spanish_medica_llm.py

import argparse
import itertools
import math
import os
from pathlib import Path
from typing import Optional
import subprocess
import sys
from datetime import datetime
from dataclasses import dataclass, field
from typing import Optional

import numpy as np
import torch
from datasets import load_dataset, concatenate_datasets
from transformers import (
    AutoModelForCausalLM,
    AutoTokenizer,
    BitsAndBytesConfig, 
    DataCollatorForLanguageModeling,
    TrainingArguments,
    Trainer
)

from accelerate import FullyShardedDataParallelPlugin, Accelerator
from torch.distributed.fsdp.fully_sharded_data_parallel import FullOptimStateDictConfig, FullStateDictConfig
from peft import LoraConfig, get_peft_model, prepare_model_for_kbit_training
#import wandb
from trl import SFTTrainer

from huggingface_hub import login


CHAT_ML_TEMPLATE_Mistral_7B_Instruct = """
{% if messages[0]['role'] == 'system' %}
    {% set loop_messages = messages[1:] %}
    {% set system_message = messages[0]['content'].strip() + '\n\n' %}
{% else %}
    {% set loop_messages = messages %}
    {% set system_message = '' %}
{% endif %}

{{ bos_token }}
{% for message in loop_messages %}
    {% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}
        {{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}
    {% endif %}

    {% if loop.index0 == 0 %}
        {% set content = system_message + message['content'] %}
    {% else %}
        {% set content = message['content'] %}
    {% endif %}

    {% if message['role'] == 'user' %}
        {{ '[INST] ' + content.strip() + ' [/INST]' }}
    {% elif message['role'] == 'assistant' %}
        {{ ' '  + content.strip() + ' ' + eos_token }}
    {% endif %}
{% endfor %}
"""




def parse_args():
    parser = argparse.ArgumentParser(description="Simple example of a training script.")
    parser.add_argument(
        "--pretrained_model_name_or_path",
        type=str,
        default=None,
        #required=True,
        help="Path to pretrained model or model identifier from huggingface.co/models.",
    )
    parser.add_argument(
        "--tokenizer_name",
        type=str,
        default=None,
        help="Pretrained tokenizer name or path if not the same as model_name",
    )
    parser.add_argument(
        "--instance_data_dir",
        type=str,
        default=None,
        #required=True,
        help="A folder containing the training data of instance images.",
    )
    parser.add_argument(
        "--class_data_dir",
        type=str,
        default=None,
        required=False,
        help="A folder containing the training data of class images.",
    )
    parser.add_argument(
        "--instance_prompt",
        type=str,
        default=None,
        help="The prompt with identifier specifying the instance",
    )
    parser.add_argument(
        "--class_prompt",
        type=str,
        default="",
        help="The prompt to specify images in the same class as provided instance images.",
    )
    parser.add_argument(
        "--with_prior_preservation",
        default=False,
        action="store_true",
        help="Flag to add prior preservation loss.",
    )
    parser.add_argument("--prior_loss_weight", type=float, default=1.0, help="The weight of prior preservation loss.")
    parser.add_argument(
        "--num_class_images",
        type=int,
        default=100,
        help=(
            "Minimal class images for prior preservation loss. If not have enough images, additional images will be"
            " sampled with class_prompt."
        ),
    )
    parser.add_argument(
        "--output_dir",
        type=str,
        default="",
        help="The output directory where the model predictions and checkpoints will be written.",
    )
    parser.add_argument("--seed", type=int, default=None, help="A seed for reproducible training.")
    parser.add_argument(
        "--resolution",
        type=int,
        default=512,
        help=(
            "The resolution for input images, all the images in the train/validation dataset will be resized to this"
            " resolution"
        ),
    )
    parser.add_argument(
        "--center_crop", action="store_true", help="Whether to center crop images before resizing to resolution"
    )
    parser.add_argument("--train_text_encoder", action="store_true", help="Whether to train the text encoder")
    parser.add_argument(
        "--train_batch_size", type=int, default=4, help="Batch size (per device) for the training dataloader."
    )
    parser.add_argument(
        "--sample_batch_size", type=int, default=4, help="Batch size (per device) for sampling images."
    )
    parser.add_argument("--num_train_epochs", type=int, default=1)
    parser.add_argument(
        "--max_train_steps",
        type=int,
        default=None,
        help="Total number of training steps to perform.  If provided, overrides num_train_epochs.",
    )
    parser.add_argument(
        "--gradient_accumulation_steps",
        type=int,
        default=1,
        help="Number of updates steps to accumulate before performing a backward/update pass.",
    )
    parser.add_argument(
        "--gradient_checkpointing",
        action="store_true",
        help="Whether or not to use gradient checkpointing to save memory at the expense of slower backward pass.",
    )
    parser.add_argument(
        "--learning_rate",
        type=float,
        default=5e-6,
        help="Initial learning rate (after the potential warmup period) to use.",
    )
    parser.add_argument(
        "--scale_lr",
        action="store_true",
        default=False,
        help="Scale the learning rate by the number of GPUs, gradient accumulation steps, and batch size.",
    )
    parser.add_argument(
        "--lr_scheduler",
        type=str,
        default="constant",
        help=(
            'The scheduler type to use. Choose between ["linear", "cosine", "cosine_with_restarts", "polynomial",'
            ' "constant", "constant_with_warmup"]'
        ),
    )
    parser.add_argument(
        "--lr_warmup_steps", type=int, default=500, help="Number of steps for the warmup in the lr scheduler."
    )
    parser.add_argument(
        "--use_8bit_adam", action="store_true", help="Whether or not to use 8-bit Adam from bitsandbytes."
    )
    parser.add_argument("--adam_beta1", type=float, default=0.9, help="The beta1 parameter for the Adam optimizer.")
    parser.add_argument("--adam_beta2", type=float, default=0.999, help="The beta2 parameter for the Adam optimizer.")
    parser.add_argument("--adam_weight_decay", type=float, default=1e-2, help="Weight decay to use.")
    parser.add_argument("--adam_epsilon", type=float, default=1e-08, help="Epsilon value for the Adam optimizer")
    parser.add_argument("--max_grad_norm", default=1.0, type=float, help="Max gradient norm.")
    parser.add_argument("--push_to_hub", action="store_true", help="Whether or not to push the model to the Hub.")
    parser.add_argument("--hub_token", type=str, default=None, help="The token to use to push to the Model Hub.")
    parser.add_argument(
        "--hub_model_id",
        type=str,
        default=None,
        help="The name of the repository to keep in sync with the local `output_dir`.",
    )
    parser.add_argument(
        "--logging_dir",
        type=str,
        default="logs",
        help=(
            "[TensorBoard](https://www.tensorflow.org/tensorboard) log directory. Will default to"
            " *output_dir/runs/**CURRENT_DATETIME_HOSTNAME***."
        ),
    )
    parser.add_argument(
        "--mixed_precision",
        type=str,
        default="no",
        choices=["no", "fp16", "bf16"],
        help=(
            "Whether to use mixed precision. Choose"
            "between fp16 and bf16 (bfloat16). Bf16 requires PyTorch >= 1.10."
            "and an Nvidia Ampere GPU."
        ),
    )

    parser.add_argument(
        "--save_n_steps",
        type=int,
        default=1,
        help=("Save the model every n global_steps"),
    )
    
    
    parser.add_argument(
        "--save_starting_step",
        type=int,
        default=1,
        help=("The step from which it starts saving intermediary checkpoints"),
    )
    
    parser.add_argument(
        "--stop_text_encoder_training",
        type=int,
        default=1000000,
        help=("The step at which the text_encoder is no longer trained"),
    )


    parser.add_argument(
        "--image_captions_filename",
        action="store_true",
        help="Get captions from filename",
    )    
    
    
    parser.add_argument(
        "--dump_only_text_encoder",
        action="store_true",
        default=False,        
        help="Dump only text encoder",
    )

    parser.add_argument(
        "--train_only_unet",
        action="store_true",
        default=False,        
        help="Train only the unet",
    )
    
    parser.add_argument(
        "--Session_dir",
        type=str,
        default="",     
        help="Current session directory",
    )    

    
    

    parser.add_argument("--local_rank", type=int, default=-1, help="For distributed training: local_rank")

    args = parser.parse_args()
    env_local_rank = int(os.environ.get("LOCAL_RANK", -1))
    if env_local_rank != -1 and env_local_rank != args.local_rank:
        args.local_rank = env_local_rank

    #if args.instance_data_dir is None:
    #    raise ValueError("You must specify a train data directory.")

    #if args.with_prior_preservation:
    #    if args.class_data_dir is None:
    #        raise ValueError("You must specify a data directory for class images.")
    #    if args.class_prompt is None:
    #        raise ValueError("You must specify prompt for class images.")

    return args


def run_training(args_imported):
    args_default = parse_args()
    #args = merge_args(args_default, args_imported)
    return(args)



TOKEN_NAME = "DeepESP/gpt2-spanish-medium"
TOKEN_MISTRAL_NAME = "mistralai/Mistral-7B-Instruct-v0.1"
SPANISH_MEDICA_LLM_DATASET = "somosnlp/spanish_medica_llm"

TOPIC_TYPE_DIAGNOSTIC = 'medical_diagnostic'
TOPIC_TYPE_TRATAMIENT = 'medical_topic'
FILTER_CRITERIA = [TOPIC_TYPE_DIAGNOSTIC, TOPIC_TYPE_TRATAMIENT]
CONTEXT_LENGTH = 256 #Max of tokens

MISTRAL_BASE_MODEL_ID = "BioMistral/BioMistral-7B"

MICRO_BATCH_SIZE = 16  #32 For other GPU BIGGER THAN T4
BATCH_SIZE = 64 #128 For other GPU BIGGER THAN T4
GRADIENT_ACCUMULATION_STEPS = BATCH_SIZE // MICRO_BATCH_SIZE

PROJECT_NAME = "spanish-medica-llm"
BASE_MODEL_NAME = "biomistral"
run_name = BASE_MODEL_NAME + "-" + PROJECT_NAME
output_dir = "./" + run_name

HUB_MODEL_ID = 'somosnlp/spanish_medica_llm'
MAX_TRAINING_STEPS = int(1500/2)
MAX_TRAINING_STEPS = 2

def loadSpanishTokenizer():
    """
    
    """
    #Load first the mistral used tokenizer
    tokenizerMistrall = AutoTokenizer.from_pretrained(TOKEN_MISTRAL_NAME)

    #Load second an spanish specialized tokenizer
    tokenizer = AutoTokenizer.from_pretrained(
        TOKEN_NAME,
        eos_token = tokenizerMistrall.special_tokens_map['eos_token'],
        bos_token = tokenizerMistrall.special_tokens_map['bos_token'],
        unk_token = tokenizerMistrall.special_tokens_map['unk_token']
    )
    tokenizer.chat_template = CHAT_ML_TEMPLATE_Mistral_7B_Instruct

    return tokenizer

def tokenize(element, tokenizer):
    outputs = tokenizer(
        element["raw_text"],
        truncation = True,
        max_length = CONTEXT_LENGTH,
        return_overflowing_tokens = True,
        return_length = True,
    )
    input_batch = []
    for length, input_ids in zip(outputs["length"], outputs["input_ids"]):
        if length == CONTEXT_LENGTH:
            input_batch.append(input_ids)
    return {"input_ids": input_batch}

def splitDatasetInTestValid(dataset):
    """
    """
    
    if dataset == None or dataset['train'] == None:
        return dataset
    elif dataset['test'] == None:
        return None
    else:
        test_eval = dataset['test'].train_test_split(test_size=0.001) 
        eval_dataset = test_eval['train']
        test_dataset = test_eval['test']
        
        return (dataset['train'], eval_dataset, test_dataset)

def loadSpanishDataset():
    spanishMedicaLllmDataset = load_dataset(SPANISH_MEDICA_LLM_DATASET, split="train")
    spanishMedicaLllmDataset = spanishMedicaLllmDataset.filter(lambda example: example["topic_type"] not in  FILTER_CRITERIA)
    spanishMedicaLllmDataset = spanishMedicaLllmDataset.train_test_split(0.2, seed=203984)
    return spanishMedicaLllmDataset

    ##See Jupyter Notebook for change CONTEXT_LENGTH size

def accelerateConfigModel():
    """
      Only with GPU support 
        RuntimeError: There are currently no available devices found, must be one of 'XPU', 'CUDA', or 'NPU'.
    """
    fsdp_plugin = FullyShardedDataParallelPlugin(
        state_dict_config=FullStateDictConfig(offload_to_cpu=True, rank0_only=False),
        optim_state_dict_config=FullOptimStateDictConfig(offload_to_cpu=True, rank0_only=False),
    )

    return Accelerator(fsdp_plugin=fsdp_plugin)

def getTokenizedDataset(dataset, tokenizer):
    if  dataset == None or tokenizer == None:
        return dataset

    return  dataset.map(
        lambda element : tokenize(element, tokenizer),
        batched = True,
        remove_columns = dataset["train"].column_names
    )

def loadBaseModel(base_model_id):

    if base_model_id in [ "", None]:
      return None
    else:
        bnb_config = BitsAndBytesConfig(
            load_in_4bit = True,
            bnb_4bit_quant_type = "nf4",
            bnb_4bit_use_double_quant = True,
            bnb_4bit_compute_dtype = torch.bfloat16
        )

        model = AutoModelForCausalLM.from_pretrained(
                base_model_id,
                quantization_config = bnb_config
            )
        
        model.gradient_checkpointing_enable()
        model = prepare_model_for_kbit_training(model)

        return model
    
def print_trainable_parameters(model):
    """
    Prints the number of trainable parameters in the model.
    """
    trainable_params = 0
    all_param = 0
    for _, param in model.named_parameters():
        all_param += param.numel()
        if param.requires_grad:
            trainable_params += param.numel()
    print(
        f"trainable params: {trainable_params} || all params: {all_param} || trainable%: {100 * trainable_params / all_param}"
    )
    
def modelLoraConfigBioMistral(model):  
    """
      r is the rank of the low-rank matrix used in the adapters, which thus controls 
      the number of parameters trained. A higher rank will allow for more expressivity, but there is a 
      compute tradeoff.
      alpha is the scaling factor for the learned weights. The weight matrix is scaled by 
      alpha/r, and thus a higher value for alpha assigns more weight to the LoRA activations.
      The values used in the QLoRA paper werer=64 and lora_alpha=16,
       and these are said to generalize well, but we will user=8 and lora_alpha=16 so that we have more emphasis on the new fine-tuned data while also reducing computational complexity.
    """
    if model == None:
        return model
    else:
        config = LoraConfig(
            r=8,
            lora_alpha=16,
            target_modules=[
                "q_proj",
                "k_proj",
                "v_proj",
                "o_proj",
                "gate_proj",
                "up_proj",
                "down_proj",
                "lm_head",
            ],
            bias="none",
            lora_dropout=0.05,  # Conventional
            task_type="CAUSAL_LM",
        )

        model = get_peft_model(model, config)
        print_trainable_parameters(model)

        accelerator = accelerateConfigModel()
       # Apply the accelerator. You can comment this out to remove the accelerator.
        model = accelerator.prepare_model(model)
        return (model)
    

# A note on training. You can set the max_steps to be high initially, and examine at what step your
# model's performance starts to degrade. There is where you'll find a sweet spot for how many steps 
# to perform. For example, say you start with 1000 steps, and find that at around 500 steps 
# the model starts overfitting - the validation loss goes up (bad) while the training 
# loss goes down significantly, meaning the model is learning the training set really well, 
# but is unable to generalize to new datapoints. Therefore, 500 steps would be your sweet spot, 
# so you would use the checkpoint-500 model repo in your output dir (biomistral-medqa-finetune) 
# as your final model in step 6 below.
    

def configAndRunTraining(basemodel, dataset, eval_dataset, tokenizer):
    if basemodel is None or dataset is None or tokenizer is None:
        return None
    else:
        tokenizer.pad_token = tokenizer.eos_token
        data_collator_pretrain = DataCollatorForLanguageModeling(tokenizer, mlm = False)
        
        training_args = TrainingArguments(
                output_dir=output_dir,
                push_to_hub = True,
                hub_private_repo = False,
                hub_model_id = HUB_MODEL_ID,
                warmup_steps = 5,
                per_device_train_batch_size = MICRO_BATCH_SIZE,
                per_device_eval_batch_size=1,
                #gradient_checkpointing=True,
                gradient_accumulation_steps = GRADIENT_ACCUMULATION_STEPS,
                max_steps = MAX_TRAINING_STEPS,
                learning_rate = 2.5e-5, # Want about 10x smaller than the Mistral learning rate
                logging_steps = 50,
                optim="paged_adamw_8bit",
                logging_dir="./logs",        # Directory for storing logs
                save_strategy = "steps",       # Save the model checkpoint every logging step
                save_steps = 50,                # Save checkpoints every 50 steps
                evaluation_strategy = "steps", # Evaluate the model every logging step
                eval_steps = 50,               # Evaluate and save checkpoints every 50 steps
                do_eval = True,                # Perform evaluation at the end of training
                report_to = None,           # Comment this out if you don't want to use weights & baises
                run_name=f"{run_name}-{datetime.now().strftime('%Y-%m-%d-%H-%M')}" ,         # Name of the W&B run (optional)
                fp16=True,  #Set for GPU T4 for more powerful GPU as G-100 or another change to false and bf16 parameter
                bf16=False
            )
        
        trainer = Trainer(
                     model= basemodel,
                     train_dataset = dataset,
                     eval_dataset = eval_dataset,
                     args = training_args,
                     data_collator = data_collator_pretrain
                )
        
        basemodel.config.use_cache = False  # silence the warnings. Please re-enable for inference!
        trainer.train()


        trainer.push_to_hub()
    



def run_training_process():
    #Loggin to Huggin Face
    login(token = os.environ.get('HG_FACE_TOKEN'))
    os.environ['WANDB_DISABLED'] = 'true'
    tokenizer = loadSpanishTokenizer()
    medicalSpanishDataset =  loadSpanishDataset()
    train_dataset, eval_dataset, test_dataset = splitDatasetInTestValid(
        getTokenizedDataset( medicalSpanishDataset, tokenizer)
       )

    base_model =  loadBaseModel(MISTRAL_BASE_MODEL_ID)
    base_model = modelLoraConfigBioMistral(base_model)

    configAndRunTraining(base_model,train_dataset, eval_dataset, tokenizer)