Feature: Add Half (float16) to data inspector

Wishlists for new functionality and features.
Post Reply
imadewins
Posts: 2
Joined: 08 Apr 2019 21:33

Feature: Add Half (float16) to data inspector

Post by imadewins » 08 Apr 2019 21:35

As the title says, this would be a super handy feature and be easy to add... (IEEE 754)

Maël
Site Admin
Posts: 1071
Joined: 12 Mar 2005 14:15

Re: Feature: Add Half (float16) to data inspector

Post by Maël » 09 Apr 2019 05:19

Will not be easy (since I need to make a special decoder for float16 to string values and back -- this doesn't come by default in most languages).

But should be well documented and useful.

P.S.: What would you need it for? I mostly know of float16 being used in deep learning, but I don't see why you'd need a hex editor for that.

imadewins
Posts: 2
Joined: 08 Apr 2019 21:33

Re: Feature: Add Half (float16) to data inspector

Post by imadewins » 09 Apr 2019 05:40

Maël wrote:
09 Apr 2019 05:19
Will not be easy (since I need to make a special decoder for float16 to string values and back -- this doesn't come by default in most languages).

But should be well documented and useful.

P.S.: What would you need it for? I mostly know of float16 being used in deep learning, but I don't see why you'd need a hex editor for that.
3D models and compressed binary animation formats use half precision floats all the time to save space, would be super helpful as I use HxD all the time for viewing the binary dumps :)

Post Reply