CoffeeVampir3

export module Vulkan;
import <stdexcept>;
import Validation;

#define GLFW_INCLUDE_VULKAN
#include <GLFW/glfw3.h>

export namespace Vulkan {
    auto CreateInstance() {
        VkInstance instance;

CoffeeVampir3

export module VulkanInstance;

#define GLFW_INCLUDE_VULKAN
#include <GLFW/glfw3.h>

import VulkanValidation;

export namespace Vulkan {
    auto CreateInstance() {
        VkInstance instance;

CoffeeVampir3

from exllamav2 import(
    ExLlamaV2Tokenizer
)

def encode_message(tokenizer: ExLlamaV2Tokenizer, role: str, message: str) -> list:
    tokens = []
    start_token = tokenizer.single_id("<start_of_turn>")
    end_token = tokenizer.single_id("<end_of_turn>")
    
    tokens.append(start_token)

CoffeeVampir3

use candle_transformers::models::quantized_llama as model;
use candle_transformers::generation::{LogitsProcessor, Sampling};
use candle_core::quantized::{gguf_file};
use candle_core::Tensor;
pub use candle_core::Device;
pub use tokenizers::Tokenizer;

pub struct StreamableModel {
    model: model::ModelWeights,
    tokenizer: tokenizers::Tokenizer,

CoffeeVampir3

CoffeeVampir3

#pragma once
#include <coroutine>
#include <utility>
#include <functional>

struct NecroAiState;

struct NecroAiState
{
	struct promise_type;

CoffeeVampir3

    let cube_mesh_handle: Handle<Mesh> = meshes.add(Rectangle::new(6., 9.));
    let cube_mesh_handle2: Handle<Mesh> = meshes.add(Rectangle::new(6., 9.));

    commands.spawn((
        SpatialBundle {
            transform: Transform::default(),
            ..default()
        },
        ParentMarker,
        // On::<Pointer<Drag>>::listener_component_mut::<Transform>(|drag, transform| {

CoffeeVampir3

use std::io::Write;
use tokenizers::Tokenizer;
use candle_core::quantized::{gguf_file};
use candle_core::Device;
use candle_transformers::generation::{LogitsProcessor, Sampling};
use candle_transformers::models::quantized_llama as model;
use model::ModelWeights;
use burnt_wick::streamable_model::StreamableModel;

fn load_model_and_tokenizer(

CoffeeVampir3

use std::io::Write;
use tokenizers::Tokenizer;
use candle_core::quantized::{gguf_file};
use candle_core::Tensor;
use candle_core::Device;
use candle_transformers::generation::{LogitsProcessor, Sampling};
use candle_transformers::models::quantized_llama as model;
use candle_examples::token_output_stream::TokenOutputStream;
use model::ModelWeights;
use std::collections::HashMap;

CoffeeVampir3

from flask import Flask, render_template
import torch
from flask_socketio import SocketIO, emit
from generation.make_instruct import get_generator_func
from generation.exllama_generator_wrapper import encode_message, encode_system, encode_header, encode_header_prefilled, encode_message_with_eot, encode_completion
from collections import deque
import time
import os,sys

sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))