Update repository

.gitattributes

@@ -25,3 +25,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zstandard filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.th filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,6 @@
+---
+tags:
+- allennlp
+---
+
+# TODO: Fill this model card

config.json

@@ -0,0 +1,160 @@
+{
+    "dataset_reader": {
+        "type": "drop",
+        "instance_format": "drop",
+        "passage_length_limit": 400,
+        "question_length_limit": 50,
+        "skip_when_all_empty": [
+            "passage_span",
+            "question_span",
+            "addition_subtraction",
+            "counting"
+        ],
+        "token_indexers": {
+            "token_characters": {
+                "type": "characters",
+                "min_padding_length": 5
+            },
+            "tokens": {
+                "type": "single_id",
+                "lowercase_tokens": true
+            }
+        }
+    },
+    "model": {
+        "type": "naqanet",
+        "answering_abilities": [
+            "passage_span_extraction",
+            "question_span_extraction",
+            "addition_subtraction",
+            "counting"
+        ],
+        "dropout_prob": 0.1,
+        "matrix_attention_layer": {
+            "type": "linear",
+            "combination": "x,y,x*y",
+            "tensor_1_dim": 128,
+            "tensor_2_dim": 128
+        },
+        "modeling_layer": {
+            "type": "qanet_encoder",
+            "attention_dropout_prob": 0,
+            "attention_projection_dim": 128,
+            "conv_kernel_size": 5,
+            "dropout_prob": 0.1,
+            "feedforward_hidden_dim": 128,
+            "hidden_dim": 128,
+            "input_dim": 128,
+            "layer_dropout_undecayed_prob": 0.1,
+            "num_attention_heads": 8,
+            "num_blocks": 6,
+            "num_convs_per_block": 2
+        },
+        "num_highway_layers": 2,
+        "phrase_layer": {
+            "type": "qanet_encoder",
+            "attention_dropout_prob": 0,
+            "attention_projection_dim": 128,
+            "conv_kernel_size": 7,
+            "dropout_prob": 0.1,
+            "feedforward_hidden_dim": 128,
+            "hidden_dim": 128,
+            "input_dim": 128,
+            "layer_dropout_undecayed_prob": 0.1,
+            "num_attention_heads": 8,
+            "num_blocks": 1,
+            "num_convs_per_block": 4
+        },
+        "regularizer": {
+            "regexes": [
+                [
+                    ".*",
+                    {
+                        "alpha": 1e-07,
+                        "type": "l2"
+                    }
+                ]
+            ]
+        },
+        "text_field_embedder": {
+            "token_embedders": {
+                "token_characters": {
+                    "type": "character_encoding",
+                    "embedding": {
+                        "embedding_dim": 64
+                    },
+                    "encoder": {
+                        "type": "cnn",
+                        "embedding_dim": 64,
+                        "ngram_filter_sizes": [
+                            5
+                        ],
+                        "num_filters": 200
+                    }
+                },
+                "tokens": {
+                    "type": "embedding",
+                    "embedding_dim": 300,
+                    "pretrained_file": "https://allennlp.s3.amazonaws.com/datasets/glove/glove.840B.300d.lower.converted.zip",
+                    "trainable": false
+                }
+            }
+        }
+    },
+    "train_data_path": "https://s3-us-west-2.amazonaws.com/allennlp/datasets/drop/drop_dataset.zip!drop_dataset/drop_dataset_train.json",
+    "validation_data_path": "https://s3-us-west-2.amazonaws.com/allennlp/datasets/drop/drop_dataset.zip!drop_dataset/drop_dataset_dev.json",
+    "trainer": {
+        "callbacks": [
+            "tensorboard"
+        ],
+        "grad_norm": 5,
+        "moving_average": {
+            "type": "exponential",
+            "decay": 0.9999
+        },
+        "num_epochs": 50,
+        "optimizer": {
+            "type": "adam",
+            "betas": [
+                0.8,
+                0.999
+            ],
+            "eps": 1e-07,
+            "lr": 0.0005
+        },
+        "patience": 10,
+        "validation_metric": "+f1"
+    },
+    "vocabulary": {
+        "min_count": {
+            "token_characters": 200
+        },
+        "only_include_pretrained_words": true,
+        "pretrained_files": {
+            "tokens": "https://allennlp.s3.amazonaws.com/datasets/glove/glove.840B.300d.lower.converted.zip"
+        }
+    },
+    "data_loader": {
+        "batch_sampler": {
+            "type": "bucket",
+            "batch_size": 16
+        }
+    },
+    "validation_dataset_reader": {
+        "type": "drop",
+        "instance_format": "drop",
+        "passage_length_limit": 1000,
+        "question_length_limit": 100,
+        "skip_when_all_empty": [],
+        "token_indexers": {
+            "token_characters": {
+                "type": "characters",
+                "min_padding_length": 5
+            },
+            "tokens": {
+                "type": "single_id",
+                "lowercase_tokens": true
+            }
+        }
+    }
+}

log/train/events.out.tfevents.1641518630.allennlp-server4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1168859509f2ce13825815b6ba6100912e3e4a8308acfe3e53e2b5e6f0286745
+size 775836

log/validation/events.out.tfevents.1641518630.allennlp-server4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49f9f6ee3e67afd73090e6e9b491af6fe040c52ef8b45530a72b102800df142a
+size 10640

metrics.json

@@ -0,0 +1,21 @@
+{
+  "best_epoch": 45,
+  "peak_worker_0_memory_MB": 11193.4921875,
+  "peak_gpu_0_memory_MB": 6340.828125,
+  "training_duration": "13:11:18.777333",
+  "epoch": 49,
+  "training_em": 0.5403715897718537,
+  "training_f1": 0.5968561358724769,
+  "training_loss": 1.685971354938646,
+  "training_reg_loss": 0.006671493937393214,
+  "training_worker_0_memory_MB": 11193.4921875,
+  "training_gpu_0_memory_MB": 6132.47216796875,
+  "validation_em": 0.46476510067114096,
+  "validation_f1": 0.49971371644295354,
+  "validation_loss": Infinity,
+  "validation_reg_loss": 0.0,
+  "best_validation_em": 0.4634018456375839,
+  "best_validation_f1": 0.4998416526845643,
+  "best_validation_loss": Infinity,
+  "best_validation_reg_loss": 0.0
+}

vocabulary/non_padded_namespaces.txt

@@ -0,0 +1,2 @@
+*tags
+*labels

vocabulary/token_characters.txt

@@ -0,0 +1,118 @@
+@@UNKNOWN@@
+e
+a
+t
+n
+o
+r
+i
+s
+h
+d
+l
+c
+u
+f
+g
+w
+m
+y
+p
+,
+.
+b
+k
+v
+1
+-
+T
+0
+2
+B
+C
+S
+3
+4
+R
+D
+A
+5
+9
+M
+q
+I
+W
+7
+6
+P
+H
+8
+F
+J
+L
+G
+N
+'
+E
+%
+x
+)
+(
+?
+O
+z
+K
+V
+U
+j
+Q
+"
+;
+:
+Y
+&
+/
+#
+$
+–
+Z
+²
+é
+X
+●
+á
+ü
+—
+ó
+ł
+í
+ö
++
+£
+ō
+°
+ä
+š
+è
+ć
+ç
+’
+Š
+ø
+€
+−
+ė
+É
+å
+ú
+č
+ñ
+ı
+Ö
+â
+ž
+ă
+ū
+Ō
+ș
+Á

weights.th

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ccfded94830c74c26457a4bc010c9621664d2ac16bfe3dec0173317ff1fe2595
+size 64354621