import os os.environ["WANDB_MODE"] = "offline" # os.environ["WANDB_DISABLED"] = "true" import json import math import random import shutil import sys import threading import time import traceback from datetime import datetime from pathlib import Path import gradio as gr import torch import transformers from .custom_scheduler import FPSchedulerTrainer from .matplotgraph import create_graph from .train_utils import get_available_loras_local, precise_cut, sliding_block_cut from datasets import Dataset, load_dataset from peft import ( LoraConfig, get_peft_model, prepare_model_for_kbit_training, set_peft_model_state_dict ) from peft.utils.other import \ TRANSFORMERS_MODELS_TO_LORA_TARGET_MODULES_MAPPING as model_to_lora_modules from transformers.models.auto.modeling_auto import ( MODEL_FOR_CAUSAL_LM_MAPPING_NAMES ) from modules import shared, utils from modules.ui import create_refresh_button from modules.evaluate import ( calculate_perplexity, generate_markdown_table, save_past_evaluations ) from modules.logging_colors import logger from modules.models import reload_model from modules.utils import natural_keys params = { "display_name": "Training PRO", "is_tab": True } non_serialized_params = { "debug_slicer": False, "Lora_sortedByTime": False, "stop_at_loss": 0, "save_steps_under_loss": 0.0, "save_checkpoint_now": False, "training_loop": False, "current_stability": 0, } MODEL_CLASSES = {v[1]: v[0] for v in MODEL_FOR_CAUSAL_LM_MAPPING_NAMES.items()} PARAMETERS = ["lora_name", "always_override", "save_steps", "micro_batch_size", "batch_size", "epochs", "learning_rate", "lr_scheduler_type", "lora_rank", "lora_alpha", "lora_dropout", "cutoff_len", "dataset", "eval_dataset", "format", "eval_steps", "raw_text_file", "higher_rank_limit", "warmup_steps", "optimizer", "hard_cut_string", "train_only_after", "stop_at_loss", "add_eos_token", "min_chars", "report_to", "precize_slicing_overlap", "add_eos_token_type", "save_steps_under_loss", "add_bos_token", "training_projection","sliding_window","warmup_ratio","grad_accumulation"] WANT_INTERRUPT = False train_log = {} train_template = {} train_log_graph = [] train_choices = ["all","q-k-v-o","q-k-v","k-v-down","q-v"] def ui(): with gr.Tab('Train LoRA', elem_id='lora-train-tab'): tmp = gr.State('') with gr.Row(): with gr.Column(): # YY.MM.DD gr.Markdown("`Ver: 23.09.22` This is enhanced version of QLora Training. [Maintained by FP](https://github.com/FartyPants/Training_PRO/tree/main)") with gr.Row(): with gr.Column(scale=5): with gr.Row(): copy_from = gr.Dropdown(label='Copy parameters from', value='None', choices=get_available_loras_local(non_serialized_params['Lora_sortedByTime']), elem_classes=['slim-dropdown']) create_refresh_button(copy_from, lambda: None, lambda: {'choices': get_available_loras_local(non_serialized_params['Lora_sortedByTime'])}, 'refresh-button') with gr.Column(): sort_byTime = gr.Checkbox(label='Sort list by Date', value=False, info='Sorts Loras by date created.', elem_classes=['no-background']) with gr.Row(): with gr.Column(scale=5): lora_name = gr.Textbox(label='Name', info='The name of your new LoRA file') with gr.Column(): always_override = gr.Checkbox(label='Override Existing Files', value=False, info='If the name is the same, checking will replace the existing file, and unchecking will load and continue from it (the rank must be the same).', elem_classes=['no-background']) with gr.Row(): with gr.Column(): lora_rank = gr.Slider(label='LoRA Rank', value=32, minimum=0, maximum=1024, step=4, info='Also called dimension count. Higher values = larger file, more content control. Smaller values = smaller file, less control. Use 4 or 8 for style, 128 or 256 to teach, 1024+ for fine-detail on big data. More VRAM is needed for higher ranks.') lora_alpha = gr.Slider(label='LoRA Alpha', value=64, minimum=0, maximum=2048, step=4, info='This divided by the rank becomes the scaling of the LoRA. Higher means stronger. A good standard value is twice your Rank.') batch_size = gr.Slider(visible= False, label='Batch Size', value=0, minimum=0, maximum=1024, step=4, info='Now Replaced with Gradient accumulation. Keeping it for sake of old saved data') micro_batch_size = gr.Slider(label='True Batch Size', value=4, minimum=1, maximum=128, step=1, info='Specifies how many text blocks per step will be trained. The higher value, the better the concept of training will be, but it requires more GPU memory and it reduces speed.') grad_accumulation = gr.Slider(label='Gradient Accumulation Steps', value=1, minimum=1, maximum=256, step=1, info="Virtually multiplies the Batch Size by averaging the learning over more than one step. Evens out loss fluctuations but also increases number of total steps.") cutoff_len = gr.Slider(label='Cutoff Length', minimum=0, maximum=2048, value=256, step=32, info='Cutoff length for text input. Essentially, how long of a line of text to feed in at a time. Higher values require drastically more VRAM.') with gr.Column(): stop_at_loss = gr.Slider(label='Stop at loss (Can be changed during training)', minimum=0.0, maximum=3.0, step=0.1, value=0.00, info='The process will automatically stop once the desired loss value is reached.') gr.Markdown(" ") epochs = gr.Number(label='Epochs', value=3, info='Number of times every entry in the dataset should be fed into training. So 1 means feed each item in once, 5 means feed it in five times, etc.') learning_rate = gr.Textbox(label='Learning Rate', value='3e-4', info='In scientific notation. 3e-4 is a good starting base point. 1e-2 is extremely high, 1e-6 is extremely low.') lr_scheduler_type = gr.Dropdown(label='LR Scheduler', value='linear', choices=['linear', 'constant', 'constant_with_warmup', 'cosine', 'cosine_with_restarts', 'polynomial', 'inverse_sqrt', 'FP_low_epoch_annealing', 'FP_half_time_annealing'], info='Learning rate scheduler - defines how the learning rate changes over time. Custom schedulers: `FP_low_epoch_annealing` constant for 1 epoch then cosine anneal. `FP_half_time_annealing` constant for half time then cosine anneal', elem_classes=['slim-dropdown']) with gr.Accordion(label='Checkpoints', open=True): with gr.Row(): with gr.Column(): save_steps = gr.Number(label='Save every n steps', value=0, info='A checkpoint will be saved every n steps. (0 = OFF)') with gr.Column(): save_steps_under_loss = gr.Slider(label='Save at 10% Loss change', value=1.8, minimum=0.0, maximum=3.0, step=0.1, info="Saves checkpoints at (or bellow) this loss and then each time loss falls by at least 10% This works independently from 'Save every n steps'") with gr.Row(): save_chackpoint_now = gr.Button('Queue Checkpoint Now') with gr.Accordion(label='Advanced Options', open=True): with gr.Row(): with gr.Column(): warmup_steps = gr.Number(label='Warmup Steps', value=100, info='Number of max steps used for a linear warmup. Value different than 0 has precedent over Warmup Ratio. The actual number of steps will be the closest multiple of graddient accumulation') warmup_ratio = gr.Slider(label='Warmup Ratio', minimum=0.0, maximum=0.2, step=0.025, value=0.0, info='Ratio of total training steps that will be used for a linear warmup. It applies only if Warmup Step is 0.') training_projection = gr.Radio(value = train_choices[4], label='LLaMA Target Projections', info='Change the targets (LORA is typically q-v)', choices=train_choices) lora_dropout = gr.Slider(label='LoRA Dropout', minimum=0.0, maximum=1.0, step=0.025, value=0.05, info='Percentage probability for dropout of LoRA layers. This can help reduce overfitting. Most users should leave at default.') optimizer = gr.Dropdown(label='Optimizer', value='adamw_torch', choices=['adamw_hf', 'adamw_torch', 'adamw_torch_fused', 'adamw_torch_xla', 'adamw_apex_fused', 'adafactor', 'adamw_bnb_8bit', 'adamw_anyprecision', 'sgd', 'adagrad'], info='Different optimizer implementation options, for advanced users. Effects of different options are not well documented yet.', elem_classes=['slim-dropdown']) with gr.Column(): train_only_after = gr.Textbox(label='Train Only After', value='', info='Only consider text *after* this string in any given chunk for training. For Alpaca datasets, use "### Response:" to only train the response and ignore the input.') add_bos_token = gr.Checkbox(label='Add BOS token', value=True, info="Adds BOS token for each dataset item") add_eos_token = gr.Checkbox(label='Add EOS token', value=False, info="Adds EOS token for each dataset item") add_eos_token_type = gr.Dropdown(label='EOS placement (raw text)', choices=['Every Block', 'Hard Cut Blocks Only'], value='Every Block', info='', allow_custom_value = False) higher_rank_limit = gr.Checkbox(label='Enable higher ranks', value=False, info='If checked, changes Rank/Alpha slider above to go much higher. This will not work without a datacenter-class GPU.') report_to = gr.Radio(label="Save detailed logs with", value="None", choices=["None", "wandb", "tensorboard"], interactive=True) with gr.Column(): with gr.Tab(label='Formatted Dataset'): with gr.Row(): with gr.Column(): with gr.Row(): dataset = gr.Dropdown(choices=utils.get_datasets('training/datasets', 'json'), value='None', label='Dataset', info='The dataset file to use for training.', elem_classes=['slim-dropdown']) create_refresh_button(dataset, lambda: None, lambda: {'choices': utils.get_datasets('training/datasets', 'json')}, 'refresh-button') with gr.Row(): eval_dataset = gr.Dropdown(choices=utils.get_datasets('training/datasets', 'json'), value='None', label='Evaluation Dataset', info='The (optional) dataset file used to evaluate the model after training.', elem_classes=['slim-dropdown']) create_refresh_button(eval_dataset, lambda: None, lambda: {'choices': utils.get_datasets('training/datasets', 'json')}, 'refresh-button') with gr.Column(): with gr.Row(): format = gr.Dropdown(choices=utils.get_datasets('training/formats', 'json'), value='None', label='Data Format', info='The format file used to decide how to format the dataset input.', elem_classes=['slim-dropdown']) create_refresh_button(format, lambda: None, lambda: {'choices': utils.get_datasets('training/formats', 'json')}, 'refresh-button') with gr.Row(): eval_steps = gr.Number(label='Evaluate every n steps', value=100, info='If an evaluation dataset is given, test it every time this many steps pass.') with gr.Tab(label="Raw text file"): with gr.Row(): raw_text_file = gr.Dropdown(choices=utils.get_datasets('training/datasets', 'txt'), value='None', label='Text file', info='The raw text file to use for training.', elem_classes=['slim-dropdown']) create_refresh_button(raw_text_file, lambda: None, lambda: {'choices': utils.get_datasets('training/datasets', 'txt')}, 'refresh-button') with gr.Row(): with gr.Column(): precize_slicing_overlap = gr.Checkbox(label='Add Overlapping blocks', value = True) sliding_window = gr.Checkbox(label='DEMENTOR Long-form Learning by FP (Highly Experimental, use low epochs)', value = False, info='Deep Memorization Enforcement Through Overlapping and Repetition. (I named it, so shush). Special process for learning long-form text using low amount of epochs.') #debug_slicer = gr.Checkbox(label='Dump sentencelist.json to logs', value = non_serialized_params['debug_slicer'], info='Debug Slicer') with gr.Column(): hard_cut_string = gr.Textbox(label='Hard Cut String', value='\\n\\n\\n', info='String that indicates a cut between logical blocks of text (ex. Ideas or Chapters). Helps prevent unwanted overlap between unrelated ideas.') min_chars = gr.Number(label='Ignore small blocks', value=0, info='Ignore Text blocks that have less or equal characters than this number.') with gr.Row(): with gr.Column(): check_dataset_btn = gr.Button('Load and Check Dataset and suggest data entries') check_dataset_txt = gr.Textbox(label='Dataset info', value='') with gr.Row(): start_button = gr.Button("Start LoRA Training", variant='primary') stop_button = gr.Button("Interrupt") output = gr.Markdown(value="Ready") with gr.Tab('Perplexity evaluation', elem_id='evaluate-tab'): with gr.Row(): with gr.Column(): models = gr.Dropdown(utils.get_available_models(), label='Models', multiselect=True) evaluate_text_file = gr.Dropdown(choices=['wikitext', 'ptb', 'ptb_new'] + utils.get_datasets('training/datasets', 'txt')[1:], value='wikitext', label='Input dataset', info='The raw text file on which the model will be evaluated. The first options are automatically downloaded: wikitext, ptb, and ptb_new. The next options are your local text files under training/datasets.') with gr.Row(): with gr.Column(): stride_length = gr.Slider(label='Stride', minimum=1, maximum=2048, value=512, step=1, info='Used to make the evaluation faster at the cost of accuracy. 1 = slowest but most accurate. 512 is a common value.') with gr.Column(): max_length = gr.Slider(label='max_length', minimum=0, maximum=8096, value=0, step=1, info='The context for each evaluation. If set to 0, the maximum context length for the model will be used.') with gr.Row(): start_current_evaluation = gr.Button("Evaluate loaded model") start_evaluation = gr.Button("Evaluate selected models") stop_evaluation = gr.Button("Interrupt") with gr.Column(): evaluation_log = gr.Markdown(value='') evaluation_table = gr.Dataframe(value=generate_markdown_table(), interactive=True) with gr.Row(): save_comments = gr.Button('Save comments', elem_classes="small-button") refresh_table = gr.Button('Refresh the table', elem_classes="small-button") # Training events all_params = [lora_name, always_override, save_steps, micro_batch_size, batch_size, epochs, learning_rate, lr_scheduler_type, lora_rank, lora_alpha, lora_dropout, cutoff_len, dataset, eval_dataset, format, eval_steps, raw_text_file, higher_rank_limit, warmup_steps, optimizer, hard_cut_string, train_only_after, stop_at_loss, add_eos_token, min_chars, report_to, precize_slicing_overlap, add_eos_token_type, save_steps_under_loss, add_bos_token, training_projection,sliding_window,warmup_ratio,grad_accumulation] def fix_old_version(batch_size_val,micro_batch_size_val, grad_accumulation_val): if batch_size_val>0: gradient_acc = batch_size_val // micro_batch_size_val print(f"Using Old version of Batch Size ({batch_size_val}) to set Gradient Accumulation: {gradient_acc}") return gradient_acc return grad_accumulation_val copy_from.change(do_copy_params, [copy_from] + all_params, all_params).then(fix_old_version,[batch_size,micro_batch_size, grad_accumulation],grad_accumulation) start_button.click(do_train, all_params, output) stop_button.click(do_interrupt, None, None, queue=False) higher_rank_limit.change(change_rank_limit, [higher_rank_limit], [lora_rank, lora_alpha]) def trigger_stop_at_loss(stop_at_loss_value): non_serialized_params.update({"stop_at_loss": stop_at_loss_value}) if non_serialized_params['training_loop']: print(f"Queue: [Stop at loss Change] to {stop_at_loss_value}") stop_at_loss.change(trigger_stop_at_loss, stop_at_loss, None) def trigger_save_checkpoint(): non_serialized_params.update({"save_checkpoint_now": True}) if non_serialized_params['training_loop']: print("Queue: [Save checkpoint] Checkpoint will be saved after the current step is finished.") else: print("Use during the training to save the checkpoint at any time.") save_chackpoint_now.click(trigger_save_checkpoint, None, None) dataset_calc_params = [save_steps,micro_batch_size, epochs, cutoff_len, dataset, format, raw_text_file, warmup_steps, hard_cut_string, min_chars, precize_slicing_overlap,sliding_window,warmup_ratio,grad_accumulation] def check_dataset(save_steps:int, micro_batch_size: int, epochs: int, cutoff_len: int, dataset:str, format:str, raw_text_file:str, warmup_steps:int, hard_cut_string:str, min_chars:int, precize_slicing_overlap:bool,sliding_window:bool,warmup_ratio:float,grad_accumulation:int): result = "Specify JSON dastaset or raw text file" total_blocks = 0 if shared.tokenizer is None: yield "Tokenizer is not available. Please Load some Model first." return if raw_text_file not in ['None', '']: logger.info("Loading raw text file dataset...") fullpath = clean_path('training/datasets', f'{raw_text_file}') fullpath = Path(fullpath) if fullpath.is_dir(): logger.info('Training path directory {}'.format(raw_text_file)) raw_text = "" file_paths = sorted(fullpath.glob('*.txt'), key=lambda path: natural_keys(path.name)) for file_path in file_paths: if file_path.is_file(): with file_path.open('r', encoding='utf-8') as file: raw_text += file.read().replace('\r', '') logger.info(f"Loaded training file: {file_path.name}") else: with open(clean_path('training/datasets', f'{raw_text_file}.txt'), 'r', encoding='utf-8') as file: raw_text = file.read().replace('\r', '') if min_chars<0: min_chars = 0 # == New more precise slicing on sentence boundary == if sliding_window: text_chunks = sliding_block_cut(raw_text, min_chars, False, cutoff_len, hard_cut_string,non_serialized_params['debug_slicer']) else: text_chunks = precise_cut(raw_text, precize_slicing_overlap, min_chars, False, cutoff_len, hard_cut_string,non_serialized_params['debug_slicer']) total_blocks = len(text_chunks) result = f"Raw Text: ({raw_text_file}.txt) has {total_blocks} blocks (with cutoff length = {cutoff_len})" del text_chunks else: if dataset in ['None', '']: yield "Select dataset or Raw text." return if format in ['None', '']: yield "Select format choice for dataset." return with open(clean_path('training/formats', f'{format}.json'), 'r', encoding='utf-8-sig') as formatFile: format_data: dict[str, str] = json.load(formatFile) def generate_prompt(data_point: dict[str, str]): for options, data in format_data.items(): if set(options.split(',')) == set(x[0] for x in data_point.items() if (type(x[1]) is str and len(x[1].strip()) > 0)): for key, val in data_point.items(): if type(val) is str: data = data.replace(f'%{key}%', val) return data raise RuntimeError(f'Data-point "{data_point}" has no keyset match within format "{list(format_data.keys())}"') def tokenize_dummy(prompt): input_ids = shared.tokenizer.encode(prompt, truncation=True, max_length=cutoff_len) labels = [1] * len(input_ids) input_ids = torch.tensor(input_ids) return { "input_ids": input_ids, "labels": labels, "attention_mask": input_ids.ne(shared.tokenizer.pad_token_id), } def generate_and_tokenize_prompt(data_point): prompt = generate_prompt(data_point) return tokenize_dummy(prompt) logger.info("Loading JSON datasets...") data = load_dataset("json", data_files=clean_path('training/datasets', f'{dataset}.json')) train_data = data['train'].map(generate_and_tokenize_prompt, new_fingerprint='%030x' % random.randrange(16**30)) total_blocks = train_data.num_rows result = f"Dataset: ({dataset}.json) has {total_blocks} blocks (with cutoff length = {cutoff_len})" if total_blocks>0: number_ofSteps = int(math.ceil(total_blocks / micro_batch_size) * epochs) num_stepsPer_epoch = int(math.ceil(number_ofSteps/epochs)) min_warm = math.ceil(100 / grad_accumulation) warmup_steps_suggest = min(int(min_warm*grad_accumulation), int(math.ceil(number_ofSteps * 0.1))) warmup_steps_suggest = min(warmup_steps_suggest,num_stepsPer_epoch) save_each_n_min = int(math.ceil(number_ofSteps/10)) save_each_n_max = int(math.ceil(number_ofSteps/5)) gradient_accumulation_max = int(total_blocks)//micro_batch_size result += f"\n[Batch Size: {micro_batch_size}, Epochs: {epochs}, Gradient Accumulation: {grad_accumulation}]\n" result += f"Total number of steps: {number_ofSteps}\n" result += f"Steps per each Epoch: {num_stepsPer_epoch}\n" result += f"Warmup steps suggestion: {warmup_steps_suggest} (Current: {int(warmup_steps)})\n" result += f"Checkpoint suggestion: Save every {save_each_n_min} - {save_each_n_max} steps (Current: {int(save_steps)})" if gradient_accumulation_max < grad_accumulation: result += f"\n\nWARNING: Gradient Accumulation {grad_accumulation} is too high: It should be below {gradient_accumulation_max}" yield result return check_dataset_btn.click(check_dataset, dataset_calc_params ,check_dataset_txt) # Evaluation events. For some reason, the interrupt event # doesn't work with the .then() syntax, so I write them one # by one in this ugly but functional way. ev = start_evaluation.click(calculate_perplexity, [models, evaluate_text_file, stride_length, max_length], evaluation_log, show_progress=False) start_evaluation.click(generate_markdown_table, None, evaluation_table, show_progress=False) start_current_evaluation.click(lambda: ['current model'], None, tmp) ev_cur = start_current_evaluation.click(calculate_perplexity, [tmp, evaluate_text_file, stride_length, max_length], evaluation_log, show_progress=False) start_current_evaluation.click(generate_markdown_table, None, evaluation_table, show_progress=False) stop_evaluation.click(None, None, None, cancels=[ev, ev_cur], queue=False) refresh_table.click(generate_markdown_table, None, evaluation_table, show_progress=True) save_comments.click( save_past_evaluations, evaluation_table, None).then( lambda: "Comments saved.", None, evaluation_log, show_progress=False) def reload_lora(): return gr.Dropdown.update(choices=get_available_loras_local(non_serialized_params['Lora_sortedByTime'])) # nonserialized items sort_byTime.change(lambda x: non_serialized_params.update({"Lora_sortedByTime": x}), sort_byTime, None).then(reload_lora,None,copy_from) #debug_slicer.change(lambda x: non_serialized_params.update({"debug_slicer": x}), debug_slicer, None) def do_interrupt(): global WANT_INTERRUPT WANT_INTERRUPT = True def do_copy_params(lora_name: str, *args): f_name = f"{shared.args.lora_dir}/{clean_path(None, lora_name)}/training_parameters.json" if Path(f_name).is_file(): with open(f_name, 'r', encoding='utf-8') as format_file: params: dict[str, str] = json.load(format_file) else: params = {} result = list() for i in range(0, len(PARAMETERS)): key = PARAMETERS[i] if key in params: result.append(params[key]) else: result.append(args[i]) return result def change_rank_limit(use_higher_ranks: bool): mult = 2 if use_higher_ranks else 1 return {"maximum": 1024 * mult, "__type__": "update"}, {"maximum": 2048 * mult, "__type__": "update"} def clean_path(base_path: str, path: str): """Strips unusual symbols and forcibly builds a path as relative to the intended directory.""" path = path.replace('\\', '/').replace('..', '_') if base_path is None: return path return f'{Path(base_path).absolute()}/{path}' def backup_adapter(input_folder): # Get the creation date of the file adapter_model.bin try: adapter_file = Path(f"{input_folder}/adapter_model.bin") if adapter_file.is_file(): logger.info("Backing up existing LoRA adapter...") creation_date = datetime.fromtimestamp(adapter_file.stat().st_ctime) creation_date_str = creation_date.strftime("Backup-%Y-%m-%d") # Create the new subfolder subfolder_path = Path(f"{input_folder}/{creation_date_str}") subfolder_path.mkdir(parents=True, exist_ok=True) # Check if the file already exists in the subfolder backup_adapter_file = Path(f"{input_folder}/{creation_date_str}/adapter_model.bin") if backup_adapter_file.is_file(): print(" - Backup already exists. Skipping backup process.") return # Copy existing files to the new subfolder existing_files = Path(input_folder).iterdir() for file in existing_files: if file.is_file(): shutil.copy2(file, subfolder_path) except Exception as e: print("An error occurred in backup_adapter:", str(e)) def calc_trainable_parameters(model): trainable_params = 0 all_param = 0 for _, param in model.named_parameters(): num_params = param.numel() # if using DS Zero 3 and the weights are initialized empty if num_params == 0 and hasattr(param, "ds_numel"): num_params = param.ds_numel all_param += num_params if param.requires_grad: trainable_params += num_params return trainable_params, all_param def do_train(lora_name: str, always_override: bool, save_steps: int, micro_batch_size: int, batch_size: int, epochs: int, learning_rate: str, lr_scheduler_type: str, lora_rank: int, lora_alpha: int, lora_dropout: float, cutoff_len: int, dataset: str, eval_dataset: str, format: str, eval_steps: int, raw_text_file: str, higher_rank_limit: bool, warmup_steps: int, optimizer: str, hard_cut_string: str, train_only_after: str, stop_at_loss: float, add_eos_token: bool, min_chars: int, report_to: str, precize_slicing_overlap: bool, add_eos_token_type: str, save_steps_under_loss: float, add_bos_token: bool, training_projection: str,sliding_window:bool,warmup_ratio:float, grad_accumulation: int): if shared.args.monkey_patch: from alpaca_lora_4bit.monkeypatch.peft_tuners_lora_monkey_patch import ( replace_peft_model_with_int4_lora_model ) replace_peft_model_with_int4_lora_model() global WANT_INTERRUPT WANT_INTERRUPT = False # == Input validation / processing == yield "Preparing the input..." lora_file_path = clean_path(None, lora_name) if lora_file_path.strip() == '': yield "Missing or invalid LoRA file name input." return lora_file_path = f"{Path(shared.args.lora_dir)}/{lora_file_path}" actual_lr = float(learning_rate) model_type = type(shared.model).__name__ if model_type in MODEL_CLASSES: model_id = MODEL_CLASSES[model_type] else: model_id = "llama" if model_type == "PeftModelForCausalLM": if len(shared.lora_names) > 0: yield "You are trying to train a LoRA while you already have another LoRA loaded. This will work, but may have unexpected effects. *(Will continue anyway in 5 seconds, press `Interrupt` to stop.)*" logger.warning("Training LoRA over top of another LoRA. May have unexpected effects.") else: yield "Model ID not matched due to LoRA loading. Consider reloading base model. *(Will continue anyway in 5 seconds, press `Interrupt` to stop.)*" logger.warning("Model ID not matched due to LoRA loading. Consider reloading base model.") else: yield "LoRA training has only currently been validated for LLaMA, OPT, GPT-J, and GPT-NeoX models. Unexpected errors may follow. *(Will continue anyway in 5 seconds, press `Interrupt` to stop.)*" logger.warning(f"LoRA training has only currently been validated for LLaMA, OPT, GPT-J, and GPT-NeoX models. (Found model type: {model_type})") time.sleep(5) if shared.args.loader == 'GPTQ-for-LLaMa' and not shared.args.monkey_patch: yield "LoRA training with GPTQ-for-LLaMa requires loading with `--monkey-patch`" return if cutoff_len <= 0 or micro_batch_size <= 0 or actual_lr <= 0 or lora_rank <= 0 or lora_alpha <= 0: yield "Cannot input zeroes." return #in new version we dumped this in favor of grad_accumulation #set it to zero fo new save batch_size = 0 gradient_accumulation_steps = grad_accumulation #batch_size // micro_batch_size shared.tokenizer.pad_token_id = 0 shared.tokenizer.padding_side = "left" def encode(text, prepend_bos_token): result = shared.tokenizer.encode(text, truncation=True, max_length=cutoff_len) # Check if the first two tokens are BOS if len(result) >= 2 and result[:2] == [shared.tokenizer.bos_token_id, shared.tokenizer.bos_token_id]: result = result[1:] if not prepend_bos_token and result[0] == shared.tokenizer.bos_token_id: result = result[1:] return result def tokenize(prompt, append_eos_token=False, prepend_bos_token = False): if train_only_after == '' or train_only_after not in prompt: input_ids = encode(prompt, prepend_bos_token) if append_eos_token and input_ids[-1] != shared.tokenizer.eos_token_id and len(input_ids) < cutoff_len: input_ids.append(shared.tokenizer.eos_token_id) input_ids = [shared.tokenizer.pad_token_id] * (cutoff_len - len(input_ids)) + input_ids labels = [1] * len(input_ids) else: ind = prompt.index(train_only_after) + len(train_only_after) before_tokens = encode(prompt[:ind], prepend_bos_token) after_tokens = encode(prompt[ind:], False) if append_eos_token and after_tokens[-1] != shared.tokenizer.eos_token_id: after_tokens.append(shared.tokenizer.eos_token_id) full_length = len(after_tokens) + len(before_tokens) if full_length > cutoff_len: after_tokens = after_tokens[:cutoff_len - len(before_tokens)] else: before_tokens = [shared.tokenizer.pad_token_id] * (cutoff_len - full_length) + before_tokens input_ids = before_tokens + after_tokens labels = [-100] * len(before_tokens) + [1] * len(after_tokens) input_ids = torch.tensor(input_ids) return { "input_ids": input_ids, "labels": labels, "attention_mask": input_ids.ne(shared.tokenizer.pad_token_id), } train_template.clear() print(f"*** LoRA: {lora_name} ***") non_serialized_params.update({"stop_at_loss": stop_at_loss}) non_serialized_params.update({"save_steps_under_loss": save_steps_under_loss+0.01}) non_serialized_params.update({"save_checkpoint_now": False}) non_serialized_params.update({"training_loop": False}) non_serialized_params.update({"current_stability": 0}) # END OF FPHAM SENTENCE SPLIT functions =================== # == Prep the dataset, format, etc == if raw_text_file not in ['None', '']: train_template["template_type"] = "raw_text" logger.info("Loading raw text file dataset...") fullpath = clean_path('training/datasets', f'{raw_text_file}') fullpath = Path(fullpath) if fullpath.is_dir(): logger.info('Training path directory {}'.format(raw_text_file)) raw_text = "" file_paths = sorted(fullpath.glob('*.txt'), key=lambda path: natural_keys(path.name)) for file_path in file_paths: if file_path.is_file(): with file_path.open('r', encoding='utf-8') as file: raw_text += file.read().replace('\r', '') logger.info(f"Loaded training file: {file_path.name}") else: with open(clean_path('training/datasets', f'{raw_text_file}.txt'), 'r', encoding='utf-8') as file: raw_text = file.read().replace('\r', '') # FPHAM PRECISE SLICING if min_chars<0: min_chars = 0 add_EOS_to_all = add_eos_token and add_eos_token_type == 'Every Block' add_EOS_to_HC = add_eos_token and add_eos_token_type != 'Every Block' #print (f"add_eos_token {add_eos_token}, add_EOS_to_all {add_EOS_to_all}, add_EOS_to_HC {add_EOS_to_HC}") # == New more precise slicing on sentence boundary == if sliding_window: text_chunks = sliding_block_cut(raw_text, min_chars, add_EOS_to_HC, cutoff_len, hard_cut_string,non_serialized_params['debug_slicer']) else: text_chunks = precise_cut(raw_text, precize_slicing_overlap, min_chars, add_EOS_to_HC, cutoff_len, hard_cut_string,non_serialized_params['debug_slicer']) train_data = Dataset.from_list([tokenize(x, add_EOS_to_all, add_bos_token) for x in text_chunks]) if add_EOS_to_all: print(f"Added EOS to {len(text_chunks)} blocks") print(f"All Data Blocks: {len(text_chunks)}") del text_chunks eval_data = None else: if dataset in ['None', '']: yield "Missing dataset choice input, cannot continue." return if format in ['None', '']: yield "Missing format choice input, cannot continue." return train_template["template_type"] = "dataset" with open(clean_path('training/formats', f'{format}.json'), 'r', encoding='utf-8-sig') as formatFile: format_data: dict[str, str] = json.load(formatFile) # == store training prompt == for _, value in format_data.items(): prompt_key = f"template_{len(train_template)}" train_template[prompt_key] = value def generate_prompt(data_point: dict[str, str]): for options, data in format_data.items(): if set(options.split(',')) == set(x[0] for x in data_point.items() if (type(x[1]) is str and len(x[1].strip()) > 0)): for key, val in data_point.items(): if type(val) is str: data = data.replace(f'%{key}%', val) return data raise RuntimeError(f'Data-point "{data_point}" has no keyset match within format "{list(format_data.keys())}"') def generate_and_tokenize_prompt(data_point): prompt = generate_prompt(data_point) return tokenize(prompt, add_eos_token, add_bos_token) logger.info("Loading JSON datasets...") data = load_dataset("json", data_files=clean_path('training/datasets', f'{dataset}.json')) train_data = data['train'].map(generate_and_tokenize_prompt, new_fingerprint='%030x' % random.randrange(16**30)) print(f"BOS: {add_bos_token} EOS: {add_eos_token}") print(f"Data Blocks: {train_data.num_rows}") if eval_dataset == 'None': eval_data = None else: eval_data = load_dataset("json", data_files=clean_path('training/datasets', f'{eval_dataset}.json')) eval_data = eval_data['train'].map(generate_and_tokenize_prompt, new_fingerprint='%030x' % random.randrange(16**30)) # == We MUST reload model if it went through any previous training, even failed one == if shared.model_dirty_from_training: selected_model = shared.model_name if selected_model: print("\033[1;31;1m(Model has been modified by previous training, it needs to be reloaded...)\033[0;37;0m") try: yield f"Reloading {selected_model}..." reload_model() if shared.model is not None: print("Model reloaded OK, continue with training.") else: return f"Failed to load {selected_model}." except: exc = traceback.format_exc() logger.error('Failed to reload the model.') print(exc) return exc.replace('\n', '\n\n') # == Start prepping the model itself == if not hasattr(shared.model, 'lm_head') or hasattr(shared.model.lm_head, 'weight'): logger.info("Getting model ready...") prepare_model_for_kbit_training(shared.model) # base model is now frozen and should not be reused for any other LoRA training than this one shared.model_dirty_from_training = True if training_projection==train_choices[0]: model_to_lora_modules["llama"] = ["gate_proj","down_proj","up_proj","q_proj","k_proj","v_proj","o_proj"] elif training_projection==train_choices[1]: model_to_lora_modules["llama"] = ["q_proj","k_proj", "v_proj", "o_proj"] elif training_projection==train_choices[2]: model_to_lora_modules["llama"] = ["q_proj","k_proj", "v_proj"] elif training_projection==train_choices[3]: model_to_lora_modules["llama"] = ["k_proj", "v_proj", "down_proj"] else: model_to_lora_modules["llama"] = ["q_proj", "v_proj"] logger.info("Preparing for training...") config = LoraConfig( r=lora_rank, lora_alpha=lora_alpha, target_modules=model_to_lora_modules[model_id], lora_dropout=lora_dropout, bias="none", task_type="CAUSAL_LM" ) # == Backup the existing adapter == if not always_override: backup_adapter(lora_file_path) # == get model trainable params model_trainable_params, model_all_params = calc_trainable_parameters(shared.model) try: logger.info("Creating LoRA model...") lora_model = get_peft_model(shared.model, config) if not always_override and Path(f"{lora_file_path}/adapter_model.bin").is_file(): logger.info("Loading existing LoRA data...") state_dict_peft = torch.load(f"{lora_file_path}/adapter_model.bin") set_peft_model_state_dict(lora_model, state_dict_peft) except: yield traceback.format_exc().replace('\n', '\n\n') return if shared.args.monkey_patch: from alpaca_lora_4bit.autograd_4bit import Autograd4bitQuantLinear from alpaca_lora_4bit.models import Linear4bitLt for _, m in lora_model.named_modules(): if isinstance(m, Autograd4bitQuantLinear) or isinstance(m, Linear4bitLt): if m.is_v1_model: m.zeros = m.zeros.half() m.scales = m.scales.half() class Tracked(): def __init__(self): self.current_steps = 0 self.max_steps = 0 self.did_save = False tracked = Tracked() actual_save_steps = math.ceil(save_steps / gradient_accumulation_steps) class Callbacks(transformers.TrainerCallback): def on_step_begin(self, args: transformers.TrainingArguments, state: transformers.TrainerState, control: transformers.TrainerControl, **kwargs): tracked.current_steps = state.global_step * gradient_accumulation_steps tracked.max_steps = state.max_steps * gradient_accumulation_steps if WANT_INTERRUPT: control.should_epoch_stop = True control.should_training_stop = True else: current_loss = float(train_log.get('loss', 0.0)) current_epoch = float(train_log.get('epoch', 0.0)) force_save = False folder_save = f"checkpoint-{tracked.current_steps}" if non_serialized_params['save_checkpoint_now']: force_save = True non_serialized_params.update({"save_checkpoint_now": False}) print(f"\033[1;31;1mSave Checkpoint manually trigerred.\033[0;37;0m") folder_save = f"checkpoint-{tracked.current_steps}-user" patience = 3 # Set the number of consecutive steps for tracking stability if gradient_accumulation_steps==1: patience = 5 min_steps = 10 if current_loss < non_serialized_params['save_steps_under_loss'] and current_loss > 0 and state.global_step > min_steps: current_stability = non_serialized_params['current_stability'] current_stability += 1 non_serialized_params.update({"current_stability": current_stability}) if current_stability >= patience: current_stability = 0 non_serialized_params.update({"current_stability": current_stability}) current_loss_dec = round(current_loss, 2) loss_str = f"{current_loss_dec:.2f}" loss_str = loss_str.replace('.', '_') new_save = (current_loss_dec-0.1) + 0.01 non_serialized_params.update({"save_steps_under_loss": new_save}) folder_save = f"checkpoint-{tracked.current_steps}-loss-{loss_str}" force_save = True else: # Reset stability if the loss goes above the threshold non_serialized_params.update({"current_stability": 0}) if state.global_step > 0 and actual_save_steps > 0 and state.global_step % actual_save_steps == 0: folder_save = f"checkpoint-{tracked.current_steps}" force_save = True if force_save: lora_model.save_pretrained(f"{lora_file_path}/{folder_save}/") print(f"\033[1;30;40mStep: {tracked.current_steps:6} \033[0;37;0m Saved: [{folder_save}]") # Save log with open(f"{lora_file_path}/{folder_save}/training_log.json", 'w', encoding='utf-8') as file: json.dump(train_log, file, indent=2) # == Save training prompt == with open(f"{lora_file_path}/{folder_save}/training_prompt.json", 'w', encoding='utf-8') as file: json.dump(train_template, file, indent=2) def on_substep_end(self, args: transformers.TrainingArguments, state: transformers.TrainerState, control: transformers.TrainerControl, **kwargs): tracked.current_steps += 1 if WANT_INTERRUPT: control.should_epoch_stop = True control.should_training_stop = True def on_log(self, args: transformers.TrainingArguments, state: transformers.TrainerState, control: transformers.TrainerControl, logs, **kwargs): train_log.update(logs) train_log.update({"current_steps": tracked.current_steps}) if WANT_INTERRUPT: print("\033[1;31;1mInterrupted by user\033[0;37;0m") print(f"\033[1;30;40mStep: {tracked.current_steps:6} \033[0;37;0m", end='') entry = { 'current_steps': int(train_log.get('current_steps',0)), 'loss': float(train_log.get('loss', 0.0)), 'learning_rate': float(train_log.get('learning_rate', 0.0)), 'epoch': float(train_log.get('epoch', 0.0)) } # Add the entry to the continuous log train_log_graph.append(entry) # Save the graph log for now, we can later generate full graph with open(f"{lora_file_path}/training_graph.json", 'w') as file: json.dump(train_log_graph, file, indent=4) if 'loss' in logs: loss = float(logs['loss']) if loss <= stop_at_loss: control.should_epoch_stop = True control.should_training_stop = True print(f"\033[1;31;1mStop Loss {stop_at_loss} reached.\033[0;37;0m") # FPHAM SAMPLE REQ Transformers error handling gradient_accumulation_max = int(train_data.num_rows)//micro_batch_size if gradient_accumulation_max < gradient_accumulation_steps: print(f"\033[1;31;1mWARNING: Current gradient accumulation is too high for the amount of training data.\033[0;37;0m") print(f"Gradient accumulation: {gradient_accumulation_steps} should be less than: {gradient_accumulation_max}. \033[1;31;1mThis could crash Accelerate/Transformers\033[0;37;0m") #min_batchSize = sample_req*micro_batch_size print(f"Preferable fix: \033[1;31;1mIncrease the size of dataset\033[0;37;0m") print(f"... or Decrerase Gradient Accumulation \033[1;31;1m{gradient_accumulation_steps}\033[0;37;0m to below {gradient_accumulation_max}") gradient_accumulation_steps = max(1,gradient_accumulation_max-1) print(f"Last resort fix for this run: Lowering Gradient accumulation to {gradient_accumulation_steps}. [Good luck]") else: print(f"Data Size Check: Gradient accumulation: {gradient_accumulation_steps} <= Blocks/Batch {gradient_accumulation_max} ... [OK]") #END OF FPHAM SAMPLE REQ # FPHAM Custom Scheduler == custom_scheduller = False lr_scheduler_type_arg = lr_scheduler_type if lr_scheduler_type == 'FP_low_epoch_annealing': custom_scheduller = True lr_scheduler_type_arg = 'cosine' elif lr_scheduler_type == 'FP_half_time_annealing': custom_scheduller = True lr_scheduler_type_arg = 'constant' args=transformers.TrainingArguments( report_to=report_to if report_to != "None" else None, per_device_train_batch_size=micro_batch_size, gradient_accumulation_steps=gradient_accumulation_steps, warmup_steps=math.ceil(warmup_steps / gradient_accumulation_steps), warmup_ratio = warmup_ratio, num_train_epochs=epochs, learning_rate=actual_lr, fp16=False if shared.args.cpu else True, optim=optimizer, logging_steps=1, evaluation_strategy="steps" if eval_data is not None else "no", eval_steps=math.ceil(eval_steps / gradient_accumulation_steps) if eval_data is not None else None, save_strategy="steps" if eval_data is not None else "no", output_dir=lora_file_path, lr_scheduler_type=lr_scheduler_type_arg, load_best_model_at_end=eval_data is not None, # TODO: Enable multi-device support ddp_find_unused_parameters=None, no_cuda=shared.args.cpu, ) if custom_scheduller: trainer = FPSchedulerTrainer( model=lora_model, train_dataset=train_data, eval_dataset=eval_data, args=args, data_collator=transformers.DataCollatorForLanguageModeling(shared.tokenizer, mlm=False), callbacks=list([Callbacks()]) ) else: trainer = transformers.Trainer( model=lora_model, train_dataset=train_data, eval_dataset=eval_data, args=args, data_collator=transformers.DataCollatorForLanguageModeling(shared.tokenizer, mlm=False), callbacks=list([Callbacks()]) ) # END OF FPHAM CUSTOM SCHEDULER lora_model.config.use_cache = False if torch.__version__ >= "2" and sys.platform != "win32": lora_model = torch.compile(lora_model) # == Save parameters for reuse == with open(f"{lora_file_path}/training_parameters.json", 'w', encoding='utf-8') as file: vars = locals() json.dump({x: vars[x] for x in PARAMETERS}, file, indent=2) # == Save training prompt == with open(f"{lora_file_path}/training_prompt.json", 'w', encoding='utf-8') as file: json.dump(train_template, file, indent=2) # == Main run and monitor loop == logger.info("Starting training...") yield "Starting..." lora_trainable_param, lora_all_param = calc_trainable_parameters(lora_model) projections_string = ", ".join([projection.replace("_proj", "") for projection in model_to_lora_modules[model_id]]) print(f"Training '{model_id}' model using ({projections_string}) projections") if lora_all_param > 0: print(f"Trainable params: {lora_trainable_param:,d} ({100 * lora_trainable_param / lora_all_param:.4f} %), All params: {lora_all_param:,d} (Model: {model_all_params:,d})") train_log.update({"base_model_name": shared.model_name}) train_log.update({"base_model_class": shared.model.__class__.__name__}) train_log.update({"base_loaded_in_4bit": getattr(lora_model, "is_loaded_in_4bit", False)}) train_log.update({"base_loaded_in_8bit": getattr(lora_model, "is_loaded_in_8bit", False)}) train_log.update({"projections": projections_string}) if stop_at_loss > 0: print(f"Monitoring loss \033[1;31;1m(Auto-Stop at: {stop_at_loss})\033[0;37;0m") if WANT_INTERRUPT: yield "Interrupted before start." return def log_train_dataset(trainer): decoded_entries = [] # Try to decode the entries and write the log file try: # Iterate over the first 10 elements in the dataset (or fewer if there are less than 10) for i in range(min(10, len(trainer.train_dataset))): decoded_text = shared.tokenizer.decode(trainer.train_dataset[i]['input_ids']) decoded_entries.append({"value": decoded_text}) # Write the log file Path('logs').mkdir(exist_ok=True) with open(Path('logs/train_dataset_sample.json'), 'w') as json_file: json.dump(decoded_entries, json_file, indent=4) logger.info("Log file 'train_dataset_sample.json' created in the 'logs' directory.") except Exception as e: logger.error(f"Failed to create log file due to error: {e}") def threaded_run(): log_train_dataset(trainer) trainer.train() # Note: save in the thread in case the gradio thread breaks (eg browser closed) lora_model.save_pretrained(lora_file_path) logger.info("LoRA training run is completed and saved.") # Save log with open(f"{lora_file_path}/training_log.json", 'w', encoding='utf-8') as file: json.dump(train_log, file, indent=2) thread = threading.Thread(target=threaded_run) thread.start() last_step = 0 start_time = time.perf_counter() while thread.is_alive(): time.sleep(0.5) if WANT_INTERRUPT: yield "Interrupting, please wait... *(Run will stop after the current training step completes.)*" elif tracked.current_steps != last_step: last_step = tracked.current_steps time_elapsed = time.perf_counter() - start_time lastloss = float(train_log.get('loss', 0.0)) non_serialized_params.update({"training_loop": True}) if lastloss > 0: lastloss_str = f", ... Current Loss: `{lastloss:.2f}`" else: lastloss_str = "" if time_elapsed <= 0: timer_info = "" total_time_estimate = 999 else: its = tracked.current_steps / time_elapsed if its > 1: timer_info = f"`{its:.2f}` it/s" else: timer_info = f"`{1.0/its:.2f}` s/it" total_time_estimate = (1.0 / its) * (tracked.max_steps) if stop_at_loss != non_serialized_params['stop_at_loss']: stop_at_loss = non_serialized_params['stop_at_loss'] print(f"Stop at loss changed \033[1;31;1m(Auto-Stop at: {stop_at_loss})\033[0;37;0m") yield f"Running... **{tracked.current_steps}** / **{tracked.max_steps}** ... {timer_info}, {format_time(time_elapsed)} / {format_time(total_time_estimate)} ... {format_time(total_time_estimate - time_elapsed)} remaining {lastloss_str}" # Saving in the train thread might fail if an error occurs, so save here if so. non_serialized_params.update({"training_loop": False}) if not tracked.did_save: logger.info("Training complete, saving...") lora_model.save_pretrained(lora_file_path) if WANT_INTERRUPT: logger.info("Training interrupted.") yield f"Interrupted by user. LoRA saved to `{lora_file_path}`." else: logger.info("Training complete!") yield f"Done! LoRA saved to `{lora_file_path}`.\n\nBefore testing your new LoRA, make sure to first reload the model, as it is currently dirty from training." create_graph(lora_file_path, lora_name) def format_time(seconds: float): if seconds < 120: return f"`{seconds:.0f}` seconds" minutes = seconds / 60 if minutes < 120: return f"`{minutes:.0f}` minutes" hours = minutes / 60 return f"`{hours:.0f}` hours"