https://bbycroft.net/llm

اين يك وبسايت ازمايشي و درك فهم
بسيار جالبه برام حتما يه سر به اين سايت خفن بزنيد

A linear model can learn non-linear/discrete patterns

اموزش يادگيري تقويتي

https://andri27-ts.github.io/Reinforcement-Learning/

Can we combine PyTorch and Scikit-Learn to create the ultimate ML library? Yes, Skorch does just that.

It allows you to leverage PyTorch's deep learning and Sklearn's ease of use. No need for complex code. You can use a simple net.fit(X, y) for training.

https://github.com/skorch-dev/skorch/

Network Neural cheat sheet

PyTorch vs. TensorFlow: the Battle of Machine Learning Frameworks

Deep learning frameworks are essential tools that simplify the development of artificial neural networks (ANNs), and their evolution has been rapid. Among these, TensorFlow and PyTorch stand out, each holding its own in various machine learning realms. But how does one decide the ideal tool for specific projects? This comprehensive guide aims to elucidate their strengths and weaknesses.

Origins and Overview:

TensorFlow: Born from the brains at Google's Brain team, TensorFlow transitioned from being a proprietary tool to an open-source marvel. As an end-to-end platform, it offers everything from basic arithmetic operations to neural network deployment. Its adaptability is evident in its compatibility with platforms such as CPUs, GPUs, TPUs, and mobile devices. Notably, industry giants like Google, Uber, and Microsoft have integrated TensorFlow into their operations.

PyTorch: Introduced in 2016, PyTorch struck a chord by marrying user-friendliness with high performance. Its Pythonic design approach and dynamic computation graphs have made it a top choice in the research community. Developed primarily in C++, its efficiency is notable, and its adoption in platforms like Tesla Autopilot and Uber’s Pyro further attests to its capabilities.

Popularity Trends in Context: PyTorch vs TensorFlow

The shifting dynamics in the popularity between PyTorch and TensorFlow over a period can be linked with significant events and milestones in the world of these frameworks:

TensorFlow’s Initial Popularity: In the early phase of our timeline, TensorFlow had a distinct edge in popularity. This can be credited to its strong backing by Google and its wide-ranging tools that catered to both beginners and professionals.

PyTorch’s Rise: Moving forward, PyTorch started gaining momentum. Its approach, which many found to be more flexible for research and experimentation, played a role in attracting attention. Additionally, as more resources and support became available for PyTorch, its user base grew.

Recent Landscape: Towards the end, both PyTorch and TensorFlow have settled into their roles in the machine learning world. TensorFlow remains a solid pick for those looking at large-scale deployments and industry solutions. PyTorch, with its emphasis on flexibility, remains higher popularity for many specially in the research domain.

Deep Dive: Static vs. Dynamic Computational Graphs

A foundational distinction between TensorFlow and PyTorch is their approach to computational graphs. TensorFlow employs static computational graphs, while PyTorch advocates for dynamic ones.

Making an Informed Choice

Choosing between TensorFlow and PyTorch isn’t about selecting the “best” framework but about finding the one that aligns best with your needs. Both frameworks offer unique advantages and have made significant strides in addressing their initial limitations. By evaluating your project’s requirements, your familiarity with Python, the need for scalability, deployment preferences, and the kind of community support you’re seeking, you can make a choice that ensures efficiency and productivity. As the world of deep learning continues to evolve, so will these frameworks, and staying updated will empower you to make informed decisions time and again.

Paper Alert: Enhancements in Multi-LoRA Composition for Text-to-Image Models 📄

Paper: https://arxiv.org/pdf/2402.16843.pdf

Code: https://github.com/maszhongming/Multi-LoRA-Composition

پیاده سازی مدل های یادگیری ماشین تنها با چند خط کد!!

به کمک کتابخونه lazypredict خیلی راحت میتونید با چند خط کد، داده هاتون رو با تمام مدل های کلاسیک یادگیری ماشین آموزش بدید.

لینک سایت اصلی که کامل این لایبرری رو توضیح داده:
https://lazypredict.readthedocs.io/en/latest/

لینک گیت هاب:
https://github.com/shankarpandala/lazypredict?tab=readme-ov-file

مقاله روی زبان فارسی فاین
PersianLLaMA: Towards Building First Persian Large Language Model

https://arxiv.org/abs/2312.15713

https://github.com/hooshvare/parsgpt

from transformers import pipeline
import hazm


normalizer = hazm.Normalizer(persian_numbers=False)

def normalize_input(text):
    text = normalizer.normalize(text)
    return text

def sents_as_output(text, num_sents=1):
    sents = hazm.sent_tokenize(text)
    if num_sents > 0:
        return " ".join(sents[:num_sents])
    return " ".join(sents[0])


generator = pipeline('text-generation', "HooshvareLab/gpt2-fa")
text = "در یک اتفاق شگفت انگیز، پژوهشگران"
text = normalize_input(text)

outputs = generator(text)
for output in outputs:
    generated = output["generated_text"]
    print(sents_as_output(generated))

خروجى:

در یک اتفاق شگفت انگیز، پژوهشگران قصد دارند با استفاده از داده‌های حاصل از چندین تلسکوپ، عکس‌هایی با وضوح مختلف از سیاره‌ی مشتری و زحل تهیه کنند

دوتا روش حل دارد

Linear SVM

RBF SVM

Linear SVM vs RBF

In Support Vector Machines (SVM), the choice between linear kernel and a Radial Basis Function (RBF) kernel depends on the data and the problem at hand. Here’s a brief comparison:
• Linear Kernel:
◦ It is a linear model.
◦ Works well when there are a lot of features, and the data is linearly separable.
◦ Faster to compute and may perform better when the number of features is very large1.
• RBF Kernel:
◦ It can handle non-linear data.
◦ Uses normal curves around the data points to define a decision boundary.
◦ Typically requires more computational resources but can capture more complex relationships2.

linear_svc = svm.SVC(kernel='linear')
linear_svc.kernel
rbf_svc = svm.SVC(kernel='rbf')
rbf_svc.kernel

Output:

'linear'
'rbf'

Understanding the black-box calculations on Neural Networks