# pip install torch peft==0.4.0 bitsandbytes==0.40.2 transformers==4.31.0 trl==0.4.7 accelerate einops
# pip install tqdm scipy

from transformers import pipeline
import torch
from datasets import load_dataset
from datasets import load_from_disk
from peft import LoraConfig, prepare_model_for_kbit_training
from transformers import (
    AutoModelForCausalLM,
    AutoTokenizer,
    TrainingArguments,
)

from trl import SFTTrainer

training_dataset = load_dataset("csv", data_files="formatted_qna_lite.csv", split="train")

print(training_dataset)
print("Is CUDA enabled?",torch.cuda.is_available())
base_model = "microsoft/phi-2"
new_model = "phi-2-qna"

tokenizer = AutoTokenizer.from_pretrained(base_model, use_fast=True)
tokenizer.pad_token=tokenizer.eos_token
tokenizer.padding_side="right"

model = AutoModelForCausalLM.from_pretrained(
    base_model,
    #quantization_config=bnb_config,
    # use_flash_attention_2=True, # Phi does not support yet.
    trust_remote_code=True,
    flash_attn=True,
    flash_rotary=True,
    fused_dense=True,
    revision="refs/pr/23",
)

model.config.use_cache = False
model.config.pretraining_tp = 1

training_arguments = TrainingArguments(
    output_dir="./results",
    num_train_epochs=3,
    per_device_train_batch_size=2,
    gradient_accumulation_steps=32,
    evaluation_strategy="steps",
    eval_steps=2000,
    logging_steps=15,
    optim="adamw_hf",
    learning_rate=2e-4,
    lr_scheduler_type="cosine",
    save_steps=2000,
    warmup_ratio=0.05,
    weight_decay=0.01,
    max_steps=-1
)

peft_config = LoraConfig(
    r=32,
    lora_alpha=64,
    lora_dropout=0.05,
    bias="none",
    task_type="CAUSAL_LM",
    target_modules= ["Wqkv", "fc1", "fc2" ] # ["Wqkv", "out_proj", "fc1", "fc2" ], - 41M params
    # modules_to_save=["embed_tokens","lm_head"]
)

trainer = SFTTrainer(
    model=model,
    train_dataset=training_dataset,
    peft_config= peft_config,
    dataset_text_field="Text",
    max_seq_length=690,
    tokenizer=tokenizer,
    args=training_arguments,
)


trainer.train()

prompt = "How old was Pascal when he lost his mother?"
pipe = pipeline(task="text-generation", model=model, tokenizer=tokenizer, max_length=250)

while prompt != 'EXIT':
    result = pipe(f"[INST] {prompt} [/INST]")
    print(result[0]['generated_text'])
    print("\n")
    prompt = input("Ask the next question .....\n")