Mirrors/text-generation-webui
1
0
Fork 0
mirror of https://github.com/oobabooga/text-generation-webui.git synced 2024-11-22 16:17:57 +01:00
Code Issues Packages Projects Releases Wiki Activity

Style changes

Browse Source
This commit is contained in:
oobabooga 2023-07-11 18:49:06 -07:00
parent bfafd07f44
commit e3810dff40
1 changed files with 8 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
modules/training.py
View File

@ -1,19 +1,17 @@
import json import json
import math import math
import random import random
import shutil
import sys import sys
import threading import threading
import time import time
import traceback import traceback
from datetime import datetime
from pathlib import Path from pathlib import Path
import gradio as gr import gradio as gr
import torch import torch
import transformers import transformers
import shutil
from datetime import datetime
from datasets import Dataset, load_dataset from datasets import Dataset, load_dataset
from peft import ( from peft import (
LoraConfig, LoraConfig,
@ -223,7 +221,7 @@ def backup_adapter(input_folder):
creation_date_str = creation_date.strftime("Backup-%Y-%m-%d") creation_date_str = creation_date.strftime("Backup-%Y-%m-%d")
# Create the new subfolder # Create the new subfolder
subfolder_path = Path(f"{input_folder}/{creation_date_str}") subfolder_path = Path(f"{input_folder}/{creation_date_str}")
subfolder_path.mkdir(parents=True, exist_ok=True) subfolder_path.mkdir(parents=True, exist_ok=True)
# Check if the file already exists in the subfolder # Check if the file already exists in the subfolder
@ -240,6 +238,7 @@ def backup_adapter(input_folder):
except Exception as e: except Exception as e:
print("An error occurred in backup_adapter:", str(e)) print("An error occurred in backup_adapter:", str(e))
def calc_trainable_parameters(model): def calc_trainable_parameters(model):
trainable_params = 0 trainable_params = 0
all_param = 0 all_param = 0
@ -252,8 +251,8 @@ def calc_trainable_parameters(model):
all_param += num_params all_param += num_params
if param.requires_grad: if param.requires_grad:
trainable_params += num_params trainable_params += num_params
return trainable_params,all_param 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, overlap_len: int, newline_favor_len: int, higher_rank_limit: bool, warmup_steps: int, optimizer: str, hard_cut_string: str, train_only_after: str, stop_at_loss: float): 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, overlap_len: int, newline_favor_len: int, higher_rank_limit: bool, warmup_steps: int, optimizer: str, hard_cut_string: str, train_only_after: str, stop_at_loss: float):
@ -559,10 +558,9 @@ def do_train(lora_name: str, always_override: bool, save_steps: int, micro_batch
yield "Starting..." yield "Starting..."
lora_trainable_param, lora_all_param = calc_trainable_parameters(lora_model) lora_trainable_param, lora_all_param = calc_trainable_parameters(lora_model)
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})")
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_name": shared.model_name})
train_log.update({"base_model_class": shared.model.__class__.__name__}) train_log.update({"base_model_class": shared.model.__class__.__name__})