### Install requirements

Source: https://github.com/m3rg-iitd/matscibert/blob/main/README.md

Execute the installation script to set up the necessary environment dependencies.

```bash
bash install_requirements.sh
```

--------------------------------

### Install Dependencies

Source: https://github.com/m3rg-iitd/matscibert/blob/main/cls/cls_baseline/README.md

Installs the required dependencies using conda and downloads a necessary spaCy model. Ensure you have conda installed and the environment.yaml file is present.

```bash
conda env create --file environment.yaml
conda activate torch1.4
python -m spacy download en_core_web_sm
```

--------------------------------

### Execute Training Scripts via Command Line

Source: https://context7.com/m3rg-iitd/matscibert/llms.txt

Commands for installing requirements and running various downstream tasks including classification, NER, and pre-training.

```bash
# Install requirements
bash install_requirements.sh

# Abstract Classification with MatSciBERT
cd cls
python cls.py \
    --model_name matscibert \
    --model_save_dir ./output \
    --preds_save_dir ./predictions \
    --seeds 0 1 2 \
    --lm_lrs 2e-5 3e-5 5e-5

# Named Entity Recognition on MATSCHOLAR dataset
cd ner
python ner.py \
    --model_name matscibert \
    --model_save_dir ./output \
    --architecture bert-crf \
    --dataset_name matscholar \
    --seeds 0 1 2 \
    --lm_lrs 2e-5 3e-5 5e-5

# NER on SOFC corpus with 5-fold cross-validation
python ner.py \
    --model_name matscibert \
    --model_save_dir ./output \
    --architecture bert-bilstm-crf \
    --dataset_name sofc \
    --fold_num 1 \
    --hidden_dim 300

# Relation Classification
cd relation_classification
python relation_classification.py \
    --model_name matscibert \
    --model_save_dir ./output \
    --seeds 0 1 2 \
    --lm_lrs 2e-5 3e-5 5e-5

# Pre-training from SciBERT
cd pretraining
python matscibert_pre_train.py \
    --train_file normalized_train.txt \
    --val_file normalized_val.txt \
    --model_save_dir ./pretrained_model
```

--------------------------------

### Define Output and Model Directories

Source: https://github.com/m3rg-iitd/matscibert/blob/main/ner/ner_inference.ipynb

Sets up directory paths for caching, output, and model storage. Ensure these directories exist or are created before running.

```python
root_dir = '../'
cache_dir = os.path.join(root_dir, '.cache')
output_dir = os.path.join(root_dir, 'ner/output_matscibert_matscholar')
model_name = os.path.join(root_dir, 'ner/models/matscholar')
to_normalize = True
```

--------------------------------

### Import Libraries for MatSciBERT

Source: https://github.com/m3rg-iitd/matscibert/blob/main/ner/ner_inference.ipynb

Imports necessary libraries for MatSciBERT, including PyTorch, Hugging Face Transformers, NumPy, Pandas, and ChemDataExtractor. Ensure these libraries are installed.

```python
import os
import pickle
import sys
import multiprocessing as mp
sys.path.append('../')

from argparse import ArgumentParser
import numpy as np
import pandas as pd
from tqdm import tqdm
from chemdataextractor.doc import Paragraph

import torch
from torch import nn

import ner_datasets
from models import BERT_CRF
from normalize_text import normalize

from transformers import (
    AutoConfig,
    AutoModelForTokenClassification,
    AutoTokenizer,
    EvalPrediction,
    Trainer,
    TrainingArguments,
    set_seed,
    AdamW,
)

import chemdataextractor as cde
```