Skip to main content

Hugging Face distilbert-base-uncased not predicting well

I have the following script where I'm trying to fine tune distilbert. It seems to train decently fast, but when I run predictions on the model, then they're all over the place. I'm pretty new to python and ML, so it's been hard debugging to figure out what's happening.

import tensorflow as tf
from datasets import load_dataset
import numpy as np
from transformers import DistilBertTokenizer, TFAutoModelForSequenceClassification, pipeline, create_optimizer, DataCollatorWithPadding


tokenizer = DistilBertTokenizer.from_pretrained("distilbert-base-uncased")

def train():
    def preprocess_function(examples):
        return tokenizer(examples["text"], truncation=True)

    dataset = load_dataset('json', data_files='full-items.json')

    tokenized = dataset.map(preprocess_function, batched=True)

    data_collator = DataCollatorWithPadding(tokenizer=tokenizer, return_tensors="tf")

    batch_size = 16
    num_epochs = 5
    batches_per_epoch = len(tokenized["train"]) // batch_size
    total_train_steps = int(batches_per_epoch * num_epochs)
    optimizer, schedule = create_optimizer(init_lr=2e-5, num_warmup_steps=0, num_train_steps=total_train_steps)

    id2label = {0: "NEGATIVE", 1: "POSITIVE"}
    label2id = {"NEGATIVE": 0, "POSITIVE": 1}

    model = TFAutoModelForSequenceClassification.from_pretrained(
        "distilbert-base-uncased", num_labels=2, id2label=id2label, label2id=label2id
    )

    tf_train_set = model.prepare_tf_dataset(
        tokenized["train"],
        shuffle=True,
        batch_size=16,
        collate_fn=data_collator,
    )

    model.compile(optimizer=optimizer, metrics="accuracy")
    model.fit(x=tf_train_set, epochs=3)
    model.save_pretrained('lease_to_own_model', save_format="tf")

def predict(text):
    model = TFAutoModelForSequenceClassification.from_pretrained(
        'lease_to_own_model'
    )

    pipe = pipeline("text-classification", model=model, tokenizer=tokenizer)

    prediction = pipe(text)

    return prediction

My json file looks like this:

[
  { "text": "tv", "label": 1 },
  { "text": "gun", "label": 0 },
]

training



source https://stackoverflow.com/questions/74868539/hugging-face-distilbert-base-uncased-not-predicting-well

Comments

Popular posts from this blog

Confusion between commands.Bot and discord.Client | Which one should I use?

Whenever you look at YouTube tutorials or code from this website there is a real variation. Some developers use client = discord.Client(intents=intents) while the others use bot = commands.Bot(command_prefix="something", intents=intents) . Now I know slightly about the difference but I get errors from different places from my code when I use either of them and its confusing. Especially since there has a few changes over the years in discord.py it is hard to find the real difference. I tried sticking to discord.Client then I found that there are more features in commands.Bot . Then I found errors when using commands.Bot . An example of this is: When I try to use commands.Bot client = commands.Bot(command_prefix=">",intents=intents) async def load(): for filename in os.listdir("./Cogs"): if filename.endswith(".py"): client.load_extension(f"Cogs.{filename[:-3]}") The above doesnt giveany response from my Cogs ...

How to show number of registered users in Laravel based on usertype?

i'm trying to display data from the database in the admin dashboard i used this: <?php use Illuminate\Support\Facades\DB; $users = DB::table('users')->count(); echo $users; ?> and i have successfully get the correct data from the database but what if i want to display a specific data for example in this user table there is "usertype" that specify if the user is normal user or admin i want to user the same code above but to display a specific usertype i tried this: <?php use Illuminate\Support\Facades\DB; $users = DB::table('users')->count()->WHERE usertype =admin; echo $users; ?> but it didn't work, what am i doing wrong? source https://stackoverflow.com/questions/68199726/how-to-show-number-of-registered-users-in-laravel-based-on-usertype

Why is my reports service not connecting?

I am trying to pull some data from a Postgres database using Node.js and node-postures but I can't figure out why my service isn't connecting. my routes/index.js file: const express = require('express'); const router = express.Router(); const ordersCountController = require('../controllers/ordersCountController'); const ordersController = require('../controllers/ordersController'); const weeklyReportsController = require('../controllers/weeklyReportsController'); router.get('/orders_count', ordersCountController); router.get('/orders', ordersController); router.get('/weekly_reports', weeklyReportsController); module.exports = router; My controllers/weeklyReportsController.js file: const weeklyReportsService = require('../services/weeklyReportsService'); const weeklyReportsController = async (req, res) => { try { const data = await weeklyReportsService; res.json({data}) console...