Upload model, config, and model card

1/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7eafd4c3b57ec8980f22603fa255774f4cf214ca91c7ef1ce544157326e82c68
+size 1330968444

README.md

@@ -0,0 +1,19 @@
+---
+language: 
+- en
+thumbnail: 
+tags:
+- question-answering
+- torchscript
+- FastNN
+license: mit
+datasets:
+- squad
+metrics:
+---
+
+# TorchScript model of bert-large-cased-whole-word-masking-finetuned-squad
+
+## Model description
+
+A serialized torchscript model of bert-large-cased-whole-word-masking-finetuned-squad with a config.pbtxt for deployment using NVIDIA Triton Inference Server.

config.pbtxt

@@ -0,0 +1,61 @@
+name: "bert-large-cased-whole-word-masking-finetuned-squad"
+platform: "pytorch_libtorch"
+max_batch_size: 64
+input [
+  {
+    name: "input__0"
+    data_type: TYPE_INT64
+    dims: [512]
+  },
+  {
+    name: "input__1"
+    data_type: TYPE_INT64
+    dims: [512]
+  },
+  {
+    name: "input__2"
+    data_type: TYPE_INT64
+    dims: [512]
+  },
+  {
+    name: "input__3"
+    data_type: TYPE_INT64
+    dims: [1]
+  },
+  {
+    name: "input__4"
+    data_type: TYPE_INT64
+    dims: [1]
+  },
+  {
+    name: "input__5"
+    data_type: TYPE_FP32
+    dims: [512]
+  }
+]
+output [
+  {
+    name: "output__0"
+    data_type: TYPE_FP32
+    dims: [512]
+  },
+  {
+    name: "output__1"
+    data_type: TYPE_FP32
+    dims: [512]
+  },
+  {
+    name: "output__2"
+    data_type: TYPE_INT64
+    dims: [1]
+  }
+]
+dynamic_batching {
+  preferred_batch_size: [ 1,2,4,8,16,32,64 ]
+  max_queue_delay_microseconds: 30000
+
+}
+version_policy: { latest { num_versions : 1 }}
+optimization {
+  graph { level: 1 }
+}