thomwolf

# Copyright (c) 2019-present, Thomas Wolf.
# All rights reserved. This source code is licensed under the MIT-style license.
""" A very small and self-contained gist to train a GPT-2 transformer model on wikitext-103 """
import os
from collections import namedtuple
from tqdm import tqdm
import torch
import torch.nn as nn
from torch.utils.data import DataLoader
from ignite.engine import Engine, Events

thomwolf

""" To use: install LLM studio (or Ollama), clone OpenVoice, run this script in the OpenVoice directory
    git clone https://github.com/myshell-ai/OpenVoice
    cd OpenVoice
    git clone https://huggingface.co/myshell-ai/OpenVoice
    cp -r OpenVoice/* .
    pip install whisper pynput pyaudio
"""

from openai import OpenAI
import time

thomwolf

""" To use: install LLM studio (or Ollama), clone OpenVoice, run this script in the OpenVoice directory
    git clone https://github.com/myshell-ai/OpenVoice
    cd OpenVoice
    git clone https://huggingface.co/myshell-ai/OpenVoice
    cp -r OpenVoice/* .
    pip install whisper pynput pyaudio
"""
from dataclasses import dataclass
from typing import Optional
import random

thomwolf

import os; import psutil; import timeit
from datasets import load_dataset

mem_before = psutil.Process(os.getpid()).memory_info().rss >> 20
wiki = load_dataset("wikipedia", "20200501.en", split='train')
mem_after = psutil.Process(os.getpid()).memory_info().rss >> 20
print(f"RAM memory used: {(mem_after - mem_before)} MB")

s = """batch_size = 1000
for i in range(0, len(wiki), batch_size):

thomwolf

model.zero_grad()                                   # Reset gradients tensors
for i, (inputs, labels) in enumerate(training_set):
    predictions = model(inputs)                     # Forward pass
    loss = loss_function(predictions, labels)       # Compute loss function
    loss = loss / accumulation_steps                # Normalize our loss (if averaged)
    loss.backward()                                 # Backward pass
    if (i+1) % accumulation_steps == 0:             # Wait for several backward steps
        optimizer.step()                            # Now we can do an optimizer step
        model.zero_grad()                           # Reset gradients tensors
        if (i+1) % evaluation_steps == 0:           # Evaluate the model when we...

thomwolf

def top_k_top_p_filtering(logits, top_k=0, top_p=0.0, filter_value=-float('Inf')):
    """ Filter a distribution of logits using top-k and/or nucleus (top-p) filtering
        Args:
            logits: logits distribution shape (vocabulary size)
            top_k >0: keep only top k tokens with highest probability (top-k filtering).
            top_p >0.0: keep the top tokens with cumulative probability >= top_p (nucleus filtering).
                Nucleus filtering is described in Holtzman et al. (http://arxiv.org/abs/1904.09751)
    """
    assert logits.dim() == 1  # batch size 1 for now - could be updated for more but the code would be less clear
    top_k = min(top_k, logits.size(-1))  # Safety check

thomwolf

from torch.optim import Optimizer

class AdamW(Optimizer):
    """
       Implements Adam algorithm with weight decay fix in PyTorch
       Paper: Fixing Weight Decay Regularization in Adam by Ilya Loshchilov, Frank Hutter
       https://arxiv.org/abs/1711.05101
    """
    def __init__(self, params, lr, b1=0.9, b2=0.999, e=1e-8, l2=0,
                 vector_l2=False, max_grad_norm=-1, **kwargs):

thomwolf

##+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
## Created by: Hang Zhang, Rutgers University, Email: zhang.hang@rutgers.edu
## Modified by Thomas Wolf, HuggingFace Inc., Email: thomas@huggingface.co
## Copyright (c) 2017-2018
##
## This source code is licensed under the MIT-style license found in the
## LICENSE file in the root directory of this source tree
##+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

"""Encoding Data Parallel"""

thomwolf

class MetaLearner(nn.Module):
    """ Bare Meta-learner class
        Should be added: intialization, hidden states, more control over everything
    """
    def __init__(self, model):
        super(MetaLearner, self).__init__()
        self.weights = Parameter(torch.Tensor(1, 2))

    def forward(self, forward_model, backward_model):
        """ Forward optimizer with a simple linear neural net

thomwolf

import json
from pytorch_pretrained_bert import cached_path

url = "https://s3.amazonaws.com/datasets.huggingface.co/personachat/personachat_self_original.json"

# Download and load JSON dataset
personachat_file = cached_path(url)
with open(personachat_file, "r", encoding="utf-8") as f:
    dataset = json.loads(f.read())