Datasets:

Prikshit7766

/

huggingface_forum

like 1

Privacy enhancing technologies in model development

https://discuss.huggingface.co/t/privacy-enhancing-technologies-in-model-development/26521

Dear Community!I am a PhD researcher at the London School of Economics and am exploring the field of privacy enhancing technology (PET) usage in ML model development (e.g., approaches like differential privacy) and their societal and organisational implications. However, I am struggling with finding data sources that allow me to analyse the diffusion of PETs (in the best case across companies or locations). One thought was that model cards on Huggingface would capture the usage of PETs - is that the case? do you have other data sources in mind I that indicate PET usage?Thanks so much for your support!

2022-11-22T16:09:41Z

Conversational QA pretrained model?

https://discuss.huggingface.co/t/conversational-qa-pretrained-model/26441

I was wondering if we’ve a pretrained model for Conversational QA. We’ve conversational AI, we’ve QA which needs context, do we’ve anything which when given a context, replies technical questions but also acts as a chat botAny help is appreciated

2022-11-21T10:52:30Z

Composition Training/Validation Split of AutoTrain

https://discuss.huggingface.co/t/composition-training-validation-split-of-autotrain/26328

Hey everyone,is there any documentation about how AutoTrain splits your data into training and validation data? I selected the option that it should automatically do so. I conducted binary text classification with BERT.It would be great to know in order to report a percentage split in a research project.Thanks!Bestrob

2022-11-18T19:04:20Z

Do the common tricks in transformers help with RNNs?

https://discuss.huggingface.co/t/do-the-common-tricks-in-transformers-help-with-rnns/25879

Does anybody know any research or work that utilizes common tricks (layer norm, masked language training, etc) commonly used with transformers with RNNs?Do these things still help improve RNNs? If not, are there reasons you think these techniques would/would not translate to rnns?

2022-11-10T17:48:33Z

Metadata of NLP datasets

https://discuss.huggingface.co/t/metadata-of-nlp-datasets/25603

Hi,I’m new to the NLP domain and HuggingFace ecosystem.I wanted to some suggestions on where to read about the meta data of datasets used for NLP.I have worked mostly with vision data so far and simple meta features shared by image datasets in general were:image resolutionNo. of training samplesNo. of classification labelsNo. of channelsWould the text data used in NLP tasks have some such features in common? Aside Number of training samples and number of classification labels. Any thoughts are welcome.Thanks!

2022-11-05T19:51:51Z

I'd like to understand on how to train a neural net with agents and evolution

https://discuss.huggingface.co/t/id-like-to-understand-on-how-to-train-a-neural-net-with-agents-and-evolution/25334

I’d like to understand on how to train a neural net with agents and evolution.It might be easier to think of a game world though i don’t create games.The training will be inside jupyter data.Say I got 10 inputs, the the agent has some value (reward store).The 10 values are unknown and to be interpreted by a NNIt needs to improve its reward though its output is just 3 options, like left/forward/rightNot every move results in a reward so training likely takes time period.Depending on their reactions agents might be in a different scenario, though at some time one selects the best (n) agents. (high reward)And then trains again until a supergood agent is able to interpren the input valuesHow does one create train such a network?,Normally in neural networks one trains a network toward a certain goal, using back prop, ea several inputs to resolve something alike DNN or a LSTM. But the rules here are so different.Anyone knows of some jupyter sample for a DNN training alike that ?.

2022-11-01T14:04:11Z

How to annotate these type of data for custom tr-ocr training

https://discuss.huggingface.co/t/how-to-annotate-these-type-of-data-for-custom-tr-ocr-training/25227

Help

2022-10-30T14:50:15Z

Online/streaming speech recognition

https://discuss.huggingface.co/t/online-streaming-speech-recognition/4456

Are there plans to implement online decoding for the speech recognition models such as wav2vec2 and XLSR? More specifically, to be able to receive audio in short chunks, and output partial transcripts as they become available.MotivationMany use cases are covered by the current wav2vec2 model in the library, involving batch recognition of pre-recorded text. However for an online application that wanted to continuously recognize speech on a live input stream, this may not be sufficient.

2021-03-17T00:22:36Z

Exploring contexts of occurrence of particular words in large datasets

https://discuss.huggingface.co/t/exploring-contexts-of-occurrence-of-particular-words-in-large-datasets/22119

Hi everybody, how are you?. I am currently working on a project where we would like to explore and be able to obtain the contexts of occurrence of particular words or n-grams in large datasets used to train language models, such asGitHub - josecannete/spanish-corpora: Unannotated Spanish 3 Billion Words Corpora.As you can imagine, the problem is that when dealing with such large datasets, conventional strategies like using libraries like pandas and the like require a lot of RAM and computing power, so here are my questions:Does the platform have any tools already available to carry out different types of searches on large datasets, which facilitates this task?Is there some kind of server/service within the platform with enough RAM and computing power that we can access to load the full datasets and use an API to interact from our Space?Thank you very much! Hernán

2022-08-25T23:55:27Z

Explaining medical diagnosis

https://discuss.huggingface.co/t/explaining-medical-diagnosis/24664

Hi there,Have any of you came across a model/set of models for explaining clinical diagnosis.I think it may be similar to text summarisation, but with few additions:it needs to use different vocab (for patients),some parts may not be necessary or even shouldn’t be explained (disturbing content that a real doctor should explain directly)Best,Michal

2022-10-19T11:59:29Z

Attention mask and token ids

https://discuss.huggingface.co/t/attention-mask-and-token-ids/15243

HI,I am taking following wonderful course,TransformesWhile we do padding we pad the sequece with 0 and ask model not to consider the padding. I was wondering if there is some token with id = 0? Because in this case we will be avoiding a token with id = 0, which is not good. Could anybody please help me here.Thank you very much.

2022-03-01T23:23:57Z

BERT from scratch without self-supervised learning

https://discuss.huggingface.co/t/bert-from-scratch-without-self-supervised-learning/24397

Suppose one copies or creates the bert-base architecture, meaning the model layers themselves and not the training curriculum (MLM and NSP). Next suppose that one adds on a classifier head to the copied bert-base architecture thatconsists of a single linear layer to make predictions over the set of classes associated with a dataset. One then randomizes the model’s parameters and begins training this model on a labeled dataset using supervised learning only.Namely, with the preprocessed data (that includes positional embeddings), this data is passed all the way through the bert-base architecture and the linear classifier layer to produce prediction set over the class set,a loss is calculated, and the weights are updated via backpropagation and stochastic gradient descent.My question is, would this be a good idea? Is there anything about this approach (compared to the self-supervised followed by task specific training curriculum) that would prevent one from obtaining decent metrics on a test set?As I understand, the bert authors had a lot of unlabelled data but suppose one had an equivalent amount of labelled data for a particular domain (let’s say sentiment about movie reviews). Is there any reason why the above approach wouldproduce poor results?

2022-10-13T17:12:22Z

Cross Lingual Transfer Learning ( XNLI )

https://discuss.huggingface.co/t/cross-lingual-transfer-learning-xnli/24383

I was reading this paper on XNLI.And I wanted to understand what does TRANSLATE-TRAIN and TRANSLATE-TEST entail.I will write down what I understood.TRANSLATE-TRAIN: In this, we train N models. N stands for 15 languages. So we train 15 separate models for each language. How do we test this model? Should we run each of these 15 models per language and jot down the average accuracy under each language? For eg: We train 15 language models, then we test each of these 15 models on the English test set and then calculate the average accuracy. Does this sound right?I have been struggling with this baseline for so long.https://arxiv.org/abs/1911.02116image1671×994 262 KB

2022-10-13T11:58:02Z

XLSR-Wav2Vec2 with punctuation

https://discuss.huggingface.co/t/xlsr-wav2vec2-with-punctuation/5775

Hi,I’ve been trying to train XLSR-Wav2Vec2 to predict transcription + “relevant” punctuation (typically we don’t keep the punctuation).The idea was to get punctuation in an end-to-end manner as the audio sample gives us additional hints to differentiate between statements, questions and exclamations vs doing an additional post-processing.The goal is to be able to speak without saying “period”, “question mark”, etc… which is unnatural.Here are my main steps:I started from the transformers examplerun_common_voiceI use the CommonVoice English dataset as it’s easier to preprocess than other languagesI useunidecodeto preprocess the text which does a lot of smart changes → Málaga becomes Malaga, François becomes Francois, etcmy regex of chars to remove is"()[\]_+/=%|` (was tricky to create, the order here matters)I have a dict of resamplers (since they’re not all 16,000)I filter by durationNot sure if the wer metric should be adapted. Maybe I should add a separator between the punctuation but based on the way it’s calculated, I feel like it should decrease regardless.So far my training loss reduces (when using the full dataset it gets to nan probably due to some corrupted examples) but I keep a wer of 1. When testing a long run, I just get an empty output.To reproduce:clonethis repopython run_common_voice.py --dataset_config_name en --output_dir ./model --overwrite_output_dir --model_name_or_path facebook/wav2vec2-large-xlsr-53 --num_train_epochs 3 --per_device_train_batch_size 16 --evaluation_strategy epoch --fp16 --freeze_feature_extractor --group_by_length --gradient_checkpointing --do_train --do_eval --save_total_limit 1 --logging_steps 100 --warmup_steps 500 --load_best_model_at_end --metric_for_best_model wer --greater_is_better False --gradient_accumulation 2 --activation_dropout 0.055 --attention_dropout 0.094 --feat_proj_dropout 0.04 --hidden_dropout 0.047 --layerdrop 0.041 --learning_rate 0.000234 --mask_time_prob 0.082 --per_device_eval_batch_size 8Feel free to give any suggestions. I’ll update if I get more interesting results.

2021-04-26T14:07:49Z

How to train relation extraction?

https://discuss.huggingface.co/t/how-to-train-relation-extraction/24280

I a little bit confused, for example I want to fine tune a NER in bert english and I realize inJohn Snow Labhave a NLP task for relation extraction, My question how we can the train the relation after fine tune the NER? can we do it in hugginface and what transformer model use in relation extraction?

2022-10-11T15:33:14Z

Problem in understanding the test phase in Few-shot learning

https://discuss.huggingface.co/t/problem-in-understanding-the-test-phase-in-few-shot-learning/24231

I am studying few-shot learning ([1703.05175] Prototypical Networks for Few-shot Learning) and its source code. But the point is that the query set and the support sets in the test phase does not make sense to me. I am trying to understand why the test phase or validation phase uses label data for the query set. It is very different from classification or semi-supervised learning.When we train the encoder, we use l2-distance, a support set, and a query set to train the network. The samples in both sets are chosen based on their labels (k-way+n-shot setting). In the test phase, we have different labels from the train set. We choose the support and query sets the same way as the training phase without updating the encoder weights. But the query set is chosen based on label.I do not understand why in the test phase we use the query set and try to predict the label based on distance from just x-way. Should not we calculate the prototypes of each label and calculate the distance of query images with all labels in the test phase? It makes more sense than calculating the distance of query samples (without considering them based on their labels). Also, the same in the training phase. no distance among all labels’ prototypes.Again, testing based on support and query set does not make sense in the test phase.

2022-10-10T15:20:05Z

`nan` training loss but eval loss does improve over time

https://discuss.huggingface.co/t/nan-training-loss-but-eval-loss-does-improve-over-time/4521

I’ve been playing around with the XLSR-53 fine-tuning functionality but I keep gettingnantraining loss.Audio files I’m using are:Down-sampled to 16kHzSet to one channel onlyVary in length between 4 to 10sI’ve set the following hyper-params:attention_dropout=0.1hidden_dropout=0.1feat_proj_dropout=0.0mask_time_prob=0.05layerdrop=0.1learning rate:on a warmup schedule to3e-4for 3 epochsat5e-4for 3 epochsback to3e-4Sadly, I’m fine-tuning the model on an unpublished corpus, so I am probably not at liberty to upload it here which might hinder reproducibility efforts greatly.Here’s what the loss and WER progression looks like:image497×815 75.2 KBAnyone know what could be happening here? The model seems to be training just fine and some testing proves that the model performs well on the language I’m training it on. So what’s up with the training loss?Pinging@patrickvonplatenand@valhallaas this might be relevant to them.

2021-03-17T19:53:43Z

LayoutLM for extraction of information from tables

https://discuss.huggingface.co/t/layoutlm-for-extraction-of-information-from-tables/7464

Can the LayoutLM model be used or tuned for table detection and extraction?The paper says that it works on forms, receipts and for document classification tasks.

2021-06-27T15:43:27Z

Is there a way to split a news article into subtopic

https://discuss.huggingface.co/t/is-there-a-way-to-split-a-news-article-into-subtopic/23436

Hello, is there a way I can perform text segmentation on news articles?For example, a news article usually contains the main topic, but when reading through, there might probably be some subtopics present in the article. Is there a way I can divide those articles into those subsections/subtopics so that a news article can contain 2,3 or more sections depending on the subtopics discussed in that particular article.In case you are curious about what I need this for, I’m performing summarization on news articles, so instead of summarizing or parsing the whole article into the model at once, I want to divide them into sections based on what is discussed in the article and then summarize each section. Basically I’m trying to imitate what is done atsummari.comI will appreciate it if someone has done something like this before, or if anybody knows a way I can work through it.

2022-09-21T10:53:57Z

Best practices for estimating FLOPs-per-token with real datasets?

https://discuss.huggingface.co/t/best-practices-for-estimating-flops-per-token-with-real-datasets/23394

Hi folks,I’m currently reading the T-Few paper on few-shot learning and in section 4.2 they provide a table and estimate of the 11B parameter model’s inference costs as follows:We summarize the costs in table 1 and discuss them below. For all estimates, we use the median number of shots (41) across the datasets we consider. Rank evaluation and our unlikelihood loss both require processing every possible output choice to attain a prediction for an unlabeled example.The median combined tokenized sequence length for the input and all possible targets is 103 for the datasets we consider.…Processing a single input and all target choices with T-Few requires 11e9×103 = 1.1e12 FLOPs, whereas few-shot ICL with GPT-3 175B requires 2×175e9×(41 × 98 + 103) = 1.4e15 FLOPs – more than 3 orders of magnitude more.My question is: why is themedianinput sequence length used for the FLOPs estimate instead of themean?I understand that a dataset can have outliers in length, but I’m curious whether using the median is common practice.Thanks!

2022-09-20T12:40:53Z

Resources on interpretability of wav2vec-style speech models

https://discuss.huggingface.co/t/resources-on-interpretability-of-wav2vec-style-speech-models/23050

Hello everyoneBig thanks to HuggingFace for creating this amazing framework, and the active community as well! I’ve been using huggingface for a while now and been reading this forum as well.I am working on multi-lingual speech models and am interested in understanding how the pre-trained wav2vec-style models represent input utterances (from a phonetics perspective if possible). For example, I would like to know how Language Identification Model like “VoxLingua107 Wav2Vec Spoken Language Identification Model” goes about representing a collection of short utterances in English vs. say Thai.The most straight-forward method I know is to take final layer output embeddings (in inference mode) and to use t-SNE to cluster. But this doesn’t seem to help as much.I am looking for literature, codes, frameworks (like Captum) and tutorials which use wav2vec-style models and focus on interpretability. Please help. Thank you

2022-09-12T23:53:24Z

Keypoint Detection Accuracy is Very Low

https://discuss.huggingface.co/t/keypoint-detection-accuracy-is-very-low/22483

Unfortunately, I cannot say too much about my data set but I am trying to predict hundreds of keypoints/landmarks on a given image/video feed.I’m having great difficulty with my model architecture, I have not been able to get an accuracy greater than 20%.My model is based on similar ones I found via GitHub and a few academic papers, however they were all predicting dozens of points vs my hundreds. I only saw two model architectures across these sources:model = tf.keras.models.Sequential([ layers.Conv2D(32, (3,3), padding='same', input_shape=(512,512,1)), layers.LeakyReLU(), layers.MaxPool2D((2,2)), layers.Conv2D(64, (3,3), padding='same'), layers.LeakyReLU(), layers.MaxPool2D((2,2)), layers.Flatten(), layers.BatchNormalization(), layers.Dense(128), layers.ReLU(), layers.Dropout(0.5), layers.Dense(64), layers.ReLU(), layers.Dropout(0.5), layers.Dense(501) ])model = tf.keras.models.Sequential([ layers.Conv2D(32, (5,5), input_shape=(512,512,1), strides=1), layers.Conv2D(32, (3,3), strides=1), layers.MaxPool2D((2,2), padding="valid"), layers.BatchNormalization(), layers.Dropout(0.2), layers.Conv2D(64, (5,5), strides=2), layers.Conv2D(64, (5,5), strides=2), layers.AveragePooling2D((2,2), padding="valid"), layers.Flatten(), layers.Dense(128), layers.ReLU(), layers.Dropout(0.5), layers.Dense(501), layers.Softmax() ])My dataset is quite large, roughly 15K; this includes augmented data. Just looking for feedback.I’ve put this in the research section because I noticed that there’s very few models out there for keypoint detection; object detection seems to be much more popular.

2022-09-03T12:16:41Z

Abstractive summarization ensemble

https://discuss.huggingface.co/t/abstractive-summarization-ensemble/19987

Hi! I was wondering if anyone could point me to papers, blog posts, etc that explain how to ensemble previously trained models for abstractive text summarization (if possible).Moreover, is anything like this already implemented in Huggingface?

2022-07-04T21:13:13Z

Using OPTForSequenceClassification

https://discuss.huggingface.co/t/using-optforsequenceclassification/22340

Hi.I’m getting an error when trying to import the OPTForSequenceClassification class:ImportError: cannot import name ‘OPTForSequenceClassification’ from ‘transformers’ (/opt/conda/lib/python3.7/site-packages/transformers/init.py)Any heads up on why this might be the case? I saw the huggingface github already added this class.

2022-08-31T09:54:05Z

Zero shot classification for automated electrocardiogram reports

https://discuss.huggingface.co/t/zero-shot-classification-for-automated-electrocardiogram-reports/21594

HiI am a person from healthcare and new to this forum, I am doing research related to classifying automated electrocardiogram reports(ECG) with pre-defined labels, After reading about zero shot classification,I feel this can support my research,And i really want to know what type transformers will support for this work. And also the steps to be taken, Thank you

2022-08-13T17:53:25Z

Can you make Q&A language model stay on topic?

https://discuss.huggingface.co/t/can-you-make-q-a-language-model-stay-on-topic/21828

I’m thinking of fine-tuning a pre-trained language model for a Q&A task. More specifically, I’d like to fine-tune the model on a single chapter in a classic college textbook. Afterward, the reader of the chapter should be able to engage in a Q&A session with the model about the content of the chapter. But how do I make sure that the model stays on topic and doesn’t go out of a tangent? I know it is possible when looking at whathttps://play.aidungeon.io/has achieved, but I don’t know if it will require me to build a model from the ground for each chapter. Can anyone tell me if I’m out of my mind or if it’s feasible?Best,

2022-08-19T12:04:58Z

How to get embedding to each n-grams from a sentence using BERT?

https://discuss.huggingface.co/t/how-to-get-embedding-to-each-n-grams-from-a-sentence-using-bert/21562

Given a set of labels with different numbers of words, such as:labels=["computer accessories", "baby", "beauty and personal care"]Is there an approach to computing label embeddings in a single BERT forward pass (considering the list of labels as a single sentence)? Or has it the same as the computational cost of a forward pass for each label?

2022-08-12T14:44:01Z

Public Research Survey

https://discuss.huggingface.co/t/public-research-survey/20979

Hello, my name is Christian Flores. I am a recent UCSD graduate gathering public opinion on the impact of AI in day-to–day life for a research project. I would appreciate it if you could spare a few minutes to share your thoughts on the topic. The survey is estimated to take 5-10 minutes to complete. Your answers are anonymous. Thank you

2022-07-28T22:26:22Z

DeBerta Paper Explained and Dissected

https://discuss.huggingface.co/t/deberta-paper-explained-and-dissected/20158

Hello Everyone , Deberta has been ruling Kaggle competitions as well as Global Benchmarks recently , if you ever used it and wonder how it works internally , I have put together a Blog post for itjarvislabs.ai – 5 Jul 22DeBerta is the new King! | Jarvislabs.aiLearn about DeBerta architecture and find out how it outperforms the SOTA Bert and RoBerta.In this blog post I explain all the novel components that deberta introduces and all of it works together to create a performance boost .You will love it if you really love the transformers

2022-07-08T17:29:32Z

Summary Decoding params

https://discuss.huggingface.co/t/summary-decoding-params/19711

Hi,When applying decoding, it is common to provide decoding params , such asmin_length , max_length, beam_size , length penalty etc.I wonder if anyone is aware of a methodology or research for determining these params as well as if these could be dynamic and not hard-coded.I have found this paper, for multi-document summarization,aclanthology.orgD13-1069.pdf294.03 KBif anyone knows any additional resources - it would be highly appreciated.Thanks!

2022-06-28T07:45:13Z

Pre-Train BERT (from scratch)

https://discuss.huggingface.co/t/pre-train-bert-from-scratch/1245

BERT has been trained on MLM and NSP objective. I wanted to train BERT with/without NSP objective (with NSP in case suggested approach is different). I haven’t performed pre-training in full sense before. Can you please share how to obtain the data (crawl and tokenization details which were used) on which BERT was trained on ?. Since it takes a lot of time, I am looking for well tested code that can yield the BERT with/without NSP in one go. Any suggestions will be helpful.I know about some projects likethese, but they won’t integrate well withtransformerswell I guess which is a must have condition in my case.

2020-09-24T13:01:31Z

How to fine tune fine tune GitHub Copilot?

https://discuss.huggingface.co/t/how-to-fine-tune-fine-tune-github-copilot/18889

We can fine tune language models like BERT, GPT-3.Can I fine tune GitHub Copilot model?I have already looked into thehttps://copilot.github.com/but cant find the details. Would really appreciate if someone had fine tuned Github Copilot.

2022-06-09T04:26:50Z

Similarity search with combined image and text?

https://discuss.huggingface.co/t/similarity-search-with-combined-image-and-text/19168

How can I do similarity match by combining both image and text?Lets stay:Product1 = Image1, Text1Product2 = Image2, Text2I want to do contrastive learning by combining both the image and text.Is there such a model?Can anyone please suggest a model?

2022-06-14T23:36:21Z

LayoutLMv3 paper review and fine tuning code

https://discuss.huggingface.co/t/layoutlmv3-paper-review-and-fine-tuning-code/19495

LayoutLMv3: Pre-training for Document AI with Unified Text and Image MaskingHi guys,Made a small video going through Layout LMV3 paper.Feel free to check it out.

2022-06-23T09:13:40Z

Grouphug: multi-task, multi-dataset training with 🤗 transformers/datasets

https://discuss.huggingface.co/t/grouphug-multi-task-multi-dataset-training-with-transformers-datasets/19177

I recently releasedgrouphug- a package optimized for training on multiple datasets/dataframes at once, with each containing an arbitary subset of tasks, built ontransformers/datasets.The need for this came from wanting a single model to predict many closely related things like message topic, sentiment, toxicity, etc, with the inference speed of a single model, and better accuracy.I have also found that co-training on a masked language modelling task results in models which generalize very well and do not start overfitting.Even for single-task modelling, the classification head is also a good deal more powerful than the usual default, and the dataset formatter may be useful to quickly turn your dataframes into the format needed.Would love to hear if this is useful for anyone else, and any suggestions you have!

2022-06-15T07:26:19Z

LSTM Encoder-Decoder not working

https://discuss.huggingface.co/t/lstm-encoder-decoder-not-working/18697

I am trying to train an LSTM Encoder-Decoder model for paraphrase generation. My model is as follows:StackedResidualLSTM( (encoder): RecurrentEncoder( (embed_tokens): Embedding(30522, 256) (dropout): Dropout(p=0.5, inplace=False) (rnn): LSTM(256, 256, num_layers=2, batch_first=True, dropout=0.5) ) (decoder): RecurrentDecoder( (embed_tokens): Embedding(30522, 128) (dropout_in_module): Dropout(p=0.5, inplace=False) (dropout_out_module): Dropout(p=0.1, inplace=False) (layers): ModuleList( (0): LSTMCell(384, 256) (1): LSTMCell(256, 256) ) (fc_out): Linear(in_features=256, out_features=30522, bias=True) ) )Following is a print of the source sentence, the sentence fed to the decoder (shifted right), the predictions, and the true sentence (labels). Everything is tokenized with BERT tokenizer:Source: [CLS] where can i get quality services in brisbane for plasterand drywall repair? [SEP] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD][PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD][PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD][PAD] [PAD] [PAD] [PAD]Decoder Input: [CLS] [CLS] where can i getquality services for plaster and drywall repairs in brisbane? [SEP][PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD][PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD][PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD]Preds:[CLS] the? [SEP]? [SEP]? [SEP]? [SEP]? [SEP]? [SEP]? [SEP]? [SEP]?[SEP]Target: [CLS] where can i get quality services for plaster anddrywall repairs in brisbane? [SEP] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD][PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD][PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD] [PAD][PAD] [PAD] [PAD] [PAD] [PAD]My loss function is a CrossEntropy between the output and labels (the padding token is switched with -100 to ignore). Something like:loss_fct = CrossEntropyLoss() loss = loss_fct(logits.view(-1, logits.size(-1)), labels.view(-1))There are two problems occurring:the loss does not go downthe generations are all the same for every entry of the same epoch (after weight updating the generations might be different than the ones from the previous epoch, but remain the same for every entry of the new epoch)Do you have any idea what might I try to fix the issue? Thanks in advance for any help you can provide.

2022-06-03T17:15:48Z

Graph2graph network for geometric shapes

https://discuss.huggingface.co/t/graph2graph-network-for-geometric-shapes/18609

Similar to Seq2Seq models, are there graph2graph models available?Context: I am working on a dimension reduction problem on shapes, where,shapes are represented as graph,vertices as nodes,connecting curves as edges.dimension reduction operation is called as Midcurve generation.Input is 2D profile, say a closed polygon. Example: thick ‘L’ profile on left in the image below.Output is 1D curve in the middle of the profile. Example: thin ‘L’ curve on the right in the image belowencoderdecoder885×278 22.1 KBWish to build encoder-decoder network which accepts graphs as input as well as output.I have training set of such input and output graphs, a supervised set.As I could not find ready graph2graph network, I converted the problem to image2image (say, pix2pix like) and then solving it. But wish to investigate if graph2graph network is available or not.More info:Short paper:MidcurveNN: Encoder-Decoder Neural Network for Computing Midcurve of a Thin Polygon, viXra.org e-Print archive, viXra:1904.0429Github repo, source code:GitHub - yogeshhk/MidcurveNN: Computation of Midcurve of Thin Polygons using Neural NetworksHow to build such encoder decode network? Please note that as both, input and output are different, this can not be AutoEncoder.Any ideas?

2022-06-01T10:27:33Z

Steps to train T5 on collections of tags

https://discuss.huggingface.co/t/steps-to-train-t5-on-collections-of-tags/18600

Hiya! I’m working on my own model of Imagen; Instead of sentence prompts, my image-pair dataset uses a series of tags to describe an image. An example would be (without quotes): “sunny_day park dog parked_motorcycle female_walking” - and there can be anywhere from a few tags to 30+ tags per image. Because Imagen uses T5 to generate embeddings, I’d need to train a T5 model from scratch based on these collections of tags instead of using transfer learning, correct? Would these tags need to be presented as an array of strings, or one large string? What else would I need to do? And if it’s possible to answer: If my dataset was about 250K, how long would it take to train a T5 large on this dataset on either a P100 or latest generation TPU? Thanks for the help!

2022-06-01T07:58:21Z

Pegasus Paraphrase Fine Tuning dataset

https://discuss.huggingface.co/t/pegasus-paraphrase-fine-tuning-dataset/18531

Hi@tuner007. I was wondering if you could update the model card forpegasus_paraphraseto include the dataset that you use to finetune Google’s checkpoint?Alex

2022-05-30T16:01:51Z

Technical Skill classification model

https://discuss.huggingface.co/t/technical-skill-classification-model/18487

A raw dataset of over 30k data points is given which contains technical skills and a lot of jargonmixed in. We need to develop a code that can clean this dataset and extract Technical (Hard) skills. Some 900 random examples of technical skills are also given to go through them to understand thepattern and sequence. How should we go about this problem?

2022-05-29T17:34:15Z

Optuna with a fine-tuned model

https://discuss.huggingface.co/t/optuna-with-a-fine-tuned-model/18113

How can I use Optuna to optimize a fine-tuned model? Is there any example?

2022-05-18T17:58:01Z

Video Classification

https://discuss.huggingface.co/t/video-classification/17995

Hi everyone,I am starting to look into the task ofclassifying videos, trying to understand what approaches are currently available.Naively speaking, I guess one could randomly (maybe better, uniformly) sample N frames from a video, perform classification on each of them, and then aggregate predictions (most frequent prediction, most confident prediction, etc.). This may be reasonable for simple classification tasks (e.g. is there a cat in this video? Is the video set indoors or outdoors?).On the other hand, this approach would lose any temporal information conveyed by the frame sequence and the sound/speech information, for which a multi-modal model that can process sequences would be required.So I was wondering if any of you can point out examples of models that have been proposed/used for video classification in any of these directions.I tried browsing the HuggingFace directory but could not find a “video classification” task category, and I have the feeling (after some web searching) that this topic is generally less covered than image or text classification.Any pointer/suggestion is very much appreciated

2022-05-16T10:43:24Z

Ideas for scoring coding assignments

https://discuss.huggingface.co/t/ideas-for-scoring-coding-assignments/17862

Hey guys , I am searching for ways to use NLP to score coding/programming assignments like what a teacher will do in an exam/test. what ideas come to mind?any papers or similar problem solutions will be much appreciated !

2022-05-12T10:18:21Z

Dynamic Programming for Byte-level BPE

https://discuss.huggingface.co/t/dynamic-programming-for-byte-level-bpe/17376

Could anyone explain the rationale behind equation (1) inNeural Machine Translation with Byte-Level Subwords?Besides, what does it exactly mean byThe design of UTF-8 encoding ensures the uniqueness of this recovery process: for a character UTF-8 encoded with multiple bytes, its trailing bytes will not make a valid UTF-8 encoded character?How exactly are the hexadecimal digits being derived in Figure 1 ?

2022-05-01T02:58:55Z

Own AI deploy webapp

https://discuss.huggingface.co/t/own-ai-deploy-webapp/17247

Hello,I am a student at the higher technicas collage Leonding in Austria.In our collage we have our own servers with kubernetes that we are allowed to use as students.Me and three colleagues have a 2 year project where we have to program a webapp where you can easily deploy and train an AI with a few clicks.Basically we should be able to do the same asAutoTrainAutonlpfrom Hugging faceonly on the servers on our school servers.I was already on Github from Hugging Face myself and looked for open source repos but didn’t really find anything.Now the question is: Can I run AutoTrain and Autonlp on the school server or are there alternatives to create such a webapp.Thank you for your answers!

2022-04-27T11:40:47Z

Bert for audio classification

https://discuss.huggingface.co/t/bert-for-audio-classification/17179

I have been thinking at a very high abstract level about using Bert for something like audio classification. Suppose I have a time series data set of sampled sounds and their labels, something like an short audio clip of a dog barking that has the label “dog_bark”. I’m wondering if it’s possible to use the Bert architecture to perform this classification?Naively, I would say that one would have to pre-train Bert from scratch since the input data is time series data represented by floats. That would also lead me to think that one would have to also reconsider how they perform the token embeddings. I don’t have any super concrete ideas, but that was where I was starting. Curious if others had similar ideas or thoughts on the matter?EDIT: I am aware that there are other models out there better suited for this that perhaps fall into ASR or audio classification like wav2vec. However, in this instance I was specifically curious about adapting bert to the task.

2022-04-25T22:30:57Z

Confidence Scores / Self-Training for Wav2Vec2 / CTC models With LM (PyCTCDecode)

https://discuss.huggingface.co/t/confidence-scores-self-training-for-wav2vec2-ctc-models-with-lm-pyctcdecode/17052

I started looking a bit into Confidence Scores / Self-Training for Speech Recognition for models like Wav2Vec2 that make use a language model usingpyctcdecode'slibraryPyCTCDecode returns alm_scorewhich can be seen as the fused score between the acoustic model (Wav2Vec2) and a language model (kenLM). This score is the sum of all per-word fusedlm_scores, so it seems reasonable to normalize the output by the number of words. Also see some questions here:confidence scores output from the LM · Issue #57 · kensho-technologies/pyctcdecode · GitHubQuestion about naming of `lm_score` parameter in `decode_logits` · Issue #63 · kensho-technologies/pyctcdecode · GitHubFirst, let’s create some Wav2Vec2 + ngram models. We’ll simply add the official 4-gram of Librispeech to the new data2vec models to create the following models:patrickvonplaten/data2vec-audio-base-10m-4-grampatrickvonplaten/data2vec-audio-base-100h-4-grampatrickvonplaten/data2vec-audio-base-960h-4-gramNow, it’s quite easy to retrieve thoselm_scoresand to compute a confidence level this way:Import all necessary libraries and load model and tokenizerfrom transformers import AutoModelForCTC, AutoProcessor from datasets import load_dataset import datasets import torch import sys model_id = "TODO: fill in" model = AutoModelForCTC.from_pretrained(model_id) processor = AutoProcessor.from_pretrained(model_id)Load Librispeech dummy data:num_samples = 4 dataset = load_dataset("patrickvonplaten/librispeech_asr_dummy", "clean", split="validation") samples = dataset[:num_samples] audio_samples = [s["array"] for s in samples["audio"]] sampling_rate = set([s["sampling_rate"] for s in samples["audio"]]).pop() text_samples = samples["text"]Predict transcription with model:# process to input_values inputs = processor(audio_samples, return_tensors="pt", sampling_rate=sampling_rate, padding=True) # forward inputs to model with torch.no_grad(): logits = model(inputs.input_values, attention_mask=inputs.attention_mask).logitsRetrieve the per word probability normalized over number of wordsoutput = processor.batch_decode(logits.numpy(), output_word_offsets=True) confidence_scores = [score / len(t.split(" ")) for score, t in zip(output.lm_score, output.text)]Define confidence score the length normalizedlm_scoreof the predictionfor i in range(num_samples): print(20 * "=" + f"Output {i}" + 20 * "=") print(text_samples[i]) print(f"{output.text[i]}: {confidence_scores[i]}") print("\n")Cool let’s run this on the new data2vec audio models:patrickvonplaten/data2vec-audio-base-10m-4-grampatrickvonplaten/data2vec-audio-base-100h-4-grampatrickvonplaten/data2vec-audio-base-960h-4-grampatrickvonplaten/data2vec-audio-base-10m-4-gram====================Output 0==================== MISTER QUILTER IS THE APOSTLE OF THE MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL MISTER QUILTER IS THE APPOSELE OF THE MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL: -2.9550299660242825 ====================Output 1==================== NOR IS MISTER QUILTER'S MANNER LESS INTERESTING THAN HIS MATTER NOR IS MISTER QUILTR'S MANNER LESS INTERESTING THAN HIS MATTER: -3.8471058156146243 ====================Output 2==================== HE TELLS US THAT AT THIS FESTIVE SEASON OF THE YEAR WITH CHRISTMAS AND ROAST BEEF LOOMING BEFORE US SIMILES DRAWN FROM EATING AND ITS RESULTS OCCUR MOST READILY TO THE MIND HE TELLS IS THAT AT THIS FESTIVE SEASON OF THE YEAR WITH CRISMIIS AND ROST BEEF LOOMING BEFORE HIS SIMILES DRAWN FROM EATING AND ITS RESULTS OCCUR MOST READILY TO THE MIND: -3.115683062281252 ====================Output 3==================== HE HAS GRAVE DOUBTS WHETHER SIR FREDERICK LEIGHTON'S WORK IS REALLY GREEK AFTER ALL AND CAN DISCOVER IN IT BUT LITTLE OF ROCKY ITHACA HE HAS GRAVED DOUBTS WHETHER SIR FREDERICK LATEN'S WORK IS RELY GREEK AFTER ALL AND CAN DESCOVER IN IT BUT LITTLE OF ROCKY ETHICA: -4.292775884726897patrickvonplaten/data2vec-audio-base-100h-4-gram====================Output 0==================== MISTER QUILTER IS THE APOSTLE OF THE MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL MISTER QUILTER IS THE APOSTLE OF THE MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL: -1.0723093529710663 ====================Output 1==================== NOR IS MISTER QUILTER'S MANNER LESS INTERESTING THAN HIS MATTER NOR IS MISTER QUILTER'S MANNER LESS INTERESTING THAN HIS MATTER: -2.6140757339617786 ====================Output 2==================== HE TELLS US THAT AT THIS FESTIVE SEASON OF THE YEAR WITH CHRISTMAS AND ROAST BEEF LOOMING BEFORE US SIMILES DRAWN FROM EATING AND ITS RESULTS OCCUR MOST READILY TO THE MIND HE TELLS US THAT AT THIS FESTIVE SEASON OF THE YEAR WITH CHRISTMAS AND ROAST BEEF LOOMING BEFORE US SIMILES DRAWN FROM EATING AND ITS RESULTS OCCUR MOST READILY TO THE MIND: -1.1805021799946347 ====================Output 3==================== HE HAS GRAVE DOUBTS WHETHER SIR FREDERICK LEIGHTON'S WORK IS REALLY GREEK AFTER ALL AND CAN DISCOVER IN IT BUT LITTLE OF ROCKY ITHACA HE HAS GRAVE DOUBTS WHETHER SIR FREDERICK LAYTON'S WORK IS REALLY GREEK AFTER ALL AND CAN DISCOVER IN IT BUT LITTLE OF ROCKY ITHACA EH: -2.069009737832042patrickvonplaten/data2vec-audio-base-960h-4-gram====================Output 0==================== MISTER QUILTER IS THE APOSTLE OF THE MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL MISTER QUILTER IS THE APOSTLE OF THE MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL: -1.0610139720694658 ====================Output 1==================== NOR IS MISTER QUILTER'S MANNER LESS INTERESTING THAN HIS MATTER NOR IS MISTER QUILTER'S MANNER LESS INTERESTING THAN HIS MATTER R: -3.11299682252419 ====================Output 2==================== HE TELLS US THAT AT THIS FESTIVE SEASON OF THE YEAR WITH CHRISTMAS AND ROAST BEEF LOOMING BEFORE US SIMILES DRAWN FROM EATING AND ITS RESULTS OCCUR MOST READILY TO THE MIND HE TELLS US THAT AT THIS FESTIVE SEASON OF THE YEAR WITH CHRISTMAS AND ROAST BEEF LOOMING BEFORE US SIMILES DRAWN FROM EATING AND ITS RESULTS OCCUR MOST READILY TO THE MIND: -1.147767963941466 ====================Output 3==================== HE HAS GRAVE DOUBTS WHETHER SIR FREDERICK LEIGHTON'S WORK IS REALLY GREEK AFTER ALL AND CAN DISCOVER IN IT BUT LITTLE OF ROCKY ITHACA HE HAS GRAVE DOUBTS WHETHER SIR FREDERICK LEIGHTON'S WORK IS REALLY GREEK AFTER ALL AND CAN DISCOVER IN IT BUT LITTLE OF ROCKY ITHACA: -1.870571726475313Alright, this actually seems to make some sense here! The 10m has consistently the lowest score and one can usually say that the “correcter” the sentence the better the score. The 960h model has the best scores for all butOutput 1for which the 100h also gives a better prediction.This already seems to work quite well, but would need some more experiments.There are a couple of questions, I’m not sure about:Right now the average probability per word is taken, isminormaxmaybe better? Also see:confidence scores output from the LM · Issue #57 · kensho-technologies/pyctcdecode · GitHub

2022-04-21T11:13:34Z

Confidence Scores / Self-Training for Wav2Vec2 / CTC models

https://discuss.huggingface.co/t/confidence-scores-self-training-for-wav2vec2-ctc-models/17050

I started looking a bit into Confidence Scores / Self-Training for Speech Recognition for models like Wav2Vec2.The most reasonable way of doing so is to do it on a per-word level basis.With the newoutput_word_offsets=Trueit’s quite easy to retrieve the logits scores of the predicted words. E.g. one could do the following:Import all necessary libraries and load model and tokenizerfrom transformers import AutoModelForCTC, AutoProcessor from datasets import load_dataset import datasets import torch import sys model_id = "TODO: fill in" model = AutoModelForCTC.from_pretrained(model_id) processor = AutoProcessor.from_pretrained(model_id)Load Librispeech dummy data:num_samples = 4 dataset = load_dataset("patrickvonplaten/librispeech_asr_dummy", "clean", split="validation") samples = dataset[:num_samples] audio_samples = [s["array"] for s in samples["audio"]] sampling_rate = set([s["sampling_rate"] for s in samples["audio"]]).pop() text_samples = samples["text"]Predict transcription with model:# process to input_values inputs = processor(audio_samples, return_tensors="pt", sampling_rate=sampling_rate, padding=True) # forward inputs to model with torch.no_grad(): logits = model(inputs.input_values, attention_mask=inputs.attention_mask).logitsCompute probabilities (log softmax here) of predicted (argmax logits):pred_ids = torch.argmax(logits, dim=-1) scores = torch.nn.functional.log_softmax(logits, dim=-1) pred_scores = scores.gather(1, pred_ids.unsqueeze(-1))[:, :, 0]Retrieve the per word probability normalized over word lengthoutput = processor.batch_decode(pred_ids, output_word_offsets=True) # add confidence def confidence_score(word_dict, index): probs = pred_scores[index, word_dict["start_offset"]: word_dict["end_offset"]] return round(torch.sum(probs).item() / (len(probs)), 4) confidence_scores = [] for i in range(num_samples): confidence_scores.append({d["word"]: confidence_score(d, i) for d in output.word_offsets[i]})Define confidence score as minimum word probfor i in range(num_samples): print(20 * "=" + f"Output {i}" + 20 * "=") print(text_samples[i]) print(f"{' '.join(confidence_scores[i].keys())}: {min(confidence_scores[i].values())}") print("\n")Cool let’s run this on the new data2vec audio models:facebook/data2vec-audio-base-10m · Hugging Facefacebook/data2vec-audio-base-100h · Hugging Facefacebook/data2vec-audio-base-960h · Hugging FaceIt should be clear that the 960h should have “more” confidence than the 100h model. However the outputs are as follows:facebook/data2vec-audio-base-10m====================Output 0==================== MISTER QUILTER IS THE APOSTLE OF THE MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL MISTER QUILTER IS THE APPOSELE OF MIDL CLASES AND WHE ER GLAD TO WELCOME HIS GASPLE: -0.5873 ====================Output 1==================== NOR IS MISTER QUILTER'S MANNER LESS INTERESTING THAN HIS MATTER NOR IS MISTERE QUILTR'S MANER LES INTRESTING THAN HIS MATER: -0.4173 ====================Output 2==================== HE TELLS US THAT AT THIS FESTIVE SEASON OF THE YEAR WITH CHRISTMAS AND ROAST BEEF LOOMING BEFORE US SIMILES DRAWN FROM EATING AND ITS RESULTS OCCUR MOST READILY TO THE MIND HE TELES IS THAT AT THIS FESTIVE CESON OF THE YEAR WITH CRISMIIS AND ROST BEF LOOMING BEFOR SEIMILIYS DRAWN FROM EATING ITS RESALTS OCARE MOST REDHILY TO MIND: -0.0 ====================Output 3==================== HE HAS GRAVE DOUBTS WHETHER SIR FREDERICK LEIGHTON'S WORK IS REALLY GREEK AFTER ALL AND CAN DISCOVER IN IT BUT LITTLE OF ROCKY ITHACA HE HAS GREAVED DOUBTS WETHER SIR FREDRICK LATEN'S WORK IS RELY GRE AFTER ALL AND CAN DESCOVER IN IT BUT LITTLE OFE ROCKY ETHICA: -0.0006facebook/data2vec-audio-base-100h====================Output 0==================== MISTER QUILTER IS THE APOSTLE OF THE MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL MISTER QUILTER IS THE APOSTLE OF MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL: -0.7656 ====================Output 1==================== NOR IS MISTER QUILTER'S MANNER LESS INTERESTING THAN HIS MATTER NOR IS MISTER QUILTER'S MANNER LESS INTERESTING THAN HIS MATTER: -0.5057 ====================Output 2==================== HE TELLS US THAT AT THIS FESTIVE SEASON OF THE YEAR WITH CHRISTMAS AND ROAST BEEF LOOMING BEFORE US SIMILES DRAWN FROM EATING AND ITS RESULTS OCCUR MOST READILY TO THE MIND HE TELLS US THAT AT THIS FESTIVE SEASON OF THE YEAR WITH CHRISTMAS AND ROAST BEEF LOOMING BEFORE SIMILES DRAWN FROM EATING ITS RESULTS OCCUR MOST READILY TO MINE: -0.0 ====================Output 3==================== HE HAS GRAVE DOUBTS WHETHER SIR FREDERICK LEIGHTON'S WORK IS REALLY GREEK AFTER ALL AND CAN DISCOVER IN IT BUT LITTLE OF ROCKY ITHACA HE HAS GRAVE DOUBTS WHETHER SIR FREDERICK LAYTON'S WORK IS REALLY GREEK AFTER ALL AND CAN DISCOVER IN IT BUT LITTLE OF ROCKY ITHICA EH: -0.0facebook/data2vec-audio-base-960h====================Output 0==================== MISTER QUILTER IS THE APOSTLE OF THE MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL MISTER QUILTER IS THE APOSTLE OF MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL: -0.938 ====================Output 1==================== NOR IS MISTER QUILTER'S MANNER LESS INTERESTING THAN HIS MATTER NOR IS MISTER QUILTER'S MANNER LESS INTERESTING THAN HIS MATTER RR: -0.6415 ====================Output 2==================== HE TELLS US THAT AT THIS FESTIVE SEASON OF THE YEAR WITH CHRISTMAS AND ROAST BEEF LOOMING BEFORE US SIMILES DRAWN FROM EATING AND ITS RESULTS OCCUR MOST READILY TO THE MIND HE TELLS US THAT AT THIS FESTIVE SEASON OF THE YEAR WITH CHRISTMAS AND ROAST BEEF LOOMING BEFORE SIMILES DRAWN FROM EATING ITS RESULTS OCCUR MOST READILY TO MIND: 0.0 ====================Output 3==================== HE HAS GRAVE DOUBTS WHETHER SIR FREDERICK LEIGHTON'S WORK IS REALLY GREEK AFTER ALL AND CAN DISCOVER IN IT BUT LITTLE OF ROCKY ITHACA HE HAS GRAVE DOUBTS WHETHER SIR FREDERIC LEYHTON'S WORK IS REALLY GREEK AFTER ALL AND CAN DISCOVER IN IT BUT LITTLE OF ROCKY ITHACA: -0.0Now as can be seen that this doesn’t seem to be too useful. Incorrect text is predicted with very high confidence by the10mmodel and there is no difference between the 960h and the 10m model really at all nor between correctly and incorrectly predicted sentences.There are a couple of questions, I’m not sure about:Is it even possible to do confidence scoring without a language model for ASR?Should the minimum (lowest prob) of all words be taken as the confidence of the transcription or the average?Should the word prob correspond to a length normalized log_sum or not normalized?

2022-04-21T10:57:24Z

Text to Speech Alignment with Transformers

https://discuss.huggingface.co/t/text-to-speech-alignment-with-transformers/16166

Hi there,I have a large dataset of transcripts (without timestamps) and corresponding audio files (avg length of one hour). My goal is to temporally align the transcripts with the corresponding audio files.Can anyone point me to resources, e.g., tutorials or huggingface models, that may help with the task? Are there any best practices for how to do it (without building an entire system from scratch)?My initial naive idea was to use a STT model to transcribe the audio (while documenting timestamps) and then performing some kind of similarity search with the transcript to align the two. However, I feel this approach might be quite error prone.I am happy for any kind of help/pointer.Simon

2022-03-28T14:00:56Z

Projected gradient descent on autoregressive models

https://discuss.huggingface.co/t/projected-gradient-descent-on-autoregressive-models/16975

I am doing text summarization along with a trained classifier (that gives a label to a outputted summarization), and I would like to find how far away certain classifier labels are from each other by using some adversarial attacks and visualizing it for summarizer’s encoder embeddings. Is there any part of the huggingface library focusing on doing i.e. projected gradient descent on (autoregressive) decoder to encoder embedding?

2022-04-19T14:22:44Z

Compute metric on Dev

https://discuss.huggingface.co/t/compute-metric-on-dev/16836

Hello, I was wondering why notebooks compute blue or rouge on dev data, not on test data? like thisnotebookdev1398×298 8.97 KB

2022-04-14T14:52:20Z

Text similarity not by cosine similarity

https://discuss.huggingface.co/t/text-similarity-not-by-cosine-similarity/8766

Hi all,I have a question.I have a dataset containing questions and answers from a specific domain. My goal is to find the find the X most similar questions to a query.for example:user: “What is python?”dataset questions: [“What is python?”, “What does python means?”, “Is it python?”, “Is it a python snake?”, “Is it a python?”]I tried encoding the questions to embeddings and calculate the cosine similarity but the problem is it gives me high similarity score for “Is it python?” for the query “What is python?” which is clearly not the same question meaning and for “What does python means?” get very low score compared to “Is it python?”Any suggestions how i can overcome this problem? maybe new approaches…

2021-07-28T13:06:30Z

Aggregate encoder states in encoder-decoder models for long sequences?

https://discuss.huggingface.co/t/aggregate-encoder-states-in-encoder-decoder-models-for-long-sequences/16625

Hi. I would like to train a text-to-text QA model for long documents.I was wondering if anyone has seen success in aggregating the encoder states of a long document in any way (e.g. pooling) before passing it to the decoder, similar to the sliding window technique done for e.g. classification with BERT. I’m well aware of models like the longformer, etc, but just wondering if this approach has any utility, and if not, why not?

2022-04-08T18:39:37Z

Why do the commit histories of Hugging Face's datasets and models appear recent? Weren't these datasets and models uploaded a while ago?

https://discuss.huggingface.co/t/why-do-the-commit-histories-of-hugging-faces-datasets-and-models-appear-recent-werent-these-datasets-and-models-uploaded-a-while-ago/16595

I’ve been going over Hugging Face for research purposes, we’ve been looking over several datasets and models. We started doing this sometime ago last year during September and October. However, recently I checked this again, and it seems like these commit histories have changed.For instance, we looked overgemback in October of last year, but it shows that its commit history started on Jan 25.I am using Hugging Face as a use case for research purposes, I want to publish this research eventually, however people will ask questions about this, so I was wondering if someone could offer an explanation for this.

2022-04-07T18:32:42Z

Incorporating structural information in a Transformer?

https://discuss.huggingface.co/t/incorporating-structural-information-in-a-transformer/16554

For a Neural Machine Translation (NMT) task, my input data has relational information. This relation could be modelled using a graphical structure. Some researchers have tried to exploit transformer for graph data. For example, here is onepaper.I want to use Transformer. But then the challenge is how can I embed structural information there? Is there any open source artefact for Relational Transformer that I can use out of the box?

2022-04-06T19:50:25Z

Can you use both copy mechanism and BPE for a NMT task?

https://discuss.huggingface.co/t/can-you-use-both-copy-mechanism-and-bpe-for-a-nmt-task/16531

I read to alleviate the problem of Out of Vocabulary (OOV), there are two techniques:BPECopy mechanismIt appears to me they are two orthogonal approaches.Can we combine the two, i.e., we use both the copy mechanism and BPE? Are there any work out there that combines the two? I cant find any.

2022-04-06T11:44:44Z

Is there an easy way to apply layer-wise decaying learning rate in huggingface trainer for RobertaMaskedForLM?

https://discuss.huggingface.co/t/is-there-an-easy-way-to-apply-layer-wise-decaying-learning-rate-in-huggingface-trainer-for-robertamaskedforlm/1599

I am pre-training RobertaMaskedForLM on my own custom dataset. I wanted to implement the layer-wise learning rate decay given inhttps://github.com/aws-health-ai/multi_domain_lm#learning-rate-controlcorresponding to the paper -An Empirical Investigation Towards Efficient Multi-Domain LanguageModel Pre-training. Is there an easy way to incorporate this decay of learning rate with layer depth towards input usingtransformers.Trainer?

2020-10-17T09:31:45Z

The discussion is about entity recognition and corefrence resolution

https://discuss.huggingface.co/t/the-discussion-is-about-entity-recognition-and-corefrence-resolution/16068

Input = " I need 3 chairs in each class and there are 10 classes, so i need 30 chairs"output = “30 chairs, 10 classes”i have used the concept of corefrence resolution and entity recognition but i am unable to simplify the statement enough so that the input can become " i need 30 chairs and 10 classrooms" or similar linguistic approach which would help me to use entity recognition and solve this problem statement.

2022-03-25T10:54:24Z

GPT2 for QA Pair Generation

https://discuss.huggingface.co/t/gpt2-for-qa-pair-generation/759

I was wondering if it were possible to somehow train GPT2 to generate question-answer pairs in a particular domain?

2020-08-18T21:59:56Z

Converting Test Case Description into Test case Steps

https://discuss.huggingface.co/t/converting-test-case-description-into-test-case-steps/15332

Hello Everyone,I am looking for model or an approach which can help us converting a test case scenario description into testcase steps for Web Functional Testing. This will help an user to follow the instructions mentioned in the test steps in order to execute a testcase for the web.For example:Test Case:Verify that Feature creation is available within Portfolio Items on RallyTest Description: (Input)I login towww.rally.comand enter username and password and click on submit button so that it takes me to the Dashboard pageI click on Epic Delivery and select SkynetI click on Portfolio and select Portfolio Items so that Features table is displayedOutputNoTest StepsValidation Steps1Login towww.rally.com2Enter Username3Enter Password4Click on submit buttonDashboard Page is displayed5Click on Epic Delivery6Select Skynet7Click on Portfolio8Select Portfolio ItemsFeatures table is displayedPlease point me in some direction where I’ll be able to achieve some results.Thanks

2022-03-04T02:43:46Z

Best Pre-training Strategy

https://discuss.huggingface.co/t/best-pre-training-strategy/15307

Hey community, I hope you’re models are converging fastI’m trying to pre-train a BERT model on short query sentences/words, and i’m wondering what’s the best pre-training strategy to adapt in this situation?Thank in advance.

2022-03-03T12:09:18Z

Relative Position Representation/Encoding for Transformer

https://discuss.huggingface.co/t/relative-position-representation-encoding-for-transformer/15018

InGPT-NeoX-20B: An Open-Source Autoregressive Language Modelpaper, why didthe authorstated thatRotary embeddings are a form of static relative positional embeddings?InHow Self-Attention with Relative Position Representations works | by ___ | Medium, could anyone explain the rationale behindthe value of the lookup indices after the 3rd element are all 6?What is the actual purpose ofskewing mechanism?image711×801 102 KBimage1738×598 134 KB

2022-02-22T08:45:31Z

How find idea for academic thesis?

https://discuss.huggingface.co/t/how-find-idea-for-academic-thesis/14933

How can I find some idea regarding NLP tasks for graduate thesis?

2022-02-19T21:57:09Z

Extractive oracle

https://discuss.huggingface.co/t/extractive-oracle/14548

Is there any official script for an extractive oracle using huggingface’s implementation of ROUGE?An extractive oracle extracts from the source the N sentences that maximize ROUGE-2 (typically). For example,thisscript computes such extractive oracle. However, since they use a different implementation of ROUGE, this might not be completely in line with my other experiments (which use the HF implementation).Thanks!

2022-02-09T12:37:40Z

A Survey to Understand Challenges of Deploying Text Classification

https://discuss.huggingface.co/t/a-survey-to-understand-challenges-of-deploying-text-classification/14345

Hello everyone,As more and more machine learning libraries are developed, it becomes much easier to build a text classifier. However, there are still a lot of challenges, ranging from collecting the training data, achieving high accuracy, making the classification fair for different groups of users, defending against malicious input, etc. As a group of researchers from MIT, we are curious about what are the challenges for industrial practitioners currently having to deploy text classifiers. If you have experience in deploying text classifiers, I wish you can spend 15-20 minutes filling out this survey to help us understand the challenges. You will also enter a lottery for a $25 gift card.Link to the SurveyWhy should you participate?Have you ever encountered a problem when deploying a text classifier, and could not find a good solution? We believe that there are common problems in the deployment process, while some of them could have a better solution. Our research is to understand the actual challenges in the deployment of text classifiers, and to establish connections between these challenges and future academic research. We will summarize the results into a position paper to call on researchers’ attention to solving problems in the practical deployment.

2022-02-02T18:36:18Z

Question Answering model on mathematical domain for the greek language

https://discuss.huggingface.co/t/question-answering-model-on-mathematical-domain-for-the-greek-language/14300

Hello to everyone. I want to build a chatbot for my students in order to answer mathematical questions for the greek language. What I want to use is a question answering bert model or a sentence pair similarity. After trying various multilingual models pretrained on the question answering task on close domain i didnt have any luck, mainly because the text has specific mathematical terminology, complementary angles, supplementary angles e.t.c.I have found a greek language bert modelnlpaueb/bert-base-greek-uncased-v1 · Hugging Facewhich was trained on the greek part of wikipedia. Should I use this model and fine tune it on the greek part of Wikipedia on articles containing mathematical text and the training that model for the question answering task? And if this is the case does anyone know any question answering dataset for the greek language like the squad dataset? If my understanding is correct auto translate the squad dataset won’t give good results since after translation tha starting point of the answer may have changed. I would appreciate if someone could give me some quide lines to follow.

2022-02-01T13:33:40Z

Finetuning German BERT for QA on biomedical domain

https://discuss.huggingface.co/t/finetuning-german-bert-for-qa-on-biomedical-domain/500

Hello there and thank you very much for this wonderful work. I am relatively new to this field, so please bear with my amateur question. I want to perform question-answering on a German Biomedical text. From what I understand up to now, I need to fine-tune German BERT on biomedical QA datasets. Is there any script/pipeline that I should be using for this?Thank you very much in advance.

2020-07-28T09:01:21Z

[Suggestions and Guidance]Finetuning Bert models for Next word Prediction

https://discuss.huggingface.co/t/suggestions-and-guidance-finetuning-bert-models-for-next-word-prediction/14043

Problem Statement : To produce a next word prediction model on legal text. The aim is to build an autocomplete model which will make use of existing typed text as well as a possible concatenation of vectors from prior clauses/paragraphs.Current Approach: Because Bert based model are based on masked language, pretrained models such asLegalBertdid not produce good accuracy for prediction of next word when the word to be predicted was marked as [MASK]. Here is an example sentence, “use of [MASK]” where “marked” is the next word to be predicted in place of “[MASK]” token. (Note that there would not be words present after the mask token, only before the token).Currently approaching the problem as a SequenceClassification problem where labels are the token ids of the words that are to be predicted next. Will also attempt to finetune gpt2 on the legal text using run_clm.py from huggingface examples directoryIs there a better way to approach this problem of next word prediction?Any suggestions and guidance would be welcome.Thank you in advance

2022-01-24T11:15:47Z

Suggestions for an open source tagging tool to build custom LayoutLMv2 datasets

https://discuss.huggingface.co/t/suggestions-for-an-open-source-tagging-tool-to-build-custom-layoutlmv2-datasets/14103

Any suggestions on an Open Source tagging tool to get data in the format expected by the LayoutLMv2 model? I take it the standard format is similar to theFUNSD dataset.

2022-01-25T16:36:55Z

Paper Notes: Deepspeed Mixture of Experts

https://discuss.huggingface.co/t/paper-notes-deepspeed-mixture-of-experts/13908

SummaryThe legends over at DeepSpeed released apaperon scaling Mixture of Experts with a bunch of cool ideas.Since they will probably release some pytorch code soon I wanted to summarize/discuss the findings so that I learn them better.I provide 0 background on Mixture of Experts, assume knowledge of Top1 vs Top2 gating, for selfish/lazy reasons. Read thedeepspeed blog postfor background.I abstract the term “acc” to encompass all types of metrics: validation perplexity, zero shot accuracy, etc.I used@srushtrick of trying to read critically (to get your brain to think harder about other peoples’ results) but I don’t want to come off as too negative. I really enjoyed this paper and am excited to read the code!The DeepSpeed team proposes:(a) (sec 4.1) architectural modifications that reduce the number of experts without hurting acc.(b) (sec 4.1) Moe 2 Moe distillation, (instead of MoE 2 dense distillation like the FAIR paper (appendix Table 9) and theSwitch paper)(c) (sec 5) Systems Optimizations to make inference fastImproved Communication Collectives for MoE Inference (hierarchical all2all)tutelstyle single-device kernels to make routing tokens to experts fast.4D parallelism!?I now cover architecture and distillation, and save systems optimizations for later because I don’t fully understand them yet.Architecture: Pyramid Residual MoEThis section is really well written. It contains two very nice ablations that motivated the changes:Phenomenon 1: “Pyramid”We compare the performance of two different half-MoE architectures. More specifically, we put MoE layers in the first half of the model and leave the second half’s layers identical to the dense model. We switch the MoE layers to the second half and use dense at the first half.The results show that deeper layers benefit more from large number of experts.This also saves a ton of parameters: 40% reduction at 1.3B dense equivalent size, which will be useful at inference time.Phenomenon 2: “Residual”we can achieve the benefit of using two experts per layer but still use one communication.They frame this as trying to get the benefits of top2 routing without the costs.But, basically MoeLayers become only half sparse – a dense ffn that process the input as does 1 expert – the results are added.Compared to top2 where 2 different sparse experts process the input, this is cheaper because there is less communication (you only need to send the input to 1 place instead of 2?)Note this does not improve acc compared to top2, just speed.Putting it all together:FAIR arch (see table 1) (52B Params)Layers: top2 gating (each token gets routed to 2 experts)512 experts at each MoE layerDeepspeed Arch: (31B params)Layers: each token processed by dense FFN and 1 expert (same FLOPs as top2 gating if same number of experts, I believe).pyramid: somewhere between 32 and 128 experts at each Moe layer – way fewer params!In terms of acc, (PIQA is the only overlapping evaluation),the 31B Deepspeed performs between the FAIR 52B and the FAIR 207B and was probably lower training cost than the 52B, even before all the systems optimizations in section 5. Nice!With the systems optimizations they say training is 5x faster than dense (to the same acc). The FAIR paper says “4x faster than dense”, but measures TFLOPS, which make the extra communication required for MoE appear to be free. So all in all this definitely seems like a better architecture.It would have been cool if Tables 2,4 had training cost and inference cost next to the few shot performances (or 1 big joined table somewhere!).Staged Knowledge Distillation: Mixture Of Students (MoS)Caveat before you read this section: in most distillation results, the student model is MUCH smaller than the teacher model, like half as large or so. Here, the student model is only 12.5% smaller than the teacher model. (3 fewer layers, 4B fewer params (31B vs 27B)).They are able to lose very little performance, which is nice, but they also didn’t really lose that much weight, and it would be interesting to try to replicate what they did with smaller students.Caveat 2: name deeply misleading. It’s normal KD but they switch to cross entropy loss halfway through that’s it!Anyways, these are the first published MoE 2 MoE Distillation results. The switch paper and FAIR paper both distill Moe 2 Dense models (since they are much easier to serve than MoE models, a gap deepspeed claims to eliminate in section 5 – the one I don’t understand yet:( ).They use the same KD loss as the other papers, but they turn it off halfway through training.They say this improves acc, but I am most interested in the speed implications. I tried MoE2MoE distillation but it was extremely slow (like 10x slower than Dense2Dense) because of teacher inference every step.If we could only run the teacher forward pass for part of the student training, that would be sweet!NextLet me know any inaccuracies, important omissions, what you ate for lunch follow up ideas!Next week I will try to tackle Section 5 (Systems optimizations) and if I don’t I will burn a 20 dollar bill and record it!

2022-01-19T21:19:55Z

Using mixup on RoBERTa

https://discuss.huggingface.co/t/using-mixup-on-roberta/306

Hello everyone!I tried to apply the technique of data augmentation, mixup, popularly used on computer vision, but in this case on NLP.The algorithm developed is in two phases:The first phase gets the representation for each sentence of the batch, computing the mean of the correspondent hidden states of the last layer. The fragment below shows the corresponding module.class LanguageModel(nn.Module): def __init__(self, pretrained_model_name, device="cuda:0", anonymized_tokens=False): super(LanguageModel, self).__init__() # Load tokenizer self.tokenizer = AutoTokenizer.from_pretrained(pretrained_model_name) # Load model self.config = AutoConfig.from_pretrained(pretrained_model_name) self.config.output_hidden_states = True self.model = AutoModel.from_pretrained(pretrained_model_name, config=self.config).to(device) def forward(self, input_ids, attention_mask): outputs = self.model( input_ids=input_ids, attention_mask=attention_mask, ) activations = torch.mean(outputs[0], axis=1) return activationsAfter that, it generates the mixup examples using the function proposed on theoriginal code, but being the input, the representations computed on the previous step, instead the images, like originally.One time the mixup examples are generated, the second phase realizes the predictions (the fragment below shows the corresponding module). Finally, is computed the loss, in the same way as in the original work.class ClassifierLayer(nn.Module): def __init__(self, num_classes, dropout_rate=0.1, petrained_size=768, device="cuda:0"): super(ClassifierLayer, self).__init__() self.layer = nn.Linear(petrained_size, num_classes, bias=True).to(device) self.drop = nn.Dropout(dropout_rate) def forward(self, z): activations = self.layer(self.drop(z)) return activationsIn the fragment of the code below, is shown a summary of the training loop proposed, however the full script used ishere:for idx_epoch in range(0, args.num_train_epochs): language_model.train() classifier_layer.train() accs = 0; ps = 0; rs = 0; f1s = 0; lss = 0 for (idx_batch, train_batch) in enumerate(train_dataloader): # 0: input_ids, 1: attention_mask, 2:token_type_ids, 3: labels batch_train = tuple(data_.to(device) for data_ in train_batch) labels_train = batch_train[-1] inputs = { 'input_ids': batch_train[0], 'attention_mask': batch_train[1], } optimizer.zero_grad() # 1st phase: conextual embeddings contextual_embeddings = language_model( input_ids=inputs['input_ids'], attention_mask=inputs['attention_mask'], ) # 2nd phase: mixup inputs, targets_a, targets_b, lam = mixup_data(contextual_embeddings, labels_train, args.alpha_mixup, use_cuda) inputs, targets_a, targets_b = map(Variable, (inputs, targets_a, targets_b)) predictions = classifier_layer(inputs) loss = mixup_criterion(criterion, predictions, targets_a, targets_b, lam) # 2nd phase: standard # predictions = classifier_layer(contextual_embeddings) # loss = criterion(predictions, labels_train) lss += loss loss.backward() optimizer.step() scheduler.step()Experimenting with this approach, the results obtained are very poor…Have any of you worked on an approximation similar to this one with successful/good results?Thanks.

2020-07-15T13:44:46Z

How does the vocabulary size count towards total parameter size of a model?

https://discuss.huggingface.co/t/how-does-the-vocabulary-size-count-towards-total-parameter-size-of-a-model/13833

TL;DR The vocabulary size changes the number of parameters of the model. If we were to compare models with different vocabulary sizes, what would be the most fair strategy, fixing the total number of parameters or having the same architecture with same number of layers, attention heads, etc.?We have a set of mini models which are pretrained from scratch using the Roberta architecture. The number of layers, hidden sizes, and number of attention heads correspond to that of the mini models in the BERT paper. We wanted to experiment with the effect of different tokenization algorithms on the downstream performance and to this end, fit BPE, WordPiece, and WordLevel tokenizers with 50K, 50K, 100K vocabulary sizes respectively in addition to character-based tokenization. The reason for the increased vocabulary size for the WordLevel tokenization is to decrease the number of OOV tokens.Later did we notice that the difference between vocabulary sizes cause a huge difference between the number of parameters. The model sizes are 20.4M, 20.4M, 33.2M, and 8.1M for BPE, WordPiece, WordLevel, and char tokenizer-based models respectively. This means that the percentages of the number of parameters coming from the vocabulary of the model are 63%, 63%, 77%, and 1% for BPE, WordPiece, WordLevel, and char tokenizer-based models respectively.My question is, is it unfair to compare the downstream performance of these models on the same task with the same dataset just because the number of parameters are different. I would assume that in a given forward pass through an input, only a very small part of the vocabulary is updated. Because, a parameter in a layer of the transformer blocks of the model is updated every step, whereas a parameter in the vocabulary is updated whenever it appears in the input text. Therefore, to say that, for example, 100K parameters from the vocabulary of WordLevel tokenizer-based model contribute to the computation of an input is not true. This means that for as long as the number of parameters in the transformer blocks of the models are comparable, it is fair to compare the performance of the models. If this assumption is incorrect, I would be happy to be corrected.Thanks for your time.

2022-01-18T07:28:48Z

Guide: The best way to calculate the perplexity of fixed-length models

https://discuss.huggingface.co/t/guide-the-best-way-to-calculate-the-perplexity-of-fixed-length-models/193

Hey all. Just thought you might be interested in a page I just added to the research docs on theperplexity of fixed-length models.Perplexity (PPL) is defined as the exponential average of a sequence’s negative log likelihoods. For at-length sequenceX, this is defined,\text{PPL}(X) = \exp \left\{ -\frac{1}{t} \sum_i^t \log p_\theta (x_i|x_{<i}) \right\}But with fixed-length models (like most transformers), we can’t always condition on the entire preceding subsequence when predicting each token.The initial instinct for many in dealing with this problem is to break the whole sequence into segments equal to the model’s max input size and calculate the likelihoods of each segment independently. This not the best approach, however, since it gives the model very little context to use for prediction at the beginning of each segment. I’ll illustrate this with the following gif where we imagine a model with a max input size of 6 adding up the log-likelihoods for the sentence, “Hugging Face is a startup based in New York City and Paris”ppl_chunked1200×160 352 KBWhen the model starts the second segment, it has to try to predict the word “in” without any context, even though we have 5 words before it that the model could be using (since we said the max input size is 6).A better approach is to instead employ asliding windowstrategy, where you continually move the context across the sequence, allowing the model to take advantage of the available context.ppl_sliding1200×160 373 KBThis is slower to compute, but will typically yield better scores and is actually much closer to the way the sequence probabilities are formally decomposed (e.g. see the the equation above).In theguide, we show how to do this in a strided way with GPT-2. When using the first, naive approach, GPT-2 gets a PPL of19.64on WikiText-2. In contrast, when we use a strided sliding window, this score improves dramatically down to16.53.

2020-07-10T17:07:49Z

Few shot automatic moderation

https://discuss.huggingface.co/t/few-shot-automatic-moderation/12102

the facebook flies shows that there’s a lack of human based moderation on social networks. what about automatic moderation and how it cope with reduced datasets availability?I’m wondering what’s the actual research status on this subject.

2021-11-20T14:59:40Z

Let's Make an Ethics Chat Bot that's Not Racist!

https://discuss.huggingface.co/t/lets-make-an-ethics-chat-bot-thats-not-racist/11905

I am a philosopher and I have studied Ethics for over 20 years (check me here j.mp/joshtedx). I am disheartened to see a few recent attempts to make Ethics AIs have not turned out well (racist ethics AIs - Google Search)This should not happen. I am quite certain I can make an Ethics AI Few shot, Q and A example or cloze or knowledge base that is not racist for you, if you can make the chat bot part. I can also easily correct the answers from any large NLP model.Let’s show the world not all Ethics AIs will end up racist!I suggest doing this as a not-for profit project that others could even then use in their chat bots to correct for unethical answers as an out-of-the-box solution!If anyone is interested please LMK! Let’s make something good for humanity!

2021-11-16T19:32:49Z

New Paper: Masked Autoencoders Are Scalable Vision Learners

https://discuss.huggingface.co/t/new-paper-masked-autoencoders-are-scalable-vision-learners/11673

(Meta-comment: I’m actually not sure which forum this would best fit into - seems like it would be useful to have a place where we can discuss new papers.)This new work by Kaiming He et al seems pretty interesting - they use a very simple setup for masking during pre-training a ViT and it looks like they get very good results across a variety of tasks.So far, I see an implementation bylucidrains.arXiv.orgMasked Autoencoders Are Scalable Vision LearnersThis paper shows that masked autoencoders (MAE) are scalable self-supervised learners for computer vision. Our MAE approach is simple: we mask random patches of the input image and reconstruct the missing pixels. It is based on two core designs....

2021-11-14T01:55:46Z

Improving performance of Wav2Vec2 fine tuning with word piece vocabulary

https://discuss.huggingface.co/t/improving-performance-of-wav2vec2-fine-tuning-with-word-piece-vocabulary/6292

Hello,I’m fine tuning XLSR-Wav2Vec2 on a 200+ hours of a speech in a language not in the original pertaining.The training progresses nicely, however when it reaches about 40 WER it starts to overfit (WER doesn’t progress much and train loss decreases while eval loss is going up).I’ve tried increasing some params of the SpecAugment, but it only helped a bit.I’ve noticed that using the Speechbrain lib implementation I’m getting a bit better results (on the expense of training stability) and was wondering if it is due to a larger vocabulary they use there. Does anyone tried to use a tokenizer with a vocabulary that contains subwords and words in addition to characters? I could’t find any experiment that uses it with Huggingface transformers W2V2.I see in the Wav2Vec 2 paper they say that:We expect performance gains by switching to a seq2seq architecture and aword piece vocabulary.https://arxiv.org/pdf/2006.11477.pdfAny suggestions on how to do that with Huggingface Transformers?P.S. my dataset is noisy and not super clean.Any help or suggestion will be very helpful.Samuel

2021-05-21T13:31:58Z

[Help needed] Extending Trainer for Meta learning

https://discuss.huggingface.co/t/help-needed-extending-trainer-for-meta-learning/635

I want to implement MAML with Glue dataset with transformers. In my case, query and support set will come from the same dataset. I’ve read some work in meta learning from HF team (Wolf et al., 18).Although I’ve implemented my training loop (withhigher) (open for other methods as well), I am still looking for a correct reference implementation of MAML or Reptile to confirm. Currently my code inherits fromTrainer. If anyone share a sample snippet that would perform MAML gradient updates, that’d be really helpful ?

2020-08-08T11:31:51Z

Detection Transformer (DETR) for text detection in documents

https://discuss.huggingface.co/t/detection-transformer-detr-for-text-detection-in-documents/10396

Hi,i do currently some experiments on text detection with a transformer based model.Do anyone have experience at this or recommendations ?My idea is to train the DetrForObjectDetection on the COCOText-v2 datasetCOCOText-v2i have tested some setups:pretrained facebook/resnet-50 with num_queries=2000 (a good value for a A4 document page)from scratch with efficentNet_b0 backbone from timm with backbone lr: 0.001 and lr: 0.01but in all cases the loss and train loss stuck at ~1.7 after ~35 epochs with 2 val steps per epochanother problem i have faiced is the COCOevaluator there seems to be a problem with numpy has no append at validation step:in COCOeval:problem:self.eval_imgs[iou_type].append(eval_imgs)one sample from my train dataloader looks like this:# pixel_values 1 example torch.Size([3, 640, 640]) # target for this example {'boxes': tensor([[0.0810, 0.8323, 0.1621, 0.1356], [0.3031, 0.3070, 0.0367, 0.0088], [0.5304, 0.3418, 0.0349, 0.0102]]), 'class_labels': tensor([0, 0, 0]), 'image_id': tensor([367969]), 'area': tensor([5295.0200, 103.8200, 105.6000]), 'iscrowd': tensor([0, 0, 0]), 'orig_size': tensor([640, 556]), 'size': tensor([640, 556])}so the data after Dataloader seems to be oksome more code:COCO_stuff:adapted from:COCOTextPytorch COCODataloaderdef collate_fn(batch): """ process on every sample in batch """ feature_extractor = DetrFeatureExtractor() pixel_values = [item[0] for item in batch] encoding = feature_extractor.pad_and_create_pixel_mask(pixel_values, return_tensors="pt") labels = [item[1] for item in batch] batch = dict() batch['pixel_values'] = encoding['pixel_values'] batch['pixel_mask'] = encoding['pixel_mask'] batch['labels'] = labels return batch class CocoTextDataset(Dataset): """MSCOCO Text V2 Dataset """ def __init__(self, path, ann_file_name, image_folder_name, feature_extractor, is_train=True, data_limit=None): self.path = path self.annotation_path = os.path.join(path, ann_file_name) self.image_folder_path = os.path.join(path, image_folder_name) self.feature_extractor = feature_extractor self.data_limit = data_limit self.dataset_length = 0 self.coco_text = COCO_Text(annotation_file=self.annotation_path) if is_train: print('Load Training Data') self.set_part = self.coco_text.train else: print('Load Validation Data') self.set_part = self.coco_text.val # create sets for train and validation self.cleaned_img_to_ann_ids = {k:v for k,v in self.coco_text.imgToAnns.items() if v and k in self.set_part} # sort out images and annotations, which are not readable or have uncorrect bound boxes self.ann_ids = list() self.image_ids = list() for entry_id in self.cleaned_img_to_ann_ids.values(): annotations = self.coco_text.loadAnns(entry_id) allowed_ann_ids = list() allowed_image_ids = list() for annotation in annotations: if annotation['legibility'] == 'legible' and len(annotation['bbox']) == 4: allowed_ann_ids.append(annotation['id']) if annotation['image_id'] not in allowed_image_ids: allowed_image_ids.append(annotation['image_id']) # if image has no annotations, skip it if allowed_image_ids and allowed_ann_ids: self.image_ids.append(allowed_image_ids) self.ann_ids.append(allowed_ann_ids) if self.data_limit: self.image_ids = self.image_ids[0:data_limit] self.ann_ids = self.ann_ids[0:data_limit] self.image_info = list() self.ann_info = list() for id in self.image_ids: info = self.coco_text.loadImgs(id) self.image_info.append(info) for id in self.ann_ids: info = self.coco_text.loadAnns(id) self.ann_info.append(info) if len(self.image_info) == len(self.ann_info): print('Dataset created sucessfully') self.dataset_length = len(self.image_info) else: print(f'Error: Number of images and annotations do not match. {len(self.image_info)} images and {len(self.ann_info)} annotations') sys.exit(0) def __len__(self): return self.dataset_length def __getitem__(self, index): image_id = self.image_ids[index] image_file = self.image_info[index] annotations = self.ann_info[index] image_path = os.path.join(self.image_folder_path, image_file[0]['file_name']) image = Image.open(image_path).convert("RGB") target = {'image_id': image_id[0], 'annotations': annotations} encoding = self.feature_extractor(images=image, annotations=target, return_tensors="pt") pixel_values = encoding["pixel_values"].squeeze() # remove batch dimension target = encoding["labels"][0] # remove batch dimension return pixel_values, target class COCODatasetLoader(pl.LightningDataModule): def __init__(self, path, ann_file_name, image_folder_name, feature_extractor, batch_size, worker, collator, data_limit=None): super().__init__() self.path = path self.ann_file_name = ann_file_name self.image_folder_name = image_folder_name self.feature_extractor = feature_extractor self.batch_size = batch_size self.worker = worker self.collator = collator self.data_limit = data_limit print(f'Data Limit is set to : {self.data_limit}') def setup(self, stage=None): self.train_dataset = CocoTextDataset(self.path, self.ann_file_name, self.image_folder_name, self.feature_extractor, is_train=True, data_limit=self.data_limit) print(f'# of training samples: {self.train_dataset.dataset_length}') self.val_dataset = CocoTextDataset(self.path, self.ann_file_name, self.image_folder_name, self.feature_extractor, is_train=False, data_limit=self.data_limit) print(f'# of validation samples: {self.val_dataset.dataset_length}') def visualize_example(self, index): print(f'Visualize Example: {index}') file_name = self.train_dataset.coco_text.loadImgs(self.train_dataset.image_ids[index])[0]['file_name'] path = os.path.join(self.train_dataset.image_folder_path, file_name) annotations = self.train_dataset.coco_text.loadAnns(self.train_dataset.ann_ids[index]) print(f'{len(annotations)} boxes in image detected') image = Image.open(path).convert("RGB") draw = ImageDraw.Draw(image, "RGBA") for annotation in annotations: box = annotation['bbox'] x,y,w,h = tuple(box) draw.rectangle((x,y,x+w,y+h), outline='red', width=1) image.show() def get_val_coco_text_dataset(self): return self.val_dataset.coco_text def train_dataloader(self): return DataLoader(self.train_dataset, batch_size=self.batch_size, shuffle=False, num_workers=self.worker, pin_memory=True, collate_fn=self.collator) def val_dataloader(self): return DataLoader(self.val_dataset, batch_size=self.batch_size, shuffle=False, num_workers=self.worker, pin_memory=True, collate_fn=self.collator)Model:class TextDetectionModel(pl.LightningModule): def __init__(self, lr, id2label, feature_extractor, coco_evaluator, sync): super().__init__() self.save_hyperparameters() self.sync_dist = sync self.lr = lr self.id2label = id2label self.feature_extractor = feature_extractor self.coco_evaluator = coco_evaluator self.num_classes = len(id2label) self.model = DetrForObjectDetection.from_pretrained("facebook/detr-resnet-50", num_queries=2000, encoder_layerdrop=0.2, decoder_layerdrop=0.2, num_labels=self.num_classes, ignore_mismatched_sizes=True, return_dict=True) def forward(self, pixel_values, pixel_mask=None, labels=None): outputs = self.model(pixel_values=pixel_values, pixel_mask=pixel_mask, labels=labels, return_dict=True) return outputs.loss, outputs.loss_dict, outputs.logits, outputs.pred_boxes def training_step(self, batch, batch_idx): pixel_values = batch["pixel_values"] pixel_mask = batch["pixel_mask"] labels = [{k: v.to(self.device) for k, v in t.items()} for t in batch["labels"]] outputs = self.model(pixel_values=pixel_values, pixel_mask=pixel_mask, labels=labels) loss = outputs[0] loss_dict = outputs[1] self.log("train_loss", loss.detach(), prog_bar=True, on_step=False, on_epoch=True, sync_dist=self.sync_dist) for k,v in loss_dict.items(): self.log("train_" + k, v.item()) return loss def validation_step(self, batch, batch_idx): pixel_values = batch["pixel_values"] pixel_mask = batch["pixel_mask"] labels = [{k: v.to(self.device) for k, v in t.items()} for t in batch["labels"]] bboxes = [entry['boxes'] for entry in labels] outputs = self.model(pixel_values=pixel_values, pixel_mask=pixel_mask, labels=labels) loss = outputs[0] loss_dict = outputs[1] logits = outputs[2] # pred_boxes = outputs[3] # compute averaged probability of each bbox proba = torch.stack([x for x in logits.softmax(-1)[0, :, :-1]]).mean() # compute COCO Output for each image # orig_target_sizes = torch.stack([target["orig_size"] for target in labels], dim=0) # results = self.feature_extractor.post_process(outputs, orig_target_sizes) # convert outputs of model to COCO api # res = {target['image_id'].item(): output for target, output in zip(labels, results)} # Coco Eval is broken currently # self.coco_evaluator.update(res) self.log("val_loss", loss.detach(), prog_bar=True, on_step=False, on_epoch=True, sync_dist=self.sync_dist) self.log("val_bbox_proba", proba.detach(), prog_bar=True, on_step=False, on_epoch=True, sync_dist=self.sync_dist) for k,v in loss_dict.items(): self.log("val_" + k, v.item()) return loss #def validation_epoch_end(self, outputs): # self.coco_evaluator.synchronize_between_processes() # self.coco_evaluator.accumulate() # self.coco_evaluator.summarize() def predict_step(self, batch, batch_idx): pixel_values = batch["pixel_values"] outputs = self.model(pixel_values=pixel_values) logits = outputs[2] pred_boxes = outputs[3] probas = logits.softmax(-1)[0, :, :-1] return {'probas': probas, 'pred_boxes': pred_boxes} def configure_optimizers(self): param_dicts = [ {"params": [p for n, p in self.named_parameters() if "backbone" not in n and p.requires_grad]}, { "params": [p for n, p in self.named_parameters() if "backbone" in n and p.requires_grad], "lr": 1e-5, # this lr is used for backbone parameters }, ] optimizer = AdamW(param_dicts, lr=self.lr, weight_decay=1e-4) scheduler = ReduceLROnPlateau(optimizer, patience=2, verbose=True) return {'optimizer': optimizer, 'lr_scheduler': scheduler, 'monitor': 'val_loss'} def optimizer_zero_grad(self, epoch, batch_idx, optimizer, optimizer_idx): optimizer.zero_grad(set_to_none=True)Trainerimport argparse import os import warnings import time import numpy as np import onnx import pytorch_lightning as pl import torch from onnxruntime.quantization import quantize_qat from pytorch_lightning.callbacks import (EarlyStopping, LearningRateMonitor, ModelCheckpoint) from pytorch_lightning.loggers import TensorBoardLogger from transformers import DetrFeatureExtractor from coco_tools.coco_torch_evaluator import CocoEvaluator from dataloader import COCODatasetLoader, collate_fn from model import TextDetectionModel def __check_for_boolean_value(val): """argparse helper function """ if val.lower() == "true": return True else: return False if __name__ == '__main__': warnings.filterwarnings("ignore") pl.seed_everything(42, workers=True) print('annotations file and image folder have to be in the same parent folder') parser = argparse.ArgumentParser(description='Text Detection Trainer') parser.add_argument("--path", help='path to generated images', type=str, required=False, default='/COCOText-v2') #set to true parser.add_argument("--ann_file_name", help='name of annotations file', type=str, required=False, default='cocotext.v2.json') parser.add_argument("--image_folder_name", help='name of image folder', type=str, required=False, default='train2014') parser.add_argument("--epochs", help='how many epochs to train the model',type=int, required=False, default=250) parser.add_argument("--batch_size", help='how big are a batch',type=int, required=False, default=8) parser.add_argument("--data_limit", help='set a fixed data limit',type=int, required=False, default=0) parser.add_argument("--worker", help='how many threads for the Dataloader',type=int, required=False, default=0) parser.add_argument("--learning_rate", help='the learning rate for the optimizer',type=float, required=False, default=1e-4) parser.add_argument("--gradient_clip", help='float for gradient clipping',type=float, required=False, default=0.1) parser.add_argument("--visualize_random_example", help='if true show an example from train set',type=__check_for_boolean_value, required=False, default=False) args = parser.parse_args() path = args.path ann_file_name = args.ann_file_name image_folder_name = args.image_folder_name epochs = args.epochs batch_size = args.batch_size data_limit = args.data_limit worker = args.worker learning_rate = args.learning_rate gradient_clip = args.gradient_clip visualize_random_example = args.visualize_random_example if data_limit == 0: data_limit = None # resource handling if torch.cuda.device_count() >= 1: batch_size = int(batch_size / torch.cuda.device_count()) accelerator = 'ddp' sync = True else: accelerator = None sync = False ### Data Part os.makedirs('text_detection_model_files', exist_ok=True) feature_extractor = DetrFeatureExtractor(format="coco_detection", do_resize=False, do_normalize=True, image_mean=[0.485, 0.456, 0.406], image_std=[0.229, 0.224, 0.225]) feat_extractor_to_save = DetrFeatureExtractor.from_pretrained("facebook/detr-resnet-50", do_resize=True, size=600) feat_extractor_to_save.save_pretrained('text_detection_model_files/transformer_model/') print('feature extractor saved succesful') data_module = COCODatasetLoader(path=path, ann_file_name=ann_file_name, image_folder_name=image_folder_name, feature_extractor=feature_extractor, batch_size=batch_size, worker=worker, collator=collate_fn, data_limit=data_limit) data_module.setup() if visualize_random_example: index = np.random.choice(len(data_module.train_dataset)) data_module.visualize_example(index) train = data_module.train_dataloader() val = data_module.val_dataloader() coco_val_dataset = data_module.get_val_coco_text_dataset() coco_evaluator = CocoEvaluator(coco_val_dataset, ['bbox']) print('Coco Evaluator created') ### Model Part id2label = {0: 'Text'} # we have only one class to detect: Text text_detection_model = TextDetectionModel(lr=learning_rate, id2label=id2label, feature_extractor=feature_extractor, coco_evaluator=coco_evaluator, sync=sync) ### Callback Part checkpoint_callback = ModelCheckpoint( dirpath="text_detection_model_files/checkpoints", filename="best-checkpoint", save_top_k=1, verbose=True, monitor="val_loss", mode="min" ) logger = TensorBoardLogger(save_dir="text_detection_model_files/Lightning_logs", name="Text_Detection") early_stopping_callback = EarlyStopping( monitor="val_loss", min_delta=0.001, patience=15, check_finite=True, verbose=True ) lr_monitor = LearningRateMonitor(logging_interval='epoch') ### Training Part trainer = pl.Trainer(logger=logger, weights_summary="full", # only if gpu mem is overheaded -> needs much more train time benchmark=True, move_metrics_to_cpu=False, val_check_interval=0.5, gradient_clip_val=gradient_clip, # set to 0.5 to avoid exploding gradients stochastic_weight_avg=True, callbacks=[ checkpoint_callback, early_stopping_callback, lr_monitor ], max_epochs=epochs, gpus=torch.cuda.device_count(), accelerator=accelerator, precision=32, # dont change for model accumulate_grad_batches=1, # optimizer step after every n batches -> better gpu mem usage / model specific progress_bar_refresh_rate=20, # profiler='pytorch', # only for debug ) trainer.fit(text_detection_model, train, val) time.sleep(2) # short delay trained_model = text_detection_model.load_from_checkpoint(trainer.checkpoint_callback.best_model_path) trained_model.eval() trained_model.freeze() ### Saving Part # ---------------------------------- # PyTorch Model - full # ---------------------------------- try: torch.save(trained_model, "text_detection_model_files/torch_text_detection_model.pt") print('Torch model saved successful') except Exception as e: print('Cannot export as PyTorch Format -- Error : ' + str(e)) # ---------------------------------- # PyTorch Model - state dict # ---------------------------------- try: torch.save(trained_model.state_dict(), "text_detection_model_files/torch_text_detection_model_state_dict.pt") print('Torch model state dict saved successful') except Exception as e: print('Cannot export as PyTorch Format with state dict -- Error : ' + str(e)) # ---------------------------------- # onnx # ---------------------------------- try: input_batch = next(iter(val)) input_sample = { "pixel_values": input_batch["pixel_values"][0].unsqueeze(0), } values = input_sample['pixel_values'] file_path = "text_detection_model_files/torch_text_detection_model.onnx" torch.onnx.export(trained_model, values, file_path, input_names=['pixel_values'], output_names=['logits', 'pred_boxes'], dynamic_axes={'pixel_values': {0: 'batch_size', 1: 'channels', 2: 'width', 3: 'height'}, 'logits': {0: 'batch_size'}, 'pred_boxes': {0: 'batch_size'}}, export_params=True, opset_version=11, enable_onnx_checker=True, verbose=False) print('Onnx model saved successful') print('Start model quantization') model_quant = "text_detection_model_files/torch_text_detection_model.quant.onnx" quantized_model = quantize_qat(file_path, model_quant) print('Quantization succesfull') except Exception as e: print('Cannot export as ONNX Format -- Error : ' + str(e)) # Predictions model = text_detection_model.load_from_checkpoint(checkpoint_path=trainer.checkpoint_callback.best_model_path) preds = trainer.predict(model, val, return_predictions=True) print(preds)@nielsrdo you have any idea or recommendations ? ^^

2021-09-29T14:51:09Z

Summarization for downstream task

https://discuss.huggingface.co/t/summarization-for-downstream-task/10011

Hi!I was wondering id anyone could point me to any work about summarization for a downstream task.For example, given an NLP pipeline, one might want to first summarize the input and then perform some tasks (eg keywords extraction, classification, etc).For very long input, a first summarization step makes the text more treatable. I know of groups / companies that do proceed in this way, in some cases.However, one might want to directly summarize the text with the downstream task in mind: for keyword extraction, this might mean to keep as many keywords as possible, for classification to keep interesting features etc.Is anyone aware of any research work in this direction? I have looked a bit and I did not find anything, but I would be surprised no previous work exists, so I am problably searching using the wrong keywords.Any idea in this direction would also be highly appreciated

2021-09-15T08:56:59Z

[Call for participation] Interactive Grounded Language Understanding in a Collaborative Environment (IGLU) Competition@NeurIPS2021

https://discuss.huggingface.co/t/call-for-participation-interactive-grounded-language-understanding-in-a-collaborative-environment-iglu-competition-neurips2021/9851

Human intelligence has the remarkable ability to quickly adapt to new tasks and environments. Starting from a very young age, humans acquire new skills and learn how to solve new tasks either by imitating the behavior of others or by following provided natural language instructions. To facilitate research in this direction, we propose theNeurIPS IGLU competition: Interactive Grounded Language Understanding in a Collaborative Environment.The primary goal of the IGLU competition is to approach the problem of how to build interactive agents that learn to solve a task while provided with grounded natural language instructions in a collaborative environment. Understanding the complexity of the challenge, we split it into sub-tasks to make it feasible for participants.This research challenge is naturally related, but not limited, to two fields of study: Natural Language Understanding and Generation (NLU/G) and Reinforcement Learning (RL). Therefore, the suggested challenge can bring two communities together to approach one of the important challenges in AI. Another important aspect of the challenge is the dedication to perform a human-in-the-loop evaluation as a final evaluation for the agents developed by contestants.The goal of our competition is to approach the following scientific challenge: How to build interactive agents that learn to solve a task while provided with grounded natural language instructions in a collaborative environment? By the interactive agent we mean that the agent is able to follow the instructions correctly, is able to ask for clarification when needed, and is able to quickly adapt newly acquired skills, just like humans are able to do while collaboratively interacting with each other.Tasks and Application Scenarios:Given the current state of the field, our main research challenge might be too complex to suggest a reasonable end-to-end solution. Therefore, we split the problem into the following concrete research tasks:Architect Task:Given target structure, generate step instructions for the BuilderSubmission system:CodaLab - CompetitionBuilder Task:Given Architect-Builder conversation, build target structure:Submission system:https://competitions.codalab.org/competitions/33828Prizes:Architect Task:1st place - 5K $2nd place - 1.5k $3rd place - 500 $Builder Task:1st place - 5K $2nd place - 1.5k $3rd place - 500 $Timeline:July 26 – Stage 1 begins;(Tentative) October 15 – Stage 1 ends;October 22 – Stage 2 begins by deploying the top-3 performing agents for human evaluation;November 26 – The results of Stage 2 are posted, and the list of winning teams per task is released;December 6 – NeurIPS 2021 begins.Upcoming workshops:To make it even easier for you to onboard, we will arrange workshops per task:Architect Task on Sep 9, at 9 am PST: the link to sign up:IGLU - Workshop (iglu-contest.net)Builder Task on Sep 10, at 10 am PST: the link to sign up:IGLU - Workshop (iglu-contest.net)During the workshops, our team will walk you through setup, available baselines, and training environment (for Builder task). You will have a great opportunity to ask any questions, which we probably can answerGuest Lectures:If you have missed our guest lectures. Here are the links to recordings:IGLU - Guest Lecture by Marc (iglu-contest.net)IGLU - Guest Lecture by Jianwei (iglu-contest.net)For more frequent updates:follow us at Twitter@IgluContestThe news section at our website:IGLU (iglu-contest.net)For questions to organizers and mentors use the slack channel:Join IGLU on Slack | SlackRegister for the competition at CodaLab:CodaLab - Competitionhttps://competitions.codalab.org/competitions/33828

2021-09-09T14:41:39Z

Implementing a custom Attention Transformer

https://discuss.huggingface.co/t/implementing-a-custom-attention-transformer/9702

Hello everyone, currently I am trying to implement a custom attention transformer, whose attention is given on Page No. 4 of thislink. They have used hugging face for the implementation, and I am not sure about how to go for approaching this problem, and how to use hugging face to implement custom attention. Can anybody guide me, about how to go about implementing this? Thanks,

2021-09-03T04:54:27Z

Collaborative Training Experiment Round 2 with Yandex and HuggingFace

https://discuss.huggingface.co/t/collaborative-training-experiment-round-2-with-yandex-and-huggingface/9674

Let’s train an even larger model together with Yandex, HuggingFaceand Neuropark!A few months ago we assembled to train a large SahajBERT. So let’s make it even larger!Join Neuropark’s discord community with this link -NeuroparkWe are about to start the training from- 2nd SeptemberThere will be a few new things to play with beside the 4x scale:sahajBERT 2.0 will start from sahajBERT 1.0 using Net2Net model expansionwe’ll try hybrid training with both GPU and TPU and see how they compareand bring along local GPU devices (see below)If you have a GPU desktop with ≥6GB memory and ≥30Mbit upload speed, we’d really appreciate if you can bring it to the collaborative run (and will help you with the setup). You can join and leave training at any time, even if it is only for a couple of hours.Also, we’d really appreciate your ideas on the training procedure:fine-tuning benchmarks that we should run: anything beside Wikiann and Soham News Category Classification?future training runs: we’ll be able to train the model in ~2 weeks. Is there any other task that you would like to pretrain a model for? What data should we use there?Let me know if you face any issues regarding joining or anything.Check our previous models onneuropark (Neuropark)Read the blog post about our previous collaborative training -Deep Learning over the Internet: Training Language Models Collaborativelypaper link -[2106.10207] Distributed Deep Learning in Open CollaborationsThanks to Yandex and Huggingface for this initiativelets train 4x !

2021-09-01T16:17:44Z

Tutorial / codebase for models interacting while training?

https://discuss.huggingface.co/t/tutorial-codebase-for-models-interacting-while-training/9554

I need guidance on how to get started on a research project. I want to train two models (the particular architectures aren’t important) in tandem, with the ability to have the two models pass input tokens and output token between one another during training. Is there a tutorial or codebase with this functionality for me to get started?

2021-08-29T00:44:06Z

10_000 samples & 10_000 labels

https://discuss.huggingface.co/t/10-000-samples-10-000-labels/8868

Hey Community, I have a data set which each sample has it’s own label, for instance :I have 10000 sample which each sample has one word as label, and each label is unique for that sentence, this made 10000 training samples with 10000 labels.Is anyone here has an idea about how to do this or a toy code?Thank you so much for your help

2021-07-31T09:57:12Z

Best way to infer continuously with Transformer?

https://discuss.huggingface.co/t/best-way-to-infer-continuously-with-transformer/8690

Hi!I’m looking for ways to infer w/ a Transformer model in a continuous manner — basically, I want it to retain some information about the previous sample in case it was part of the same text segment.One approach I’m trying out now is inferring with intersecting windows (stride < length), and aggregating encoder embeddings of the overlapping part of the sequence (i.e. use information from window N to infer N+1). I use summing to aggregate instead of mean/dot product, as it gives the closest result to inferring as usual, but the result still doesn’t account for earlier context, meaning the approach doesn’t work.Has this problem been addressed already? Is the typical solution to just increase input length bound? (What if I don’t have enough compute to train a model with large input lengths?)

2021-07-26T12:11:00Z

The (hidden) meaning behind the embedding of the padding token?

https://discuss.huggingface.co/t/the-hidden-meaning-behind-the-embedding-of-the-padding-token/3212

So noticed that the transformers contain different embeddings for PAD tokens, and I know pad tokens typically are simply ignored for the most part (if at all present). However, as a forward pass using a batch typically contain dozens of padding tokens it would be interesting to see if these in fact hold any meaningful information (as padding tokens do attend to the sequence). Does anyone know of any research which has been conducted on what information might be present here?One might legitimately ask why this is relevant isn’t padding tokens simply a convenience for efficient processing because we need the same tensor shape? This is naturally correct, but quite a few studies have clustered the sentence embedding and it seems relevant to ask what influence the padding embeddings have on this.For a short demonstration that they indeed have different embeddings:import transformers tokenizer = transformers.AutoTokenizer.from_pretrained( "bert-base-uncased") model = transformers.BertModel.from_pretrained( "bert-base-uncased") input_ = tokenizer(["this is a sample sentence"], return_tensors="pt", # add some padding padding="max_length", max_length=128, truncation=True) output = model(**input_) # extract padding token embedding pad_tok_id = [i for i, t in enumerate(input_["input_ids"][0]) if t == 0] embedding_pad1 = output[0][0][pad_tok_id[0]] embedding_pad2 = output[0][0][pad_tok_id[1]] embedding_pad1.shape #embedding size embedding_pad1[0:10] embedding_pad2[0:10]tensor([-0.5072, -0.4916, -0.1021, -0.1485, -0.4096, 0.0536, -0.1111, 0.0525, -0.0748, -0.4794], grad_fn=<SliceBackward>) tensor([-0.6447, -0.5780, -0.1062, -0.1869, -0.3671, 0.0763, -0.0486, 0.0202, -0.1334, -0.5716], grad_fn=<SliceBackward>)

2021-01-15T09:22:01Z

Language model to search an answer in a huge collection of (unrelated) paragraphs

https://discuss.huggingface.co/t/language-model-to-search-an-answer-in-a-huge-collection-of-unrelated-paragraphs/2210

I want to build a question/answer language model to search a large collection of paragraphs.Say 10k paragraphs. And find relevant answers in them.There are 2 issues I don’t know how to solve.existing solutions often identify an answer from a short paragraph. I don’t know how to deal with a lot of paragraphs. A naive approach would be going through each paragraph and identify an answer in each of them.existing solutions will generate an answer even when fed with an unrelated paragraph. they don’t give a confidence number. If I have 10k paragraphs to search an answer from, and only 3 paragraphs have an answer, using existing solutions won’t let me to rule out unrelated paragraphs.Is there a way to generate a document embedding first (using both a question and a paragraph ), and I can use the embedding to find candidate paragraphs first and then do the actual answer search. And when there is no answer, I’d like to get a confidence number that 's below my answer threshold.Are there any papers dealing with this problem?

2020-11-25T18:59:48Z

Address extraction and formated using Places API (Google Maps API)

https://discuss.huggingface.co/t/address-extraction-and-formated-using-places-api-google-maps-api/7998

I am currently playing around with Places API from Google.I just have curious about the technique that they were using to make this happens (do only NER enough for this). I thought they firstly detect where is the address in my input, then parsing these into sub-level (like what I am gonna describe as below).When I input the text: "I wanna deliver this package to#KA B C "It gave me the result was super tremendous with 3 administrative_area_level, even it did format my input text (also correct their spelling/grammar mistake) to the address one. In detail, it somehow will be likestreet_name:{#k} administrative_area_level_1: {A} administrative_area_level_2:{B} ... formated_address: #K, A, B, CGoogle DevelopersOverview  |  Places API  |  Google DevelopersProvide type-ahead predictions for text-based geographic searches, by returning places such as businesses, addresses and points of interest as a user types.

2021-07-04T14:25:56Z

Finetuning for fp16 compatibility

https://discuss.huggingface.co/t/finetuning-for-fp16-compatibility/977

t5 and pegasus don’t really work in fp16 because they create activations that overflow fp16 bits. (they were trained in bfloat 16 which has larger range) Has anyone read/seen/heard anything about finetuning/scaling models so that their activations can fit in fp16. (or generally to encourage smaller magnitude activations?I tried one experiment on google/pegasus-xsum where I finetune with summarization lm loss and add some additional losses based on the magnitude of hidden states, but I haven’t weighted them (the model instantly forgets how to summarize) so I’m looking around.

2020-09-03T17:26:08Z

What can transformers learn without position encoding?

https://discuss.huggingface.co/t/what-can-transformers-learn-without-position-encoding/6554

So it obviously makes sense that attention mechanisms don’t have any inherent sense of position without encoding it explicitly, and for sequence prediction this seems critical. But, for example, word2vec via CBOW or skip gram is able to learn word embeddings without explicit position encoding. So my question is basically if we train a BERT model without the position encoding on the Masked LM task (something very similar to word2vec it seems to me), what is BERT capable of learning if anything? Would it be better than word2vec for creating word embeddings?

2021-06-03T15:52:11Z

Project Description

https://discuss.huggingface.co/t/project-description/6444

Hi@Madsyour project looks very interesting, would you mind adding a description?huggingface.coMads/wav2vec2-xlsr-large-53-kor-financial-engineering · Hugging FaceWe’re on a journey to advance and democratize artificial intelligence through open source and open science.

2021-05-28T09:33:44Z

Does it make sense to generate sentences with Transofmrer's encoder?

https://discuss.huggingface.co/t/does-it-make-sense-to-generate-sentences-with-transofmrers-encoder/6311

Quite a few vision+language papers pretrain BERT-based model with image-text data and finetune for image captioning task. But there is no decoder involved to generate sentences. Does that make sense? And what’s the main difference between using T’s encoder to do the sentence generation and do it with a T’ decoder?

2021-05-22T15:26:15Z

PEGASUS model overfitting

https://discuss.huggingface.co/t/pegasus-model-overfitting/6246

Hey everyone, I would like to see any scientific evidence regarding model overfitting available for thePEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarizationmodel.If anyone can point me to some resources or provide an answer, I’d greatly appreciate itThanks and stay safe

2021-05-19T07:05:14Z

Classification Heads in BERT and DistilBERT for Sequence Classification

https://discuss.huggingface.co/t/classification-heads-in-bert-and-distilbert-for-sequence-classification/6146

Hi,I have been using BertForSequenceClassification and DistilBertForSequenceClassification recently and I have noticed that they have different classification heads.BertForSequenceClassification has a dropout layer and a linear layer, whereas DistilBertForSequenceClassification has two linear layers and a dropout layer.Is there a particular reason for this?Thanks in advance!

2021-05-12T16:16:52Z

Collaborative Training Experiment of an Albert Model for Bengali

https://discuss.huggingface.co/t/collaborative-training-experiment-of-an-albert-model-for-bengali/5991

Huggingface is launching a collaborative training experiment of an Albert Model for Bengali language with our community. We are actively looking for participants who will help us to train the model.So what do you need in order to participate-A Google Colab accountThat’s everything you need.[Although if you want to use the power of your own GPUs, Huggingface will also provide a script for that.]How you can contribute?If you are a native Bengali speaker, that would be a great help, we are looking for participants who will check the performance of the tokenizer, sentence splitter, etc.You might want to help us preprocessing the dataset. We are using the Wikidump and OSCAR Bengali dataset to train the model, if you have some suggestions on preprocessing these feel free to contribute in that part.Now the main part, distributive training. You have been provided a google colab script in order to start the training and if your kernel crashes just restart the training script. (Non native speakers can participate)Join our discord community link-https://discord.gg/GD9G4j8fJU[A separate slack channel from Huggingface will be provided where you will get to know more about the distributive training framework and other related things.]We are aiming to start this collaborative training experiment from -May 7thPlease do participate in this first Huggingface collaborative training experiment specifically the native bengali speakers.

2021-05-05T06:15:19Z

Task-specific fine-tuning of GPT2

https://discuss.huggingface.co/t/task-specific-fine-tuning-of-gpt2/5700

Hi thereIn the Seq2Seq examples (transformers/examples/legacy/seq2seq at master · huggingface/transformers · GitHub) why there is no mention of GPT-x? it seems to me that, it shouldn’t be difficult to fine-tune this model usingGPT2LMHeadModelfor particular text-to-text tasks.Wondering if anyone has any thoughts on this.Thanks!

2021-04-22T19:37:46Z

Is causal language modeling (CLM) vs masked language modeling (MLM) a common distinction in NLP research?

https://discuss.huggingface.co/t/is-causal-language-modeling-clm-vs-masked-language-modeling-mlm-a-common-distinction-in-nlp-research/5665

Thehuggingface documentationstates:GPT and GPT-2 are fine-tuned using a causal language modeling (CLM) loss while BERT and RoBERTa are fine-tuned using a masked language modeling (MLM) loss.I have two questions regarding this statement:Is this a common distinction you’d find in the NLP literature (any literature on this distinction)?Is it a sensible distinction in your opinion? I have two questions While I totally agree with CLM, I don’t understand why you would call BERT & co. “masked language models”, since causal language models do the actual masking in next token prediction?Thanks!

2021-04-21T14:30:19Z

Any ways to visualize attention of the LXMERT?

https://discuss.huggingface.co/t/any-ways-to-visualize-attention-of-the-lxmert/5579

I would like to observe the attention between an input RoI and each word in an input sentence of LXMERT. If a framework that facilitates what I want do exists, please let me know. If not, could you tell me which of the tensors from LXMERT I should watch?

2021-04-17T17:06:53Z

Human Evaluation and Statistical significance

https://discuss.huggingface.co/t/human-evaluation-and-statistical-significance/5374

Hello,I have recently conducted a human evaluation of a chatbot via a survey. I wonder how I can prove that the results are statistically significant.More specifically, I compared the generated responses of two chatbots and calculated each one’s win rate. Moreover, participants were asked to rate each model according to “relevance” and “fluency” using a scale ranging from 1 to 5.According to some references (e.g.DodecaDialogue paper), they prove that the results are statistically significant using binomial testing.How can I apply binomial testing in the aforementioned case?@patrickvonplaten

2021-04-08T20:21:45Z