Spaces:
Runtime error
Runtime error
File size: 2,502 Bytes
f481cbe 3eea93e f481cbe f6c29be f481cbe 1cd5d29 f481cbe 1cd5d29 f481cbe 694a287 f481cbe 5d09bfd f481cbe 694a287 f481cbe 9b37590 f481cbe 9b37590 f481cbe 694a287 f481cbe 9b37590 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 |
import os
from uuid import uuid4
import pandas as pd
from datasets import load_dataset
import subprocess
# from dotenv import load_dotenv,find_dotenv
# load_dotenv(find_dotenv(),override=True)
# Load dataset
dataset_name = 'ai-aerospace/ams_data_train_generic_v0.1_100'
dataset=load_dataset(dataset_name)
# Write dataset files into data directory
data_directory = './fine_tune_data/'
# Create the data directory if it doesn't exist
os.makedirs(data_directory, exist_ok=True)
# Write the train data to a CSV file
train_data='train_data'
train_filename = os.path.join(data_directory, train_data)
dataset['train'].to_pandas().to_csv(train_filename, columns=['text'], index=False)
# Write the validation data to a CSV file
validation_data='validation_data'
validation_filename = os.path.join(data_directory, validation_data)
dataset['validation'].to_pandas().to_csv(validation_filename, columns=['text'], index=False)
# Define project parameters
username='ai-aerospace'
project_name='./llms/'+'ams_data_train-100_'+str(uuid4())
repo_name='ams-data-train-100-'+str(uuid4())
# model_name='TinyLlama/TinyLlama-1.1B-Chat-v0.1'
model_name='mistralai/Mistral-7B-v0.1'
# Save parameters to environment variables
os.environ["project_name"] = project_name
os.environ["model_name"] = model_name
os.environ["repo_id"] = username+'/'+repo_name
os.environ["train_data"] = train_data
os.environ["validation_data"] = validation_data
os.environ["data_directory"] = data_directory
print("project_name:", project_name)
print("model_name:", model_name)
print("repo_id:", username+'/'+repo_name)
print("train_data:", train_data)
print("validation_data:", validation_data)
print("data_directory:", data_directory)
# Set .venv and execute the autotrain script
# To see all parameters: autotrain llm --help
# !autotrain llm --train --project_name my-llm --model TinyLlama/TinyLlama-1.1B-Chat-v0.1 --data_path . --use-peft --use_int4 --learning_rate 2e-4 --train_batch_size 6 --num_train_epochs 3 --trainer sft
command="""
autotrain llm --train \
--project_name ${project_name} \
--model ${model_name} \
--data_path ${data_directory} \
--train_split ${train_data} \
--valid_split ${validation_data} \
--use-peft \
--learning_rate 2e-4 \
--train_batch_size 6 \
--num_train_epochs 3 \
--trainer sft \
--push_to_hub \
--repo_id ${repo_id} \
--token $HUGGINGFACE_TOKEN
"""
# Use subprocess.run() to execute the command
subprocess.run(command, shell=True, check=True) |