notebook.py

# -*- coding: utf-8 -*-
"""Untitled11.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/11JsYh0UnnMH0hbxcRt1bpgBHI-DxoQuH
"""

import json
import os

if not os.path.exists("/tmp/credentials.json"):
    raise AssertionError("Please provide a /tmp/credentials.json")

credentials = json.load(open("/tmp/credentials.json"))

# github_access_token = credentials["github_token"]
os.environ["WANDB_API_KEY"] = credentials["wandb_api_key"]
# hf_auth_token = credentials["hf_auth_token"]

del credentials

# Commented out IPython magic to ensure Python compatibility.
# %%capture
# 
# !pip install transformers peft wandb

import tempfile

import wandb

def download_lora(wandb_run_id, epoch_num):
    run = wandb.Api().run(wandb_run_id)
    [artifact] = [artifact_ for artifact_ in run.logged_artifacts() if artifact_.metadata.get("epoch_num") == epoch_num]
    download_dir = tempfile.mkdtemp()
    artifact.download(root=download_dir)
    return download_dir

from concurrent.futures import ThreadPoolExecutor
from concurrent import futures
import os
import tempfile
from dataclasses import dataclass
from peft import PeftModel
from tqdm import tqdm
from typing import Any, List
import requests


CLOUD_LORA_URL = "https://my-cloud-run-service-m3v6o3jpna-uc.a.run.app"


class CloudLora:
    def __init__(self, lora_dir: str):
        self.lora_dir = lora_dir

    @staticmethod
    def from_lora_dir(lora_dir: str):
        return CloudLora(lora_dir=lora_dir)

    @staticmethod
    def from_peft_model(peft_model: PeftModel):
        tmpdir = tempfile.mkdtemp()
        peft_model.save_pretrained(tmpdir)
        return CloudLora(lora_dir=tmpdir)

    def map_prompts(self, prompts: List[str]):
        response = requests.post(f"{CLOUD_LORA_URL}/upload-files", files=dict(
            adapter_config=open(os.path.join(self.lora_dir, "adapter_config.json"), "rb"),
            adapter_model=open(os.path.join(self.lora_dir, "adapter_model.bin"), "rb"),
        ))
        if response.status_code != 200:
            print("uuid request error", response.status_code, response.text)
        cloud_weights_uuid = response.json()["uuid"]

        def fetch_response(prompt):
            response = requests.post(
                f"{CLOUD_LORA_URL}/generate/{cloud_weights_uuid}",
                headers={"Content-Type": "application/json"},
                json={"prompt": prompt}
            )

            if response.status_code != 200:
                print("generation request error", prompt, response.status_code, response.text)
                return None

            return prompt + response.json()["output"]["choices"][0]["text"]

            # print(prompt, response.status_code, response.text)

        fs = []
        with ThreadPoolExecutor() as executor:
            for prompt in prompts:
                fut = executor.submit(fetch_response, prompt)

                fs.append(fut)

            # results = list(executor.map(fetch_response, prompts))

        results = []
        for fut in futures.as_completed(fs):
            result = fut.result()
            print(result)
            results.append(result)

        return results

# https://wandb.ai/ericyu3-fine-tuning/fine-tuning/runs/gzfqekzc?workspace=user-ericyu3_
run_id = "ericyu3-fine-tuning/fine-tuning/gzfqekzc"

epoch_to_load = 10
lerp_ratio = 1.0

lora_id = str(hash((run_id, epoch_to_load, lerp_ratio)))
lora_dir = download_lora(run_id, epoch_num=epoch_to_load)

CloudLora.from_lora_dir(lora_dir).map_prompts([
    "What is the output of 'print(5 // 2)'?",
    "What is the purpose of the 'pass' keyword in Python?",
    "How can you convert a string to a list of characters?",
    "What is the difference between 'is' and '=='?",
    "How can you reverse a list in Python?",
    "What is the result of '5 * 3'?",
    "How would you remove duplicates from a list?",
    "What is the output of 'print(5 % 2)'?",
    "How would you check the length of a list?",
    "What is the output of 'print(5 ** 2)'?",
    "How do you concatenate two strings?",
    "What is a lambda function?",
    "What is the difference between a tuple and a list?",
    "What is a dictionary in Python?",
    "How do you create a function in Python?",
    "What is the difference between a method and a function?",
    "How do you check the type of a variable?",
    "What is the output of 'print('Hello' * 3)'?",
    "How do you create a class in Python?",
    "What is inheritance in object-oriented programming?",
    "How do you handle exceptions in Python?",
    "What is the difference between a set and a list?",
    "How do you iterate over a dictionary?",
    "What is the use of the 'global' keyword?",
    "How do you write a recursive function?",
    "What is a decorator in Python?",
    "What is the purpose of the '__init__' method?",
    "How do you access the value of a dictionary using a key?",
    "How do you copy a list?",
    "How do you convert a string to a float?",
    "What is the purpose of a docstring?",
    "What is list comprehension?",
    "What is the purpose of the '__name__' variable?",
    "What is slicing in Python?",
    "How do you merge two dictionaries?",
    "How can you sort a list in descending order?",
    "What is the output of 'print('Hello'.upper())'?",
    "How can you convert an integer to a string?",
    "How do you add an element to a set?",
    "What is a module in Python?",
    "What is the use of the 'break' keyword?",
    "How do you find the index of an element in a list?",
    "What is the output of 'print(5 + 2)'?",
    "How do you create a virtual environment in Python?",
    "What is the difference between 'and' and 'or' operators?",
    "How do you remove an element from a list?",
    "What is the purpose of the 'else' clause in a loop?",
    "How do you install a package using pip?",
    "What is the output of 'print('hello'.capitalize())'?",
    "What is the purpose of the 'return' statement?",
    "How do you access elements in a tuple?",
    "What is a generator in Python?",
    "How do you open a file in Python?",
    "What is the use of the 'continue' keyword?",
    "How do you check if a key is present in a dictionary?",
    "What is the output of 'print('hello world'.split())'?",
    "What is the difference between 'print' and 'return'?",
    "How do you concatenate two lists?",
    "What is the purpose of the 'yield' keyword?",
    "How do you create a set in Python?",
    "What is a namespace in Python?",
    "How do you find the maximum value in a list?",
    "What is the output of 'print('Python'.lower())'?",
    "What is the difference between shallow and deep copy?",
    "How do you create a dictionary in Python?",
    "What is the use of the 'assert' keyword?",
    "How do you append an element to a list?",
    "What is the purpose of a constructor?",
    "What is the difference between 'append' and 'extend' methods in a list?",
    "How do you check if a string starts with a certain substring?",
    "How can you generate a random number in Python?",
    "What is the output of 'print('5' + '3')'?",
    "What is a list in Python?",
    "What is polymorphism in object-oriented programming?",
    "How do you remove an element from a set?",
    "What is the purpose of the '__str__' method?",
    "What is the output of 'print('hello'.islower())'?",
    "How do you read a file in Python?",
    "What is an instance variable?",
    "How do you write a list to a file?",
    "What is the use of the 'import' keyword?",
    "What is the output of 'print(5 - 2)'?",
    "How do you create a tuple in Python?",
    "What is the difference between 'remove' and 'pop' methods in a list?",
    "How do you write a for loop?",
    "What is the use of the 'in' keyword?",
    "What is the output of 'print('hello'.replace('h', 'j'))'?",
    "How do you check if a string ends with a certain substring?",
    "What is a method in object-oriented programming?",
    "What is an attribute in object-oriented programming?",
    "How do you find the minimum value in a list?",
    "What is the output of 'print(10 / 2)'?",
    "What is encapsulation in object-oriented programming?",
    "How do you convert a list to a tuple?",
    "What is an iterator in Python?",
    "How do you create a list in Python?",
    "What is the purpose of the '__call__' method?",
    "What is the use of the 'def' keyword?",
    "How do you check if an element is present in a list?",
    "What is the output of 'print('hello world'.title())'?",
    "How do you create an instance of a class?",
    "What is the use of the 'while' loop?",
    "What is the output of 'print(bool(0))'?",
    "How do you write a conditional statement?",
    "What is the difference between a class variable and an instance variable?",
    "How do you read input from the user?",
    "What is the use of the 'not' keyword?",
    "How do you get the keys from a dictionary?",
    "What is the output of 'print(list(range(5)))'?",
    "What is a default argument in a function?",
    "How do you create a string in Python?",
    "What is the difference between 'str' and 'repr'?",
    "What is the use of the 'enumerate' function?",
    "What is the output of 'print('hello'.isalpha())'?",
    "What is the purpose of the 'self' parameter in a class method?",
    "How do you get the values from a dictionary?",
    "What is a comprehension in Python?",
    "What is the output of 'print(' '.join(['hello', 'world']))'?",
    "How do you write a while loop?",
])

}