config.py

import os
from pathlib import Path

REPO_DIR = Path(os.path.realpath(os.path.join(os.path.dirname(__file__), '..')))

# Text extraction
url = 'VedCodes/Easy_Share'
extraction_path = REPO_DIR / "VedCodes/Easy_Share.jsonl"

# Text processing
min_length = 100

# HF repo
hf_repo = "VedCodes/Easy_Share"

# Dataset
context_length = 2048
batch_size = 1000
test_size = 0.1
shuffle = True

# Training
model_name = 'bigscience/bloom-3b'
lora_r = 16 # attention heads
lora_alpha = 32 # alpha scaling
lora_dropout = 0.05
lora_bias = "none"
lora_task_type = "CAUSAL_LM" # set this for CLM or Seq2Seq

## Trainer config
per_device_train_batch_size = 1 
gradient_accumulation_steps = 1
warmup_steps = 100 
num_train_epochs=3
weight_decay=0.1
learning_rate = 2e-4 
fp16 = True
logging_steps = 1
overwrite_output_dir = True
evaluation_strategy = "no"
save_strategy = "no"
push_to_hub = False

## Data collator
mlm =False

## Generate
max_new_tokens = 50
temperature = 0.5
do_sample = False