Add float16 to MLOperandType and ArrayBufferView compatibility table

- Note TC39 Float16Array proposal is WIP
- Update issue links

Related #373

(Bikeshed syntax not supported inside tables, thus plain HTML.)

Author: anssiko

index.bs

@@ -3152,6 +3152,9 @@ console.log('Output value: ' + result.outputs.output);
     <tr>
         <td>{{MLOperandType/float32}}
         <td>{{Float32Array}}
+    <tr>
+        <td>{{MLOperandType/float16}}
+        <td><p class="issue">TC39 <a href="https://tc39.es/proposal-float16array/">`Float16Array`</a> is work in progress. <a href="https://github.com/webmachinelearning/webnn/issues/373">[Issue webnn#373]</a></p>
     <tr>
         <td>{{MLOperandType/int32}}
         <td>{{Int32Array}}
@@ -3166,8 +3169,6 @@ console.log('Output value: ' + result.outputs.output);
         <td>{{Uint8Array}}
 </table>
 
-Issue(webmachinelearning/webnn#127): clarify the usage of {{ArrayBufferView}} for {{MLOperandType/float16}}.
-
 <h2 id="acknowledgements">Acknowledgements</h2>
 
 This specification follows the concepts of the Android Neural Networks API C