sri-harini-m/M26CS7.501-Assignment-1-Models

ANLP Assignment 1 - Model Checkpoints (Configs C1 to C5)

This repository contains PyTorch model checkpoints (.pt) for various configurations (C1 through C5) developed as part of Assignment 1 for the Advanced Natural Language Processing (ANLP) course.

Model Details

• Name: Sriharini Margapuri (2026701030)
• Course: M26CS7.501 - Advanced Natural Language Processing
• Assignment: Assignment 1: Transformers from Scratch, Architectural Variants, and Byte Latent Transformers (BLT)

Repository Files

The repository stores multiple configurations, including both final and best-performing checkpoints based on validation loss:

• Config C1: model_C1.pt, model_C1_best.pt
• Config C2: model_C2.pt, model_C2_best.pt
• Config C3: model_C3.pt, model_C3_best.pt
• Config C4: model_C4.pt, model_C4_best.pt
• Config C5: model_C5.pt, model_C5_best.pt

How to Load the Model

You can load any of the checkpoint files directly using PyTorch:

import torch

# Load a specific configuration checkpoint
checkpoint = torch.load("model_C5_best.pt", map_location="cpu")

# Depending on how your training script saved the checkpoint, 
# it may contain model weights, optimizer states, or configuration dictionaries.
print(checkpoint.keys())