lucataco

import os
import datetime
from huggingface_hub import hf_hub_download
from safetensors.torch import load_file
import torch

sf_filename = hf_hub_download("bigscience/bloom-560m", filename="model.safetensors")
pt_filename = hf_hub_download("bigscience/bloom-560m", filename="pytorch_model.bin")

lucataco

import os
import datetime
from huggingface_hub import hf_hub_download
from safetensors.torch import load_file
import torch

sf_filename = "./model.safetensors"
pt_filename = hf_hub_download("google/flan-t5-large", filename="pytorch_model.bin")

lucataco

# Install Ubuntu 22.04
sudo apt-get update
sudo apt-get upgrade -y

# Install tools
sudo apt install git htop btop

# Install miniconda
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
chmod +x Miniconda3-latest-Linux-x86_64.sh

lucataco

from transformers import AutoTokenizer, AutoModelForCausalLM
import transformers
import torch
import time

model = "tiiuae/falcon-7b"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",

lucataco

from cog import BasePredictor, Input, Path
import os
import torch
import time
from diffusers import (DDIMScheduler, 
    DiffusionPipeline,
    DPMSolverMultistepScheduler,
    EulerAncestralDiscreteScheduler, 
    EulerDiscreteScheduler,
    HeunDiscreteScheduler, 

lucataco

import io
import time
import json
import base64
import requests
from PIL import Image
# Start SDXL locally:
# docker run -d -p 5000:5000 --gpus=all r8.im/stability-ai/sdxl@sha256:39ed52f2a78e934b3ba6e2a89f5b1c712de7dfea535525255b1aa35c5565e08b

url = "http://localhost:5000/predictions"

lucataco

import io
import time
import json
import base64
import requests

# Start SDXL locally:
# docker run -d -p 5000:5000 --gpus=all r8.im/stability-ai/stable-video-diffusion@sha256:3f0457e4619daac51203dedb472816fd4af51f3149fa7a9e0b5ffcf1b8172438

lucataco

import time
import json
import requests


# Start Llama2 13b locally:
# docker run -d -p 5000:5000 --gpus=all r8.im/meta/llama-2-13b@sha256:078d7a002387bd96d93b0302a4c03b3f15824b63104034bfa943c63a8f208c38


url = "http://localhost:5000/predictions"

lucataco

import os
import time
import torch
from typing import Iterator
from threading import Thread
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer

#Change this to 512, 1024, 2048
MAX_NEW_TOKENS = 512
MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))

lucataco

from diffusers import DiffusionPipeline
import torch
import time

# load both base & refiner
t1 = time.time()
base = DiffusionPipeline.from_pretrained(
    "stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16, variant="fp16", use_safetensors=True
)
base.to("cuda")