metadata
tags:
- summarization
- summary
- booksum
- long-document
- long-form
license: apache-2.0
datasets:
- kmfoda/booksum
metrics:
- rouge
widget:
- text: >-
large earthquakes along a given fault segment do not occur at random
intervals because it takes time to accumulate the strain energy for the
rupture. The rates at which tectonic plates move and accumulate strain at
their boundaries are approximately uniform. Therefore, in first
approximation, one may expect that large ruptures of the same fault
segment will occur at approximately constant time intervals. If subsequent
main shocks have different amounts of slip across the fault, then the
recurrence time may vary, and the basic idea of periodic mainshocks must
be modified. For great plate boundary ruptures the length and slip often
vary by a factor of 2. Along the southern segment of the San Andreas fault
the recurrence interval is 145 years with variations of several decades.
The smaller the standard deviation of the average recurrence interval, the
more specific could be the long term prediction of a future mainshock.
example_title: earthquakes
- text: ' A typical feed-forward neural field algorithm. Spatiotemporal coordinates are fed into a neural network that predicts values in the reconstructed domain. Then, this domain is mapped to the sensor domain where sensor measurements are available as supervision. Class and Section Problems Addressed Generalization (Section 2) Inverse problems, ill-posed problems, editability; symmetries. Hybrid Representations (Section 3) Computation & memory efficiency, representation capacity, editability: Forward Maps (Section 4) Inverse problems Network Architecture (Section 5) Spectral bias, integration & derivatives. Manipulating Neural Fields (Section 6) Edit ability, constraints, regularization. Table 2: The five classes of techniques in the neural field toolbox each addresses problems that arise in learning, inference, and control. (Section 3). We can supervise reconstruction via differentiable forward maps that transform Or project our domain (e.g, 3D reconstruction via 2D images; Section 4) With appropriate network architecture choices, we can overcome neural network spectral biases (blurriness) and efficiently compute derivatives and integrals (Section 5). Finally, we can manipulate neural fields to add constraints and regularizations, and to achieve editable representations (Section 6). Collectively, these classes constitute a ''toolbox'' of techniques to help solve problems with neural fields There are three components in a conditional neural field: (1) An encoder or inference function € that outputs the conditioning latent variable 2 given an observation 0 E(0) =2. 2 is typically a low-dimensional vector, and is often referred to aS a latent code Or feature code_ (2) A mapping function 4 between Z and neural field parameters O: Y(z) = O; (3) The neural field itself $. The encoder € finds the most probable z given the observations O: argmaxz P(2/0). The decoder maximizes the inverse conditional probability to find the most probable 0 given Z: arg- max P(Olz). We discuss different encoding schemes with different optimality guarantees (Section 2.1.1), both global and local conditioning (Section 2.1.2), and different mapping functions Y (Section 2.1.3) 2. Generalization Suppose we wish to estimate a plausible 3D surface shape given a partial or noisy point cloud. We need a suitable prior over the sur- face in its reconstruction domain to generalize to the partial observations. A neural network expresses a prior via the function space of its architecture and parameters 0, and generalization is influenced by the inductive bias of this function space (Section 5).'
example_title: scientific paper
- text: >-
Is a else or outside the cob and tree written being of early client rope
and you have is for good reasons. On to the ocean in Orange for time. By's
the aggregate we can bed it yet. Why this please pick up on a sort is do
and also M Getoi's nerocos and do rain become you to let so is his brother
is made in use and Mjulia's's the lay major is aging Masastup coin present
sea only of Oosii rooms set to you We do er do we easy this private
oliiishs lonthen might be okay. Good afternoon everybody. Welcome to this
lecture of Computational Statistics. As you can see, I'm not socially my
name is Michael Zelinger. I'm one of the task for this class and you might
have already seen me in the first lecture where I made a quick appearance.
I'm also going to give the tortillas in the last third of this course. So
to give you a little bit about me, I'm a old student here with better
Bulman and my research centres on casual inference applied to biomedical
disasters, so that could be genomics or that could be hospital data. If
any of you is interested in writing a bachelor thesis, a semester paper
may be mastathesis about this topic feel for reach out to me. you have my
name on models and my email address you can find in the directory I'd Be
very happy to talk about it. you do not need to be sure about it, we can
just have a chat. So with that said, let's get on with the lecture.
There's an exciting topic today I'm going to start by sharing some slides
with you and later on during the lecture we'll move to the paper. So bear
with me for a few seconds. Well, the projector is starting up. Okay, so
let's get started. Today's topic is a very important one. It's about a
technique which really forms one of the fundamentals of data science,
machine learning, and any sort of modern statistics. It's called cross
validation. I know you really want to understand this topic I Want you to
understand this and frankly, nobody's gonna leave Professor Mineshousen's
class without understanding cross validation. So to set the stage for
this, I Want to introduce you to the validation problem in computational
statistics. So the problem is the following: You trained a model on
available data. You fitted your model, but you know the training data you
got could always have been different and some data from the environment.
Maybe it's a random process. You do not really know what it is, but you
know that somebody else who gets a different batch of data from the same
environment they would get slightly different training data and you do not
care that your method performs as well. On this training data. you want to
to perform well on other data that you have not seen other data from the
same environment. So in other words, the validation problem is you want to
quantify the performance of your model on data that you have not seen. So
how is this even possible? How could you possibly measure the performance
on data that you do not know The solution to? This is the following
realization is that given that you have a bunch of data, you were in
charge. You get to control how much that your model sees. It works in the
following way: You can hide data firms model. Let's say you have a
training data set which is a bunch of doubtless so X eyes are the features
those are typically hide and national vector. It's got more than one
dimension for sure. And the why why eyes. Those are the labels for
supervised learning. As you've seen before, it's the same set up as we
have in regression. And so you have this training data and now you choose
that you only use some of those data to fit your model. You're not going
to use everything, you only use some of it the other part you hide from
your model. And then you can use this hidden data to do validation from
the point of you of your model. This hidden data is complete by unseen. In
other words, we solve our problem of validation.
example_title: transcribed audio - lecture
- text: >-
Transformer-based models have shown to be very useful for many NLP tasks.
However, a major limitation of transformers-based models is its O(n^2)O(n
2) time & memory complexity (where nn is sequence length). Hence, it's
computationally very expensive to apply transformer-based models on long
sequences n > 512n>512. Several recent papers, e.g. Longformer, Performer,
Reformer, Clustered attention try to remedy this problem by approximating
the full attention matrix. You can checkout 🤗's recent blog post in case
you are unfamiliar with these models.
BigBird (introduced in paper) is one of such recent models to address this
issue. BigBird relies on block sparse attention instead of normal
attention (i.e. BERT's attention) and can handle sequences up to a length
of 4096 at a much lower computational cost compared to BERT. It has
achieved SOTA on various tasks involving very long sequences such as long
documents summarization, question-answering with long contexts.
BigBird RoBERTa-like model is now available in 🤗Transformers. The goal of
this post is to give the reader an in-depth understanding of big bird
implementation & ease one's life in using BigBird with 🤗Transformers.
But, before going into more depth, it is important to remember that the
BigBird's attention is an approximation of BERT's full attention and
therefore does not strive to be better than BERT's full attention, but
rather to be more efficient. It simply allows to apply transformer-based
models to much longer sequences since BERT's quadratic memory requirement
quickly becomes unbearable. Simply put, if we would have ∞ compute & ∞
time, BERT's attention would be preferred over block sparse attention
(which we are going to discuss in this post).
If you wonder why we need more compute when working with longer sequences,
this blog post is just right for you!
Some of the main questions one might have when working with standard
BERT-like attention include:
Do all tokens really have to attend to all other tokens? Why not compute
attention only over important tokens? How to decide what tokens are
important? How to attend to just a few tokens in a very efficient way? In
this blog post, we will try to answer those questions.
What tokens should be attended to? We will give a practical example of how
attention works by considering the sentence 'BigBird is now available in
HuggingFace for extractive question answering'. In BERT-like attention,
every word would simply attend to all other tokens.
Let's think about a sensible choice of key tokens that a queried token
actually only should attend to by writing some pseudo-code. Will will
assume that the token available is queried and build a sensible list of
key tokens to attend to.
>>> # let's consider following sentence as an example >>> example =
['BigBird', 'is', 'now', 'available', 'in', 'HuggingFace', 'for',
'extractive', 'question', 'answering']
>>> # further let's assume, we're trying to understand the representation
of 'available' i.e. >>> query_token = 'available' >>> # We will initialize
an empty `set` and fill up the tokens of our interest as we proceed in
this section. >>> key_tokens = [] # => currently 'available' token doesn't
have anything to attend Nearby tokens should be important because, in a
sentence (sequence of words), the current word is highly dependent on
neighboring past & future tokens. This intuition is the idea behind the
concept of sliding attention.
example_title: bigbird blog intro
- text: >-
To be fair, you have to have a very high IQ to understand Rick and Morty.
The humour is extremely subtle, and without a solid grasp of theoretical
physics most of the jokes will go over a typical viewer's head. There's
also Rick's nihilistic outlook, which is deftly woven into his
characterisation- his personal philosophy draws heavily from Narodnaya
Volya literature, for instance. The fans understand this stuff; they have
the intellectual capacity to truly appreciate the depths of these jokes,
to realise that they're not just funny- they say something deep about
LIFE. As a consequence people who dislike Rick & Morty truly ARE idiots-
of course they wouldn't appreciate, for instance, the humour in Rick's
existential catchphrase 'Wubba Lubba Dub Dub,' which itself is a cryptic
reference to Turgenev's Russian epic Fathers and Sons. I'm smirking right
now just imagining one of those addlepated simpletons scratching their
heads in confusion as Dan Harmon's genius wit unfolds itself on their
television screens. What fools.. how I pity them. 😂
And yes, by the way, i DO have a Rick & Morty tattoo. And no, you cannot
see it. It's for the ladies' eyes only- and even then they have to
demonstrate that they're within 5 IQ points of my own (preferably lower)
beforehand. Nothin personnel kid 😎
example_title: Richard & Mortimer
parameters:
max_length: 64
min_length: 8
no_repeat_ngram_size: 3
early_stopping: true
repetition_penalty: 3.5
length_penalty: 0.3
encoder_no_repeat_ngram_size: 3
num_beams: 4
model-index:
- name: pszemraj/long-t5-tglobal-base-16384-book-summary
results:
- task:
type: summarization
name: Summarization
dataset:
name: kmfoda/booksum
type: kmfoda/booksum
config: kmfoda--booksum
split: test
metrics:
- name: ROUGE-1
type: rouge
value: 36.4085
verified: true
- name: ROUGE-2
type: rouge
value: 6.0646
verified: true
- name: ROUGE-L
type: rouge
value: 16.7209
verified: true
- name: ROUGE-LSUM
type: rouge
value: 33.3405
verified: true
- name: loss
type: loss
value: .nan
verified: true
- name: gen_len
type: gen_len
value: 252.8099
verified: true
- task:
type: summarization
name: Summarization
dataset:
name: samsum
type: samsum
config: samsum
split: test
metrics:
- name: ROUGE-1
type: rouge
value: 30.9047
verified: true
- name: ROUGE-2
type: rouge
value: 7.4715
verified: true
- name: ROUGE-L
type: rouge
value: 22.3962
verified: true
- name: ROUGE-LSUM
type: rouge
value: 26.9094
verified: true
- name: loss
type: loss
value: .nan
verified: true
- name: gen_len
type: gen_len
value: 46.7973
verified: true
- task:
type: summarization
name: Summarization
dataset:
name: cnn_dailymail
type: cnn_dailymail
config: 3.0.0
split: test
metrics:
- name: ROUGE-1
type: rouge
value: 30.5942
verified: true
- name: ROUGE-2
type: rouge
value: 7.252
verified: true
- name: ROUGE-L
type: rouge
value: 17.7156
verified: true
- name: ROUGE-LSUM
type: rouge
value: 27.2881
verified: true
- name: loss
type: loss
value: .nan
verified: true
- name: gen_len
type: gen_len
value: 125.2507
verified: true
- task:
type: summarization
name: Summarization
dataset:
name: xsum
type: xsum
config: default
split: test
metrics:
- name: ROUGE-1
type: rouge
value: 20.3648
verified: true
- name: ROUGE-2
type: rouge
value: 3.4126
verified: true
- name: ROUGE-L
type: rouge
value: 13.6168
verified: true
- name: ROUGE-LSUM
type: rouge
value: 15.8313
verified: true
- name: loss
type: loss
value: .nan
verified: true
- name: gen_len
type: gen_len
value: 82.2177
verified: true
- task:
type: summarization
name: Summarization
dataset:
name: billsum
type: billsum
config: default
split: test
metrics:
- name: ROUGE-1
type: rouge
value: 39.6378
verified: true
- name: ROUGE-2
type: rouge
value: 13.0017
verified: true
- name: ROUGE-L
type: rouge
value: 23.0255
verified: true
- name: ROUGE-LSUM
type: rouge
value: 32.9943
verified: true
- name: loss
type: loss
value: 1.9428048133850098
verified: true
- name: gen_len
type: gen_len
value: 162.3588
verified: true
long-t5-tglobal-base-16384 + BookSum
- summarize long text and get a SparkNotes-esque summary of arbitrary topics!
- generalizes reasonably well to academic & narrative text.
- A simple example/use case on ASR is here. An example notebook can be found here.
Cheeky Proof-of-Concept
A summary of the infamous navy seals copypasta:
The narrator tells us that he's graduated from the Navy seals and has been involved in many secret raids. He's also one of the best snipers in the entire U.S. military. He promises to "wipe you out with precision" when they meet again.
Model description
A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset:
- 30+ epochs of fine-tuning from the base model on V100/A100 GPUs
- all training used 16384 token input / 1024 max output
Read the paper by Guo et al. here: LongT5: Efficient Text-To-Text Transformer for Long Sequences
How-To in Python
Install/update transformers pip install -U transformers
Summarize text with pipeline:
from transformers import pipeline
summarizer = pipeline(
'summarization',
'pszemraj/long-t5-tglobal-base-16384-book-summary',
)
long_text = "Here is a lot of text I don't want to read. Replace me"
result = summarizer(long_text)
print(result[0]['summary_text'])
Pass other parameters related to beam search textgen when calling summarizer to get even higher quality results.
Intended uses & limitations
- The current checkpoint is fairly well converged but will be updated if further improvements can be made.
- Compare performance to LED-base trained on the same dataset (API gen parameters are the same).
- while this model seems to improve upon factual consistency, do not take summaries to be foolproof and check things that seem odd.
Training and evaluation data
kmfoda/booksum dataset on HuggingFace - read the original paper here. Summaries longer than 1024 LongT5 tokens were filtered out to prevent the model from learning to generate "partial" summaries.
NOTE: early checkpoints of this model were trained on a "smaller" subsection of the dataset as it was filtered for summaries of 1024 characters. This was subsequently caught and adjusted to 1024 tokens and then trained further for 10+ epochs.
Training procedure
Updates:
- July 22, 2022: updated to a fairly converged checkpoint
- July 3, 2022: Added a new version with several epochs of additional training that is more performant in general.
Training hyperparameters
The following hyperparameters were used during the most recent training round*:
- learning_rate: 0.0005
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- gradient_accumulation_steps: 128
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.01
- num_epochs: 2
*Prior training sessions used roughly similar parameters; multiple sessions were required as this takes aeons to train
Training results
Framework versions
- Transformers 4.20.1
- Pytorch 1.10.0+cu113
- Datasets 2.3.2
- Tokenizers 0.12.1