Motoki Wu tokestermw

I hereby claim:

I am tokestermw on github.
I am motoki (https://keybase.io/motoki) on keybase.
I have a public key whose fingerprint is 26C6 F8AB C16D 50E4 3A97 05C2 B235 7159 51D6 074D

To claim this, I am signing this object:

Where A is a class (e.g. definite article), and B is another class (e.g. indefinite article). O is the null class.

The cat had a dog .
 A   O   O  B  O  O

v1

| Real | Predicted | Verdict

Supplementary Materials for Generating Wikipedia by Summarizing Long Sequences

	# train_grpo.py
	#
	# See https://github.com/willccbb/verifiers for ongoing developments
	#
	import re
	import torch
	from datasets import load_dataset, Dataset
	from transformers import AutoTokenizer, AutoModelForCausalLM
	from peft import LoraConfig
	from trl import GRPOConfig, GRPOTrainer

	import torch
	import torch.nn as nn
	import torch.nn.functional as F

	# helpers

	def make_unit_length(x, epsilon=1e-6):
	norm = x.norm(p=2, dim=-1, keepdim=True)
	return x.div(norm + epsilon)

	import random


	def augmentation_fun(x, augment_by=3):
	# augment the original data point by 3
	return [x + random.random() * 2 - 1 for i in range(augment_by)]


	def train_loop(dataset, do_augment=False):
	# emit one data point at a time

	"""
	To use it inside ELMo script

	To get the embeddings:

	allennlp elmo sample_sents.txt out1.hdf5 --top
	python -c "import h5py; f = h5py.File('out1.hdf5'); print(f['0'][:], f['0'].shape)"

	To get probabilities:

	import tensorflow as tf

	from tensorflow.python.ops.metrics_impl import _streaming_confusion_matrix


	# almost the same as
	def confusion_matrix(labels, predictions, num_classes, weights=None):
	total_cm, update_op = _streaming_confusion_matrix(
	labels, predictions, num_classes, weights=weights)

	import torch, torch.nn as nn
	from torch.autograd import Variable

	text = ['BOS', 'How', 'are', 'you', 'EOS']
	seq_len = len(text)
	batch_size = 1
	embedding_size = 1
	hidden_size = 1
	output_size = 1