Відео

Very interesting video. Greetings from Popayan, Colombia.

print(model.predict(np.array([[10.0]])))

Personal anecdote from early 2024: Using OpenAI ChatGPT 4 LLM prompting plus using the ChatGPT Python Code Interpreter module _together_ enabled the AI to solve 1 difficult problem from the ARC challenge fairly quickly. So LLM alone = no general AI, but LLM + code interpreter is much more capable and hence more intelligent . So does combining LLM + code interpreter + prompt rewriting + Self-refine + ... move us even more closer to general AI?.... That's a Q I would have asked in hindsight

great , I am more clear about it , but still I dont know should I move to MLE or I should continue working as SWE ?

Isn't c and c++ faster than python?

i was having troubles with the code till i used this (from keras_preprocessing.text import Tokenizer sentences = [ 'i love my dog', 'i love my cat', 'i hate dog!' ] tokenizer = Tokenizer(num_words=100) tokenizer.fit_on_texts(sentences) word_index = tokenizer.word_index print(word_index) )

What a great learning experience! Thank you Laurence Moroney

안녕하세용

Hindi version sounded pretty good to me. All the examples are pretty amazing

from flask import Flask, request, jsonify from pyngrok import ngrok import pandas as pd import numpy as np import re # إعداد Flask app = Flask(__name__) # تحميل البيانات ومعالجتها كما هو في الكود الأصلي PATH = '/content/mobile_recommendation_system_dataset.csv' data = pd.read_csv(PATH) # هنا نضع بقية كود معالجة البيانات كما هو في المثال السابق def recommend(brand=None, system=None, min_price=None, max_price=None, top_n=10): filtered_data = data if brand: filtered_data = filtered_data[filtered_data['Brand'] == brand] if system: filtered_data = filtered_data[filtered_data['System'] == system] if min_price is not None: filtered_data = filtered_data[filtered_data['price'] >= min_price] if max_price is not None: filtered_data = filtered_data[filtered_data['price'] <= max_price] if filtered_data.empty: return [] recommendations = filtered_data.sort_values(by='ratings', ascending=False).head(top_n) return list(recommendations.name.values) @app.route('/recommend', methods=['GET']) def get_recommendations(): brand = request.args.get('brand') system = request.args.get('system') min_price = request.args.get('min_price', type=int) max_price = request.args.get('max_price', type=int) top_n = request.args.get('top_n', type=int, default=10) recommendations = recommend(brand, system, min_price, max_price, top_n) return jsonify(recommendations) # تشغيل ngrok لإتاحة الخادم عبر الإنترنت url = ngrok.connect(5000) print(f"Public URL: {url}") app.run()

i love you

Great lecture! Really easy to follow. Even being familiar with Lin Alg, this gave me some cool insights on Data Science.

amei 💙

amei 💙

amei !!!! 💙

Good

very clear thank you

Im another one who don't catch why 128 neurons in the middle layer. Great Video, thanks.

1:14 lol they compared TPU runtimes with GPU runtimes

Thanks sir

0:16

sentences = [ 'كم سعر الراجحي', 'ما هي قيمة الراجحي؟', 'هل تعرف سعر أرامكو؟' ] {'سعر': 1, 'كم': 2, 'الراجحي': 3, 'ما': 4, 'هي': 5, 'قيمة': 6, 'الراجحي؟': 7, 'هل': 8, 'تعرف': 9, 'أرامكو؟': 10} its putting الراجحي and الراجحي? as two tokens, is that becuase of arabic?

Every step has been followed but keep getting this error "RuntimeError: Input tensor has type kTfLiteFloat32: it requires specifying NormalizationOpt ions metadata to preprocess input images." Raspberry Pi 4 Bullseye, 32-bit

Karmel are you sitting down while making this video? JW 😊

Obama!

Interesting

the following command fails in JupyterLab: file = files.upload()

This is so clear and easy to understand. Thank you, now I have a better understanding about NLP.

Palm pilots are back?

print(Hashing == Tokenization ) whats the output??

wow just wow

Start with random values, then the optimizer "will generate new parameters to see if it can do better". How? Either this glosses over the core of the solution, or there is no core and the answer is arrived at completely randomly. Not very clever, if so.

this is huge!!

token_list = tokenizer.texts_to_sequences([line])[0], why we put [line] instead of line simply here?

The "adam = Adam(lr=0.01) ValueError: Argument(s) not recognized: {'lr': 0.01}" can be fixed as follows: adam = Adam(learning_rate=0.01)

bless you

its not Win 11 USB plug and play and no driver self installation - the Ai is not shown in the Win 11 Taskmanager - if this comes in 11 it would be nice - also parallel Coral USB Sticks in use would be nice - USB 1,2,3, ... to geht more Ai power 🙂 ... on my Win 11 PC i have VM with a Linux Mint Cinnamon virtual PC - it would be fine if this Windows Ai would be pushed through into the Virtual Mint PC ... like my mouse, Soundblaster, Wifi as a Bridge ... and so on

Debugging

Perfect, Thanks for clarification of the new KerasCV library

hello today i saw this video i dont understand one things what is validation data ? i have to create a model for full face detection only face without any background

i tried to to use it but the pip install tflite-model-maker, downloaded about 50GB and reached google's storage limit

@tensorflow amazing host btw!

I cant find seedbank, could someone out there please send me the link?

I wish one day I could have a chance to work together Mr.Moroney

Why did google drop support for running Tensorflow on GPU using Windows? And how does Keras 3 change that?

Thank you

انا استفدت ايه

Not while your testing and whom you used asv the lang model.

Can you do the model training on a windows pc instead of the RPi, in my case it is RPi 3b, then just download the finished model/product on the RPi after?

What a great tut, thanks dude ....