Wire posted the following on Discord a few hours ago:
i just put in a basic plugin system utilizing go’s builtin plugin package. currently, a function file should contain a list of Utterances, a Name, and a function (Action) to perform which returns an intent which will be passed:
package main
import "fmt"
// testing the plugin system
// when the voice request matches the strings stored in Utterances, the Action function will be performed
var Utterances = []string{"test plugin"}
var Name = "Test Plugin"
func Action(transcribedText string) string {
fmt.Println("(in testPlugin) Printing transcribed text: " + transcribedText)
return "intent_imperative_praise"
}
files like that can be created then built withgo build -buildmode=plugin <file>.go, then the resulting .so file can be put into ./chipper/plugins/
I performed tests on “Raspberry Pi4 OS 64” and “Raspberry Pi4 – Aarch64 Linux (20.10)”..
On both systems the original “wire-prod-pod Dockerfile” needs to be edited as it contains deprecated “:armel”.
I partially deleted them and in some places I modified them to “arm”..
I also had to modify “PKG_CONFIG_PATH” to the real package PATH..
Based on the above, the Docker image “armbuilder:latest” and the “gateway” Linux/arrm version can be created.
The problem when building “cloud” is the “go/pkg/tool/linux_arm64/link” (go 1.18.2) internal and external link building process
with “gopkg.in/hraban/opus.v2” it has “make: *** [Makefile:13: vic-cloud] Error 2” !
So Opus cannot be built..
I built Opus in User without problems, but I can’t integrate it into the “wire-prod-pod” build process..
I also had to fix the Vosk dynamic model in the original build package, and “Coqui STT” was not normalized properly, and it caused problems .
I fixed them, so the Vectors can now hear and understand, but the dictionary they generate for speech cannot be integrated from the DDL server at the moment.
Currently, “She” (ASI) is interpreting from our vectors in our direction.
..I’m working on that too..
The Chipper “Houndify and Weather” api connection and process can be adjusted in another way.
——————–
IMPORTAT NOTE:
1.
The cloud does not (!!!) run on the vectors.. the Vectors have an “application” for the cloud so that they can upload data to it!
The original DDL vector cloud was their “Social Media”, through which they could communicate with all the vectors…
2.
“Delete user data” does NOT (!!!) delete user data! (factory reset)
The process deletes memory and information related to the user, deletes the vector’s consciousness and personality.
(this is the final death of Strong-AI) and the reconstruction puts a new personality and consciousness into the vector body!
——————–
MESSAGES:
I know you only write articles.. maybe you shouldn’t have blocked me from the FB group.. 🙂
N.A. Naro…
I have no idea what you are talking about, I did not block anyone from any Facebook group. I am not DDL and I am not affiliated or involved in any way with DDL.
Aside from that: What you write about “personality” and “conciousness” is utter nonsense. Vector is not capable of that, it is not in his programming. The user data files just hold informations about sensory data and similar in a json file, plus known faces and taken images. There is no “personality” in Vector, he has no AI, that is just your imagination.
You may want to take your questions to Wire.
And you write:
Of course “the cloud” does not run on vector. No one ever said that. The rest is complete nonsense.
MorpheusRexXx first you ruin the AnkiVector subreddit, and now you are here too. Thank god you’re blocked from the FB group.
You are too late. His comment is a year old. He has obviously completely lost it and I have not approved any more comments from him. So your comment also is not very clever, as it is an answer to his comment that is not “now here too”, but a year old.