Sadikshya Dhakal

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay60
Wrapping up my 60 Days of Learning by building an EduRAG pipeline with PDF processing, embeddings, FAISS indexing, and semantic search. Grateful for the journey and excited for what's next!
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay60
Today I explored Granger Causality and revised the RAG pipeline, strengthening my understanding of time series prediction and LLM-based retrieval systems. @lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay59
Today I learned ARMA and ARIMA for time series forecasting. I explored AR, MA, differencing, ACF/PACF, and the ADF stationarity test, then built an ARIMA model to forecast future sales using Python.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay58
Today I explored SVM decision boundaries and maximum margin, learning how Support Vector Machines choose the optimal hyperplane for better classification. I also visualized decision boundaries using Python.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay57
Today I learned the basics of SVM, including hyperplanes, support vectors, and how SVM finds the best decision boundary for binary classification. I also visualized the decision boundary using Python.
@lftechnology

#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay56
Today I learned about Sentence Transformers and sentence embeddings, generated vector representations for text using all-MiniLM-L6-v2, and measured semantic similarity between sentences using cosine similarity.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay55
Today I explored advanced NLP concepts by learning the difference between Stemming and Lemmatization, performing Part-of-Speech Tagging, and implementing Named Entity Recognition using NLTK.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay54
Today I explored the basics of Natural Language Processing (NLP) using NLTK, learned text preprocessing, performed word stemming with PorterStemmer, and downloaded WordNet for future NLP tasks.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay53
Today I fine-tuned VGG16 using transfer learning, evaluated the model with a confusion matrix, achieved 96.73% validation accuracy, and saved the trained model for future inference.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay52
Today I explored VGG19 architecture and learned how stacked convolutional layers, max pooling, and transfer learning help build powerful image classification models.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay51
Today I evaluated a ResNet18 transfer learning model using a confusion matrix, classification report, and prediction visualization, achieving around 95% validation accuracy.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay50
Today I started Transfer Learning with ResNet18 in PyTorch. I learned how to prepare image datasets, apply data augmentation, create DataLoaders, and use pretrained models for efficient image classification.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay49
Today I learned about Pooling (Max & Average) and Batch Normalization in CNNs. I explored how pooling reduces feature size while BatchNorm normalizes features for faster and more stable model training.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay48
Today I learned how Stride and Padding work in CNNs. I explored how they affect feature map size and visualized the convolution process using numerical examples.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay47
Today I explored CNN feature maps and learned how different filters detect horizontal, vertical, and diagonal patterns through numerical examples.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay46
Today I learned how Convolution works in CNN through a numerical example. I also explored kernels, feature maps, and CNN layers used for image processing.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay45
Today I built a CNN for digit classification, trained it, and tried the numerical implementation step by step. Achieved 98% test accuracy.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay44
Today I learned about the CNN architecture, including convolution, ReLU, pooling, and fully connected layers. I also explored the MNIST dataset for handwritten digit classification.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay43
Today I learned about the Perceptron, its types, and how NVIDIA GPUs accelerate deep learning in Google Colab. I also explored GPU setup and monitoring using nvidia-smi.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog

[Sadikshya Dhakal @SadikshyaDkl]

#LSPPDay42
Today I learned how Backpropagation works in a Neural Network, built a simple neural network in PyTorch, and trained it on the MNIST dataset.
@lftechnology
#60DaysOfLearning2026 #LearningWithLeapfrog