perryism

import google.auth
from google.auth.transport.requests import Request
import requests

def get_credentials() -> str:
    credentials, _project_id = google.auth.default(scopes=['https://www.googleapis.com/auth/cloud-platform'])
    request = Request()
    credentials.refresh(request)
    return credentials

perryism

from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig, HfArgumentParser, TrainingArguments, pipeline, logging, TextStreamer
from peft import LoraConfig, PeftModel, prepare_model_for_kbit_training, get_peft_model
import os,torch, platform, warnings
from datasets import load_dataset     
from trl import SFTTrainer

class Lora:
    @classmethod
    def prepare(cls, model_name):
        """

perryism

import unittest
from langchain.prompts import PromptTemplate
from langchain.llms import OpenAI
import logging
import sys
logging.basicConfig(stream=sys.stdout, level=logging.INFO)
logger = logging.getLogger(__name__)

class Agent:
    def __init__(self, role, objective):

perryism

import argparse
import logging
from enum import EnumMeta

logger = logging.getLogger(__name__)

class Blackmore:
    def __init__(self, name, functions):
        self.name = name
        self.functions = dict([[f.__name__, f] for f in functions])

perryism

from google.oauth2.credentials import Credentials
from google.cloud import bigquery
import google.oauth2.credentials

import os

PROJECT = os.environ.get("PROJECT")

# gcloud auth print-access-token
ACCESS_TOKEN = os.environ.get("ACCESS_TOKEN")

perryism

from langchain.chains import SequentialChain
from langchain.prompts import ChatPromptTemplate
from langchain.chains import LLMChain
from langchain.chat_models import ChatOpenAI

llm = ChatOpenAI(temperature=0.9)

class MyChain:
    def __init__(self, llm=None, template_engine=None):
        self.llm = llm

perryism

JAVA_HOME=$(java_home -v 18)
CERT_FOLDER=/Users/plee/.certs
KEYSTORE=$JAVA_HOME/lib/security/cacerts

for cert in $(ls $CERT_FOLDER)
do
  sudo keytool -import -alias $(basename $cert .cer) -keystore $KEYSTORE -file $CERT_FOLDER/$cert
done

#sudo keytool -list -keystore $KEYSTORE | grep -i <cert name>

perryism

inputs = """red pepper pasta
garlic noodles
pizza
biryani
roti curry
flat rice noodles
hakka noodles
bhel puri
sushi
ramen"""

perryism

def handle(body):
    body["foo"] = "bar"
    return body

perryism

Overview

I need to create a vertex ai endpoint with different functions. There are two options. 1. one endpoint per function. 1. one endpoint for all functions. This is an attempt to handle different functions in a single endpoint.

Base image

The base image requires three files.

1. requirements.txt - it will be used to build the base docker image
2. handler.py - it is the fastapi handler which we have direct access to request and response objects.