,9 @@
   "layer_norm_epsilon": 1e-06,
   "model_type": "t5",
   "n_positions": 512,
-  "num_decoder_layers": 6,
-  "num_heads": 8,
-  "num_layers": 6,
+  "num_decoder_layers": 24,
+  "num_heads": 16,
+  "num_layers": 24,
   "output_past": true,
   "pad_token_id": 0,
   "relative_attention_max_distance": 128,
@@ -54,7 +54,7 @@
     }
   },
   "torch_dtype": "float32",
-  "transformers_version": "4.26.1",
+  "transformers_version": "4.27.3",
   "use_cache": true,
-  "vocab_size": 32124
+  "vocab_size": 32114
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3e974ce2a7ef3ca156a0d1a53de155fb1888878d2b50f8c85023f36f9ec6d737
-size 242063449
+oid sha256:e03d92b495880b29acf2dd0169c643fa8d587d387644233cf69c35317c3f6728
+size 2950791169

runs/Mar28_11-55-35_d5f806247e8b/1680004539.0123274/events.out.tfevents.1680004539.d5f806247e8b.17127.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3905897a2e5a058794a58beedbb7cb33019bc193b3ad7e22e7dbc995a6f29c82
+size 5981

runs/Mar28_11-55-35_d5f806247e8b/events.out.tfevents.1680004538.d5f806247e8b.17127.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f911b8b5b60d6b65c5e03373723c63b2fa08ac902a503485357b21e4b56f8554
+size 4976

special_tokens_map.json

@@ -13,17 +13,7 @@
     ":ARGM-CAU",
     ":ARGM-DIR",
     ":ARGM-EXT",
-    ":ARGM-NEG",
-    ":graph1",
-    ":graph2",
-    ":graph3",
-    ":graph4",
-    ":graph5",
-    ":graph6",
-    ":graph7",
-    ":graph8",
-    ":graph9",
-    ":graph10"
+    ":ARGM-NEG"
   ],
   "eos_token": "</s>",
   "pad_token": "<pad>",

tokenizer.json

@@ -1067,96 +1067,6 @@
       "rstrip": false,
       "normalized": false,
       "special": true
-    },
-    {
-      "id": 32114,
-      "content": ":graph1",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 32115,
-      "content": ":graph2",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 32116,
-      "content": ":graph3",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 32117,
-      "content": ":graph4",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 32118,
-      "content": ":graph5",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 32119,
-      "content": ":graph6",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 32120,
-      "content": ":graph7",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 32121,
-      "content": ":graph8",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 32122,
-      "content": ":graph9",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 32123,
-      "content": ":graph10",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": false,
-      "special": true
     }
   ],
   "normalizer": {

tokenizer_config.json

@@ -104,7 +104,6 @@
   "eos_token": "</s>",
   "extra_ids": 100,
   "model_max_length": 512,
-  "name_or_path": "t5-small",
   "pad_token": "<pad>",
   "special_tokens_map_file": null,
   "tokenizer_class": "T5Tokenizer",

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a30bd09f4afc5c21d55974f45dcdf6d0bc7c83e942122064a692247db45e17d0
-size 3579
+oid sha256:3e4675273a31e5eebae2b1b2c3989e5c7676ef6477893030095707614a5e8872
+size 3707