I was woundering if there was a way to increase the prompt length and output and if so how i can do it.
Thanks

I want it to use binary code and talk like a cat pls pls pls plsssss

See this PR here for more detail
ggerganov/llama.cpp#732

You can either add the translation layer into the C++ code or the JavaScript frontend.

Can I use a larger model? Can I use display-based reasoning? May be I can have these options.

Hey seems like the gpt is not made to deal with memory. it can not remember context. how can i make it to remember context ?

https://github.com/ohmplatform/FreedomGPT/releases/download/v1.1.2/freedomgpt-1.1.2.Setup.exe

it says it has access to latest info. but ask for today's date. it goes blank

Hi
When I use 1.1.3, the size cannot be recognized. It's not clear to me what the problem is.

As a user of FreedomGPT, I would like to suggest adding the ability to start a new conversation without having to quit and restart the tool. Currently, if I want to start a new conversation, I have to quit the tool and then restart it.

This feature would be particularly useful for users who want to switch between conversations or start a new conversation without losing the context of the previous conversation. It would make FreedomGPT more user-friendly and bring it in line with other conversational AI tools that allow for starting a new conversation from within the same interface.

It would be nice to have bigger unrestricted model but I do not know how to train it not to follow ethical standards.
Can someone do a guide for that?

I find emoji in index.js: 🗣 and 🗽.
I... actually wish to speak with something else other than Statue of Liberty 😂

Consider this is JavaScript, maybe add support to paste base64 image in somewhere for example settings to change human and AI avatar?

I made the app work on windows, but am wondering, if its intended, that there is no context between individual queries like ChatGPT has?

when i ask with any language(such chinese), freedomGPT only answer English, how to become answer with ask language?

Description

I've been waiting for an hour, and there's no response. How many hours it typically takes to load a model?

Stuck on this screen for an hour. But the Websocket request still receive and sends payload.

Memory usage is also low. There's no heavy activity inside the FreedomGPT.

Stack

• FreedomGPT 1.1.3
• Windows 10
• 11th Gen Intel(R) Core(TM) i7-1165G7 @ 2.80GHz 2.80 GHz
• 16,0 GB

Similar Issues

• #72

Any option for spanish offline?

I have started the servers and the application, and when I write a message, I get no reply, only these error messages

Use the main directory, instead of the home path.I don't like it when the c disk is cluttered, and even more so the weights can weigh a lot.

//index.ts
//before:
const homeDir = app.getPath("home");
const MODEL_LOCATION = homeDir + "/FreedomGPT";
//after:
const rootDir = process.cwd();
const MODEL_LOCATION = rootDir + "/weights";

I asked if it can generate images, and said yes, but when it sends me an image, it delivers nothing, an empty space.

so it can't?

Hi, seems that this use the standard alpaca 7B weights. But when I use this weights with llama.cpp it seems to be “censored”.

So what is the Pre Prompt or how to make use this in the same way with llama.cpp?

I've just downloaded the client and it seems to be stuck in an infinite loop:

That is to say it never answers and seems to be stuck as such.

By the way, if this is a true project I would be really happy to help!

I would like to suggest an improvement to Freedom GPT. It would be great if the model could be executed on Colab, This would allow for easy access and execution of the model on a cloud-based environment, without the need for local installations.

Please consider this suggestion and let me know if there are any plans to implement it in the future. Thank you for your hard work on Freedom GPT!

Ask him to answer in Chinese, but the Chinese answer is garbled and unable to contact the context, unable to remember the previous question and answer
Ask him to answer in Chinese, but the Chinese answer is garbled and unable to contact the context, unable to remember the previous question and answer

I saw that it's using the cpu only which is slower.
Can you implement a way to choose if the user wants to use cpu or gpu?

I'm stuck to the loading screen after selecting the AI Model.

""DevTools failed to load source map: Could not parse content for file:///C:/Users/Username/AppData/Local/FreedomGPT/app-1.1.3/resources/app/.webpack/renderer/main_window/index.js.map: Unexpected end of JSON input"

PC specs:
16 GB RAM DDR3
Windows 10
AMD FX8320 3.5 ghz

First my English is not the best, i have instald it yesterday and i have tested it, i think the freedomgpt have a realy good potencial but in this inicial phase i found some things that can be adjusted:

-newer gpt version(gpt neo)
-beter translation
-memory
-speed optimize

So the freedomgpt ia using a variant of gpt-2, i think this gpt version is a litle obsolent and some of their answeres arent good enought, gpt 3 or 3.5 have a realy Better potencial and performance improvements, gpt 4 is good but very slower:
Newer gpt version (3.5)

The freedomgpt seems to get slower when i talk in some non-english language, and all the Times i talk to him in portuguese he talk to me in English or spanhich, some translater can be an option like detect the language, traduct the prompt and give it to gpt, gpt give a new prompt to the translater to traduct to the language used by the user:
Beter translation and speed.

Ik the freedomgpt cant memorize the prompts, bc if he did her answers can get nonsense, Thats becouse of some gpt-2 limitations, this can be changed by using a newer gpt version or something i cant think now but probably worse that a newer gpt:
Memorysing prompts.

For now this seems to be realy much slower that chatgpt, im using a r5 3600 rx5700xt 16g RAM and SSD, this can be inproved by adding setings to chose use cpu or gpu or maybe the two. Maybe cpu for translater and gpu for gpt if u can not use botw for gpt:
Speed optimize.

Update 1:

I have searched more and l have found that gpt 3 have not an opensource version so the newer version Open sourse is "GPT-NEO" search for it i think this project can be much Better if it use a powerful gpt version.
This can be an option to enable, becouse its realy Better but is slower.
And my solution to the translation, is good but it need to be an option becouse if u ask something and gpt need to know the language ur talking or if u ask him to translate something u gonna get a realy wrong answer.

Update 2:

the diference betwin freedomgpt and chat gpt making the skane game in console c#:

prompt: can u make me a console c# code to a snake game?

freedomgpt:

using System;

class SnakeGame {
public static void Main()
{
int n1 = Console.ReadLine(); // Number of players

// Drawing the screen with score and other info's
Console.SetTitle("SNAKE GAME");

for (int i=0;i

chatgpt:
using System;
using System.Collections.Generic;
using System.Linq;

class SnakeGame
{
static void Main(string[] args)
{
Console.CursorVisible = false;
Console.SetWindowSize(40, 20);
Console.SetBufferSize(40, 20);

    var snake = new List<Point>()
    {
        new Point(10, 10),
        new Point(9, 10),
        new Point(8, 10),
        new Point(7, 10),
        new Point(6, 10)
    };

    var food = GenerateFood(snake);
    ...
    ...
    ....
    The entire game code working

and i can not ask freedomgpt to continue the code becouse he has no memory.

I just noticed when installing that the Application is installed in AppData\Local
but the Model itself is in the Users directory.
for me personally that's problematic because i don't have much space on my C drive anymore and since the model is quite big (and i assume it will rather go bigger than smaller) one should be able to put it on a bigger hard drive than C.

i think it's quite common sense to only use the C drive being a very fast but as a result also small drive for programs and the operation system while mass data lies on bigger HDDs, so it'd be nice to make the path to the model .bin file configurable somehow.

either that, or at least put it in the same directory as the application is and use a relative path, so one could simply move it from AppData\Local to any other location.

Have written a simple prompt (What is 9 + 10?). Judging by the CPU usage, it's not doing anything other than blinking the cursor.

build log

The community wants to support this project but there is a lot of doubt. Can we please post or reference the code for the binary that is used to spawn model?

In the repository's about section, "This codebase is for an React..." , should be "This codebase is for a React...".

I've followed the instruction as far as I can tell and i get the GUI to open but it sticks at 0MB downloaded

Any ideas?

All fine and well until trying to run the application:

C:\Users\User\freedom-gpt-electron-app>yarn start
yarn run v1.22.19
$ electron-forge start
"electron-forge" is not recognized as an internal or external command,
program or executable batch file.
error Command failed with exit code 1.
info Visit https://yarnpkg.com/en/docs/cli/run for documentation about this command.

From what I can see, FreedomGPT is just an electron frontend for alpaca, but the "About" section of this repository says that it "executes the FreedomGPT LLM".

I suggest editing that in order to not mislead users or be more precise about what this is and what it isn't.

I used this Windows build of Freedom GPT:
https://github.com/ohmplatform/freedom-gpt-electron-app/issues/13

90% or more of the time, if I instruct the AI to write a story, the story starts off great and interesting but halfway through it will start repeating the same few lines of text over and over and over again. I'd rather it just stop writing.

Is this something that you are aware of and is there anything you can do to make this behavior stop?

I recently tried out the FreedomGPT app on Windows.
my specs:
CPU: Intel i9-10900K 3.70GHz 20 CPUs
16 RAM
Graphics: NVidia GeForce GTX 1660 SUPER

i tried typing a command which fits 2 lines, so 25+ words. but the AI didn't even start typing. it was just on hold. I checked the GPU and it was not even using 10% of my GPU. and nothing of my CPU. ideally it'd use both to the maximum extend.

also i noticed that when minimizing the app the GPU consumption ends completely. so it's kind of paused. ideally it should run in background so that you can at least watch a video or search something in the browser whiel you're waiting for a response.

i think a lot of potential users have no ML machine, so it'd make sense to optimize. i hope what i suggest is not too complicated. please give me some feedback.

Hi.

I hope you are having a great new year so far.

Is there a task list for what is needed for Intel Mac support? Maybe some of us could help?

Thanks in advance.

It downloaded me 4GB of dependancies. I tried Linux and Windows release. So how to delete all the dependancies now? From Windows and from Linux?

forge.config.ts:5:32 - error TS2307: Cannot find module './devconst' or i…

5 import { DEVELOPER_DATA } from "./devconst";
~~~~~~~~~~~~
◼ Preparing native dependencies
◼ Running generateAssets hook

An unhandled rejection has occurred inside Forge:
TSError: ⨯ Unable to compile TypeScript:
forge.config.ts:5:32 - error TS2307: Cannot find module './devconst' or its corresponding type declarations.

5 import { DEVELOPER_DATA } from "./devconst";
~~~~~~~~~~~~
at createTSError (/local/node/freedom-gpt/node_modules/ts-node/src/index.ts:859:12)
at reportTSError (/local/node/freedom-gpt/node_modules/ts-node/src/index.ts:863:19)
at getOutput (/local/node/freedom-gpt/node_modules/ts-node/src/index.ts:1077:36)
at Object.compile (/local/node/freedom-gpt/node_modules/ts-node/src/index.ts:1433:41)
at Module.m._compile (/local/node/freedom-gpt/node_modules/ts-node/src/index.ts:1617:30)
at Module._extensions..js (node:internal/modules/cjs/loader:1213:10)
at Object.require.extensions. [as .ts] (/local/node/freedom-gpt/node_modules/ts-node/src/index.ts:1621:12)
at Module.load (node:internal/modules/cjs/loader:1037:32)
at Function.Module._load (node:internal/modules/cjs/loader:878:12)
at Module.require (node:internal/modules/cjs/loader:1061:19)
at require (node:internal/modules/cjs/helpers:103:18)
at exports.default (/local/node/freedom-gpt/node_modules/@electron-forge/core/src/util/forge-config.ts:127:22)
at async Task.task (/local/node/freedom-gpt/node_modules/@electron-forge/core/src/api/start.ts:63:29)
at async Task.run (/local/node/freedom-gpt/node_modules/listr2/dist/index.cjs:978:11)
at async /local/node/freedom-gpt/node_modules/p-map/index.js:57:22
error Command failed with exit code 1.
info Visit https://yarnpkg.com/en/docs/cli/run for documentation about this command.

The error message is The program is not executable.
My environment is
almalinux 8
node.js version
v18.12.1
The directory is
node_modules
src
assets
build
The
What am I missing?

after pnpm i ,a error found when run npm start or npm build make-win.

d:\nodejs\yarn.cmd run start
yarn run v1.22.15
$ electron-forge start
√ Checking your system
√ Locating application
√ Locating application
× Loading configuration
  › ⨯ Unable to compile TypeScript:
  forge.config.ts:51:5 - error TS2322: Type 'MakerZIP' is not assignable to type 'ForgeConfigMaker'.
    Property 'config' is missing in type 'MakerZIP' but required in type 'IForgeResolvableMaker'.
    51     new MakerZIP({}, ["darwin"]),
           ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    node_modules/@electron-forge/shared-types/dist/index.d.ts:109:5
    109     config: any;
            ~~~~~~
    'config' is declared here.
    forge.config.ts:51:18 - error TS2554: Expected 0 arguments, but got 2.
    51     new MakerZIP({}, ["darwin"]),
                        ~~~~~~~~~~~~~~
    forge.config.ts:52:5 - error TS2322: Type 'MakerDMG' is not assignable to type 'ForgeConfigMaker'.
    Property 'config' is missing in type 'MakerDMG' but required in type 'IForgeResolvableMaker'.
    52     new MakerDMG({}, ["darwin"]),
           ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    node_modules/@electron-forge/shared-types/dist/index.d.ts:109:5
    109     config: any;
            ~~~~~~
    'config' is declared here.
    forge.config.ts:52:18 - error TS2554: Expected 0 arguments, but got 2.
    52     new MakerDMG({}, ["darwin"]),
                        ~~~~~~~~~~~~~~
    forge.config.ts:53:5 - error TS2322: Type 'MakerSquirrel' is not assignable to type 'ForgeConfigMaker'.
    Property 'config' is missing in type 'MakerSquirrel' but required in type 'IForgeResolvableMaker'.
     53     new MakerSquirrel(
            ~~~~~~~~~~~~~~~~~~
     54       {
        ~~~~~~~
    ...
     60       ["win32"]
        ~~~~~~~~~~~~~~~
     61     ),
        ~~~~~
    node_modules/@electron-forge/shared-types/dist/index.d.ts:109:5
    109     config: any;
            ~~~~~~
    'config' is declared here.
    forge.config.ts:54:7 - error TS2554: Expected 0 arguments, but got 2.
     54       {
              ~
     55         name: "FreedomGPT",
        ~~~~~~~~~~~~~~~~~~~~~~~~~~~
    ...
     59       },
        ~~~~~~~~
     60       ["win32"]
        ~~~~~~~~~~~~~~~
■ Preparing native dependencies
■ Running generateAssets hook

An unhandled rejection has occurred inside Forge:
TSError: ⨯ Unable to compile TypeScript:
forge.config.ts:51:5 - error TS2322: Type 'MakerZIP' is not assignable to type 'ForgeConfigMaker'.
  Property 'config' is missing in type 'MakerZIP' but required in type 'IForgeResolvableMaker'.

51     new MakerZIP({}, ["darwin"]),
       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

  node_modules/@electron-forge/shared-types/dist/index.d.ts:109:5
    109     config: any;
            ~~~~~~
    'config' is declared here.
forge.config.ts:51:18 - error TS2554: Expected 0 arguments, but got 2.

51     new MakerZIP({}, ["darwin"]),
                    ~~~~~~~~~~~~~~
forge.config.ts:52:5 - error TS2322: Type 'MakerDMG' is not assignable to type 'ForgeConfigMaker'.
  Property 'config' is missing in type 'MakerDMG' but required in type 'IForgeResolvableMaker'.

52     new MakerDMG({}, ["darwin"]),
       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

  node_modules/@electron-forge/shared-types/dist/index.d.ts:109:5
    109     config: any;
            ~~~~~~
    'config' is declared here.
forge.config.ts:52:18 - error TS2554: Expected 0 arguments, but got 2.

52     new MakerDMG({}, ["darwin"]),
                    ~~~~~~~~~~~~~~
forge.config.ts:53:5 - error TS2322: Type 'MakerSquirrel' is not assignable to type 'ForgeConfigMaker'.
  Property 'config' is missing in type 'MakerSquirrel' but required in type 'IForgeResolvableMaker'.

 53     new MakerSquirrel(
        ~~~~~~~~~~~~~~~~~~
 54       {
    ~~~~~~~
...
 60       ["win32"]
    ~~~~~~~~~~~~~~~
 61     ),
    ~~~~~

  node_modules/@electron-forge/shared-types/dist/index.d.ts:109:5
    109     config: any;
            ~~~~~~
    'config' is declared here.
forge.config.ts:54:7 - error TS2554: Expected 0 arguments, but got 2.

 54       {
          ~
 55         name: "FreedomGPT",
    ~~~~~~~~~~~~~~~~~~~~~~~~~~~
...
 59       },
    ~~~~~~~~
 60       ["win32"]
    ~~~~~~~~~~~~~~~
at createTSError (E:\web\FreedomGPT\node_modules\.pnpm\[email protected][email protected]\node_modules\ts-node\src\index.ts:859:12)
    at reportTSError (E:\web\FreedomGPT\node_modules\.pnpm\[email protected][email protected]\node_modules\ts-node\src\index.ts:863:19)
    at getOutput (E:\web\FreedomGPT\node_modules\.pnpm\[email protected][email protected]\node_modules\ts-node\src\index.ts:1077:36)
    at Object.compile (E:\web\FreedomGPT\node_modules\.pnpm\[email protected][email protected]\node_modules\ts-node\src\index.ts:1433:41)
    at Module.m._compile (E:\web\FreedomGPT\node_modules\.pnpm\[email protected][email protected]\node_modules\ts-node\src\index.ts:1617:30)
    at Module._extensions..js (node:internal/modules/cjs/loader:1153:10)
    at Object.require.extensions.<computed> [as .ts] (E:\web\FreedomGPT\node_modules\.pnpm\[email protected][email protected]\node_modules\ts-node\src\index.ts:1621:12)
    at Module.load (node:internal/modules/cjs/loader:981:32)
    at Function.Module._load (node:internal/modules/cjs/loader:822:12)
    at Module.require (node:internal/modules/cjs/loader:1005:19)
    at require (node:internal/modules/cjs/helpers:102:18)
    at exports.default (E:\web\FreedomGPT\node_modules\.pnpm\@[email protected]\node_modules\@electron-forge\core\src\util\forge-config.ts:127:22)
    at async Task.task (E:\web\FreedomGPT\node_modules\.pnpm\@[email protected]\node_modules\@electron-forge\core\src\api\start.ts:63:29)
    at async Task.run (E:\web\FreedomGPT\node_modules\.pnpm\[email protected]\node_modules\listr2\dist\index.cjs:978:11)
    at async E:\web\FreedomGPT\node_modules\.pnpm\[email protected]\node_modules\p-map\index.js:57:22
error Command failed with exit code 1.
info Visit https://yarnpkg.com/en/docs/cli/run for documentation about this command.

Process finished with exit code 1

I copy pasted this:

git clone https://github.com/ohmplatform/FreedomGPT.git
cd FreedomGPT
yarn install
yarn start:prod

and got an endless loading.

Thanks for playing with the current models and approaches and making your results available!
There are build instructions for linux, but I do not see a linux binary linked somewhere - could you upload a build with the next release? Or is it not possible (sorry for not checking myself, I think I got stuck because of incompatible node versions or so - know to few about that ecosystem).

Is there any way we can download a larger LLM if we have enough RAM and resources to support it? I currently have 128GB for reference.

I would like to suggest that FreedomGPT should have the ability to edit a prompt like you can do in ChatGPT. Currently, in FreedomGPT, once a prompt is submitted, there is no way to edit it or continue the conversation from that point onwards.

As a user of FreedomGPT, I find it inconvenient that I cannot edit a prompt once it has been submitted. This limits my ability to fine-tune the conversation and achieve the desired outcome. However, I have noticed that ChatGPT has a feature that allows users to edit prompts, and I think it would be great if FreedomGPT could have a similar functionality.

Getting the following error with: yarn start

`yarn run v1.22.19
$ electron-forge start
/home/ubuntu/freedom-gpt/node_modules/listr2/dist/index.cjs:90
this.path = [...task.listr.path ?? [], task.title].join(" > ");
^

SyntaxError: Unexpected token ?
at Module._compile (internal/modules/cjs/loader.js:723:23)
at Object.Module._extensions..js (internal/modules/cjs/loader.js:789:10)
at Module.load (internal/modules/cjs/loader.js:653:32)
at tryModuleLoad (internal/modules/cjs/loader.js:593:12)
at Function.Module._load (internal/modules/cjs/loader.js:585:3)
at Module.require (internal/modules/cjs/loader.js:692:17)
at require (internal/modules/cjs/helpers.js:25:18)
at Object. (/home/ubuntu/freedom-gpt/node_modules/@electron-forge/cli/dist/electron-forge.js:10:18)
at Module._compile (internal/modules/cjs/loader.js:778:30)
at Object.Module._extensions..js (internal/modules/cjs/loader.js:789:10)
error Command failed with exit code 1.
info Visit https://yarnpkg.com/en/docs/cli/run for documentation about this command.`

Please support android devices

WARNING in ./node_modules/ws/lib/buffer-util.js 113:23-44
Module not found: Error: Can't resolve 'bufferutil' in '/home/jet/freedom-gpt/node_modules/ws/lib'
@ ./node_modules/ws/lib/websocket.js 31:21-45
@ ./node_modules/ws/index.js 3:18-44
@ ./node_modules/engine.io/build/server.js 12:13-26
@ ./node_modules/engine.io/build/engine.io.js 5:17-36
@ ./node_modules/socket.io/dist/index.js 36:20-40
@ ./src/index.ts 15:20-40

WARNING in ./node_modules/ws/lib/validation.js 117:24-49
Module not found: Error: Can't resolve 'utf-8-validate' in '/home/jet/freedom-gpt/node_modules/ws/lib'
@ ./node_modules/ws/lib/sender.js 11:30-53
@ ./node_modules/ws/index.js 8:19-42
@ ./node_modules/engine.io/build/server.js 12:13-26
@ ./node_modules/engine.io/build/engine.io.js 5:17-36
@ ./node_modules/socket.io/dist/index.js 36:20-40
@ ./src/index.ts 15:20-40

Hello, it is possible to provide fully offline version or at least URLs to download the data a where to copy them manually? After I run it, it wants to download another 4212 MB of data. I'd like to use this in fully offline environment without internet access and this seems to be not possible in current version.

Following problem:

After I selected the AI Model the program get stuck in the "Loading Model" screen forever. I waited several minutes and nothing happens. In the terminal (I run the Linux / yarn version) appears following message, when I select the model:
Child process exited with code 126 and signal null
Any ideas?

When I try to ask something to the ai with some accents (like "é"), the "é" in the AI reply aren't recognized

Also, when I don't use accents at all, the AI remove every accents on it sentence

But, when I the letters without the accents, the AI is responding with the accents

Hi. Does the application(arm64) work on Intel Mac?