vector
listlengths
1.02k
1.02k
text
stringlengths
2
11.8k
[ 0.033163633, 0.02203055, -0.0021279508, -0.0020248666, -0.03269239, -0.010433585, -0.017215047, 0.03119031, 0.025240883, 0.0310725, -0.017612658, 0.0017413853, -0.00055131596, -0.026301177, -0.02588884, 0.011317164, -0.03501915, -0.029025542, -0.026286451, 0.0013870336, 0.013...
from transformers import TapasTokenizer, TFTapasForQuestionAnswering import pandas as pd model_name = "google/tapas-base-finetuned-wtq" model = TFTapasForQuestionAnswering.from_pretrained(model_name) tokenizer = TapasTokenizer.from_pretrained(model_name) data = {"Actors": ["Brad Pitt", "Leonardo Di Caprio", "George C...
[ 0.038421955, -0.02673783, -0.02488529, 0.004700636, 0.0013912277, -0.019590236, -0.0033878128, 0.012858372, -0.03109932, 0.030078234, 0.040785037, 0.0068339733, 0.017577242, -0.035475396, -0.04014321, 0.048778668, -0.04463598, 0.002875447, -0.044140026, -0.014499401, 0.016381...
In case of a conversational set-up, then each table-question pair must be provided sequentially to the model, such that the prev_labels token types can be overwritten by the predicted labels of the previous table-question pair. Again, more info can be found in this notebook (for PyTorch) and this notebook (for TensorFl...
[ 0.015748007, -0.0018984413, -0.027148908, 0.003626006, -0.00804488, -0.025549501, -0.034148023, 0.0004066868, -0.014654395, 0.02933613, 0.000115448645, 0.031906117, -0.012398821, -0.044920094, -0.016855288, 0.019261234, -0.020095112, -0.0027511166, -0.03578844, -0.028597943, ...
Text classification task guide Masked language modeling task guide TAPAS specific outputs [[autodoc]] models.tapas.modeling_tapas.TableQuestionAnsweringOutput TapasConfig [[autodoc]] TapasConfig TapasTokenizer [[autodoc]] TapasTokenizer - call - convert_logits_to_predictions - save_vocabulary
[ 0.0052463645, -0.00729514, -0.006340636, 0.006514492, -0.0046634353, 0.02408078, 0.0037191578, 0.009054155, 0.006139508, 0.08126923, 0.005696345, 0.04216182, -0.0014488011, -0.023385355, -0.025130734, 0.037907455, -0.0006434382, -0.052579552, -0.021067273, -0.019417346, 0.012...
TapasModel [[autodoc]] TapasModel - forward TapasForMaskedLM [[autodoc]] TapasForMaskedLM - forward TapasForSequenceClassification [[autodoc]] TapasForSequenceClassification - forward TapasForQuestionAnswering [[autodoc]] TapasForQuestionAnswering - forward
[ -0.0076160273, -0.02888368, -0.021117786, 0.01795693, -0.011512599, -0.01235731, -0.020504689, 0.0002620564, 0.0044313287, 0.074988574, 0.009032963, 0.039973926, 0.021908, -0.017970555, -0.05013771, 0.054388516, -0.013467697, -0.045287434, -0.022548346, 0.0038999782, 0.011860...
TFTapasModel [[autodoc]] TFTapasModel - call TFTapasForMaskedLM [[autodoc]] TFTapasForMaskedLM - call TFTapasForSequenceClassification [[autodoc]] TFTapasForSequenceClassification - call TFTapasForQuestionAnswering [[autodoc]] TFTapasForQuestionAnswering - call
[ 0.02157659, -0.017025623, -0.025936093, -0.010677835, 0.004550967, -0.03349158, -0.0019054412, -0.018071314, -0.01090612, 0.041297447, -0.01184135, -0.0075370786, 0.03493493, -0.038351838, 0.005011218, 0.0017139766, 0.0018833491, -0.02776237, -0.060031522, -0.005143771, -0.02...
LXMERT Overview The LXMERT model was proposed in LXMERT: Learning Cross-Modality Encoder Representations from Transformers by Hao Tan & Mohit Bansal. It is a series of bidirectional transformer encoders (one for the vision modality, one for the language modality, and then one to fuse both modalities) pretrained using ...
[ -0.0038921777, 0.00033864385, -0.008129327, -0.04289645, -0.033267245, -0.021598212, -0.004323392, 0.01654363, -0.03284728, 0.044366326, -0.007769357, 0.017023591, 0.03536707, -0.021193245, 0.002523541, -0.02089327, -0.011219071, -0.017203575, -0.06365473, 0.020098336, -0.016...
Bounding boxes are not necessary to be used in the visual feature embeddings, any kind of visual-spacial features will work. Both the language hidden states and the visual hidden states that LXMERT outputs are passed through the cross-modality layer, so they contain information from both modalities. To access a m...
[ 0.054471683, 0.0128149465, -0.05042417, 0.016777592, -0.032693453, 0.018971082, -0.005092032, 0.014649383, -0.0063389265, 0.059380922, -0.0004647295, -0.0051899552, 0.002870796, -0.03930004, 0.004096474, -0.015772242, -0.016281445, 0.0022718294, -0.021386532, -0.04010954, 0.0...
Resources Question answering task guide
[ 0.0069392812, -0.011194564, -0.029699314, 0.0141078485, -0.021566397, -0.02967234, 0.020312065, 0.03479756, -0.0070067183, 0.06619629, -0.0037326454, 0.04661255, 0.029132843, -0.032882348, -0.03528311, 0.049390957, 0.020797612, -0.034878485, -0.035364036, 0.0011911084, -0.030...
LxmertModel [[autodoc]] LxmertModel - forward LxmertForPreTraining [[autodoc]] LxmertForPreTraining - forward LxmertForQuestionAnswering [[autodoc]] LxmertForQuestionAnswering - forward TFLxmertModel [[autodoc]] TFLxmertModel - call TFLxmertForPreTraining [[autodoc]] TFLxmertForPreTraining - call
[ -0.00014696429, 0.004849009, -0.012582092, 0.036635518, -0.025475975, -0.051731426, 0.0018204084, 0.0092628095, -0.008061113, 0.05383602, -0.019707829, 0.028321072, 0.018265793, -0.044586204, -0.044586204, 0.021045934, -0.0027541593, -0.033777427, -0.08137762, 0.01054895, -0....
LxmertConfig [[autodoc]] LxmertConfig LxmertTokenizer [[autodoc]] LxmertTokenizer LxmertTokenizerFast [[autodoc]] LxmertTokenizerFast Lxmert specific outputs [[autodoc]] models.lxmert.modeling_lxmert.LxmertModelOutput [[autodoc]] models.lxmert.modeling_lxmert.LxmertForPreTrainingOutput [[autodoc]] models.lxmert.modelin...
[ 0.017895794, -0.032893483, -0.016809005, -0.009665178, 0.0076220143, -0.002706105, 0.0060316795, -0.019344846, 0.0060932646, 0.05471621, -0.0046840613, -0.028720213, 0.038139053, -0.008527672, 0.038805615, 0.030575, -0.0059556044, -0.006723602, -0.054861113, 0.004557269, 0.01...
DPT Overview The DPT model was proposed in Vision Transformers for Dense Prediction by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun. DPT is a model that leverages the Vision Transformer (ViT) as backbone for dense prediction tasks like semantic segmentation and depth estimation. The abstract from the paper is the f...
[ 0.0091290455, -0.023479, 0.009106412, -0.013474774, -0.026119635, 0.036425646, -0.036908507, 0.0110076675, 0.021140154, 0.038960654, -0.01875604, -0.051605515, 0.04590175, 0.0013985925, 0.010373916, 0.06271126, -0.03075206, -0.015526922, -0.09572672, 0.00553024, 0.0065638595,...
DPT architecture. Taken from the original paper. This model was contributed by nielsr. The original code can be found here. Usage tips DPT is compatible with the [AutoBackbone] class. This allows to use the DPT framework with various computer vision backbones available in the library, such as [VitDetBackbone] or [D...
[ 0.017458001, -0.011414847, -0.019518923, -0.0011010973, -0.028746534, -0.016846374, -0.026725503, 0.010457516, 0.022497289, 0.04337243, -0.007605466, 0.0062359506, 0.051908635, -0.0345703, 0.0027506659, 0.03507556, 0.024903912, -0.026645724, -0.043957468, -0.02776261, -0.0172...
If you're interested in submitting a resource to be included here, please feel free to open a Pull Request and we'll review it! The resource should ideally demonstrate something new instead of duplicating an existing resource. DPTConfig [[autodoc]] DPTConfig DPTFeatureExtractor [[autodoc]] DPTFeatureExtractor - c...
[ 0.020834085, -0.027459865, -0.025003757, 0.0012307318, -0.014950847, -0.020634169, 0.0114309015, 0.022047859, 0.016093224, 0.047123015, -0.020277176, -0.0012307318, 0.051920995, -0.029758899, 0.011802173, 0.040068842, 0.020719847, -0.019777387, -0.033700097, -0.0145938555, -0...
Resources A list of official Hugging Face and community (indicated by 🌎) resources to help you get started with DPT. Demo notebooks for [DPTForDepthEstimation] can be found here. Semantic segmentation task guide Monocular depth estimation task guide
[ 0.04217103, -0.004818486, 0.005564646, -0.029296989, -0.020090636, 0.025837183, 0.008211472, -0.0054310053, -0.0050783427, 0.06349413, 0.02999489, 0.016155664, -0.016749622, 0.0059915534, 0.0037753468, 0.008649517, -0.021501288, -0.07353203, -0.0695525, 0.008649517, -0.009250...
CANINE uses no less than 3 Transformer encoders internally: 2 "shallow" encoders (which only consist of a single layer) and 1 "deep" encoder (which is a regular BERT encoder). First, a "shallow" encoder is used to contextualize the character embeddings, using local attention. Next, after downsampling, a "deep" en...
[ 0.069971345, -0.0018509016, 0.0125062885, 0.0000034768202, -0.05069293, 0.017492836, 0.019989738, -0.0039195567, -0.0011323163, 0.045524925, 0.048950907, 0.021949517, -0.036843833, -0.034143694, -0.01077878, 0.03754064, -0.032314565, -0.021528527, -0.043782897, -0.024852892, ...
Model checkpoints: google/canine-c: Pre-trained with autoregressive character loss, 12-layer, 768-hidden, 12-heads, 121M parameters (size ~500 MB). google/canine-s: Pre-trained with subword loss, 12-layer, 768-hidden, 12-heads, 121M parameters (size ~500 MB). Usage example CANINE works on raw characters, so i...
[ 0.038939457, -0.015971377, -0.009897235, -0.040592685, -0.017388431, -0.008908249, 0.030260002, -0.024739398, -0.02401611, 0.0649483, 0.033153154, 0.009594635, -0.028621534, -0.021388656, -0.022126704, -0.0073103737, -0.012768245, -0.04245257, -0.08183486, 0.019720666, 0.0156...
CANINE Overview The CANINE model was proposed in CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting. It's among the first papers that trains a Transformer without using an explicit tokenization step (such as Byte Pair Enc...
[ 0.040417846, -0.013018106, -0.0011823225, -0.016304927, -0.014582575, 0.006803288, 0.010779049, -0.0010459697, 0.007973052, 0.054799482, 0.004510408, 0.008367757, -0.0024059096, -0.025404684, 0.0065557, 0.030887503, -0.020051042, -0.033212677, -0.04934537, 0.009027992, 0.0105...
from transformers import CanineModel import torch model = CanineModel.from_pretrained("google/canine-c") # model pre-trained with autoregressive character loss text = "hello world" use Python's built-in ord() function to turn each character into its unicode code point id input_ids = torch.tensor([[ord(char) for char i...
[ 0.028144155, 0.03697093, 0.015327002, 0.010878364, -0.05826235, -0.007564799, -0.02388587, -0.0063803755, 0.00007755153, 0.047489736, -0.0006036681, -0.015665408, -0.03398167, -0.031330816, -0.05933397, 0.028623562, -0.02750964, -0.08014598, -0.039593577, 0.0008821486, -0.020...
For batched inference and training, it is however recommended to make use of the tokenizer (to pad/truncate all sequences to the same length): thon
[ 0.029339885, -0.0065374076, -0.027326362, 0.011924231, -0.022384083, 0.0025250737, 0.0057594557, -0.013976977, 0.00087437825, 0.05172397, -0.010165669, -0.0034452137, 0.023691565, -0.038335357, 0.00025209878, -0.016147396, -0.035589647, -0.01808247, -0.049501248, -0.034099117, ...
Resources Text classification task guide Token classification task guide Question answering task guide Multiple choice task guide
[ 0.025606617, 0.007096059, 0.024127983, -0.010965381, -0.03487917, -0.032308836, 0.008270674, 0.011683969, 0.007890652, 0.038859043, 0.024777476, 0.0325023, 0.017964708, -0.045657992, -0.011469774, 0.0166519, -0.012305824, -0.06881864, -0.04739919, 0.0247222, -0.0060216314, ...
from transformers import CanineTokenizer, CanineModel model = CanineModel.from_pretrained("google/canine-c") tokenizer = CanineTokenizer.from_pretrained("google/canine-c") inputs = ["Life is like a box of chocolates.", "You never know what you gonna get."] encoding = tokenizer(inputs, padding="longest", truncation=True...
[ 0.012449311, -0.0064811665, 0.0046582315, -0.012230422, -0.021095427, 0.004189672, -0.010308304, 0.032258764, -0.005407242, 0.045556273, 0.0005942494, 0.04268335, -0.012948652, -0.0046719117, -0.027908346, 0.030234043, -0.022572927, -0.06254753, -0.06933309, -0.005044707, -0....
CanineConfig [[autodoc]] CanineConfig CanineTokenizer [[autodoc]] CanineTokenizer - build_inputs_with_special_tokens - get_special_tokens_mask - create_token_type_ids_from_sequences CANINE specific outputs [[autodoc]] models.canine.modeling_canine.CanineModelOutputWithPooling CanineModel [[autodoc]] Canin...
[ 0.026265187, -0.022698851, 0.0016022933, 0.010669775, -0.01737858, -0.04832093, -0.008667366, 0.014265344, -0.040252823, 0.03630647, -0.023181183, -0.027434478, 0.02123724, -0.0268206, -0.024335857, 0.042445246, -0.0014305538, -0.027682953, -0.03554643, 0.02980229, 0.00519603...
The architecture is similar than llava architecture except that the multi-modal projector takes a set of concatenated vision hidden states and has an additional layernorm layer on that module. We advise users to use padding_side="left" when computing batched generation as it leads to more accurate results. Simply make...
[ 0.023771105, -0.007498142, -0.034044802, 0.026151229, -0.0049560796, -0.020170793, 0.031724934, 0.0104394015, -0.024207963, 0.059021033, -0.003991979, -0.0015949085, -0.0074529494, -0.022370147, -0.00016805856, 0.017459258, 0.0020713098, -0.03410506, -0.034918517, -0.009136359,...
VipLlava Overview The VipLlava model was proposed in Making Large Multimodal Models Understand Arbitrary Visual Prompts by Mu Cai, Haotian Liu, Siva Karthik Mustikovela, Gregory P. Meyer, Yuning Chai, Dennis Park, Yong Jae Lee. VipLlava enhances the training protocol of Llava by marking images and interact with the mo...
[ 0.03513639, 0.004099727, -0.030892309, 0.013692222, -0.010574119, -0.027037987, -0.023429073, 0.042873904, -0.024915947, 0.03860095, 0.016341165, 0.009960604, -0.008430424, -0.039207246, -0.045039255, -0.008574781, 0.0008350126, -0.008228325, -0.027211215, -0.046425078, 0.010...
A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions.###Human: <image>\n<prompt1>###Assistant: <answer1>###Human: <prompt2>###Assistant: The original code can be found here. This model was contributed by Younes Belkad...
[ -0.003865014, -0.0037053027, -0.017192036, 0.014466297, -0.019151162, -0.039949127, 0.0038259735, -0.014977373, 0.00055632775, 0.017688917, 0.0032918279, 0.033049595, 0.018313564, -0.0020797963, -0.0503694, 0.01767472, -0.034668006, -0.016979089, -0.043469872, 0.015502646, 0....
NLLB Updated tokenizer behavior DISCLAIMER: The default behaviour for the tokenizer was fixed and thus changed in April 2023. The previous version adds [self.eos_token_id, self.cur_lang_code] at the end of the token sequence for both target and source tokenization. This is wrong as the NLLB paper mentions (page 48, 6....
[ 0.05569324, -0.006403592, -0.013887989, -0.008173145, -0.021107484, -0.051793862, -0.025247036, 0.036902774, -0.005121461, 0.072449245, 0.020867305, -0.022124711, -0.019482743, -0.029527871, -0.030347304, 0.004273771, 0.004644635, -0.03823082, -0.022845248, -0.045294907, 0.03...
Note the model has not been explicitly trained to process multiple images in the same prompt, although this is technically possible, you may experience inaccurate results. For better results, we recommend users to prompt the model with the correct prompt format: A chat between a curious human and an artificial intel...
[ 0.035344396, 0.006291839, -0.03653007, -0.014693896, -0.011320369, -0.042119678, 0.047229372, 0.008179744, 0.009612434, 0.05262137, -0.0123296045, 0.018095654, 0.031787373, -0.041950297, -0.039268415, -0.006002478, -0.039381336, -0.045930777, -0.01997297, 0.0029818306, -0.001...
from transformers import NllbTokenizer tokenizer = NllbTokenizer.from_pretrained("facebook/nllb-200-distilled-600M") tokenizer("How was your day?").input_ids [13374, 1398, 4260, 4039, 248130, 2, 256047] 2: '' 256047 : 'eng_Latn' New behaviour thon from transformers import NllbTokenizer tokenizer = NllbTokenizer.from...
[ 0.025060242, -0.011869893, -0.049751326, 0.0024669785, 0.011962184, -0.019920407, 0.018330183, -0.021610022, -0.0148657635, 0.054493602, 0.01081211, 0.00037093417, -0.00033676918, -0.051540326, -0.013509813, 0.0023036965, -0.00025557188, -0.02581276, -0.03842097, -0.0078091407,...
For more details, feel free to check the linked PR and Issue. Overview The NLLB model was presented in No Language Left Behind: Scaling Human-Centered Machine Translation by Marta R. Costa-jussà, James Cross, Onur Çelebi, Maha Elbayad, Kenneth Heafield, Kevin Heffernan, Elahe Kalbassi, Janice Lam, Daniel Licht, Jean ...
[ 0.008487725, -0.00024654123, -0.017484438, 0.03615193, -0.014664367, -0.011287161, 0.019699225, 0.0068954164, 0.008707828, 0.04778988, -0.009773952, 0.014898227, 0.019300288, -0.011851175, -0.016535243, 0.03769265, -0.040774096, -0.052962303, -0.02447271, 0.020896036, 0.00922...
Enabling the old behaviour can be done as follows: thon from transformers import NllbTokenizer tokenizer = NllbTokenizer.from_pretrained("facebook/nllb-200-distilled-600M", legacy_behaviour=True)
[ -0.020252764, -0.006451602, -0.018204408, 0.00626047, -0.0026992867, -0.01883911, 0.01689173, -0.004139988, -0.024984179, 0.04116908, 0.0017769854, 0.0039777067, 0.0046448647, -0.039005324, -0.01400672, -0.032629453, -0.035543315, -0.03462011, -0.058075234, 0.035024013, -0.00...
Generating from any other language than English English (eng_Latn) is set as the default language from which to translate. In order to specify that you'd like to translate from a different language, you should specify the BCP-47 code in the src_lang keyword argument of the tokenizer initialization. See example below fo...
[ 0.028072031, 0.008109233, -0.008325386, 0.022689117, -0.010005804, -0.024599634, 0.0042254473, 0.0210854, 0.0115397945, 0.04258916, -0.00023619972, 0.0052957544, 0.024292836, -0.04889247, -0.040859934, 0.02367924, -0.022912243, -0.04919927, -0.012760014, -0.034751862, -0.0024...
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer tokenizer = AutoTokenizer.from_pretrained("facebook/nllb-200-distilled-600M") model = AutoModelForSeq2SeqLM.from_pretrained("facebook/nllb-200-distilled-600M") article = "UN Chief says there is no military solution in Syria" inputs = tokenizer(article, ret...
[ 0.014498531, 0.014154664, -0.009937038, 0.03160764, -0.009642296, -0.00935457, 0.008814208, -0.0018158252, 0.011810759, 0.04847815, -0.0013395001, -0.00054913363, 0.02240044, -0.043649983, -0.016351199, 0.03820426, -0.030316386, -0.06742589, -0.027509313, -0.014063435, 0.0052...
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer tokenizer = AutoTokenizer.from_pretrained( "facebook/nllb-200-distilled-600M", token=True, src_lang="ron_Latn" ) model = AutoModelForSeq2SeqLM.from_pretrained("facebook/nllb-200-distilled-600M", token=True) article = "Şeful ONU spune că nu există o s...
[ 0.015604513, -0.014539036, -0.032314964, 0.026596455, -0.020985844, -0.005755597, 0.008867733, 0.017519673, 0.0069019957, 0.052221842, -0.007195339, 0.0018342383, -0.0021022933, -0.06484572, -0.0064434363, 0.01324428, -0.01885489, -0.048634287, -0.05168236, 0.023211207, -0.01...
Resources Translation task guide Summarization task guide NllbTokenizer [[autodoc]] NllbTokenizer - build_inputs_with_special_tokens NllbTokenizerFast [[autodoc]] NllbTokenizerFast
[ 0.020849405, -0.015705494, -0.015676677, -0.032016154, -0.015921626, -0.0017974867, -0.0105327675, -0.024451582, -0.034609724, 0.06622244, 0.00036697186, -0.024667712, 0.054493744, -0.020273056, 0.02626708, -0.015157964, -0.0058895606, -0.015157964, -0.083570525, -0.007236775, ...
DeiT Overview The DeiT model was proposed in Training data-efficient image transformers & distillation through attention by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou. The Vision Transformer (ViT) introduced in Dosovitskiy et al., 2020 has shown that one can match...
[ 0.01283544, -0.015659535, -0.011161902, -0.051461298, -0.035174187, -0.01627217, -0.035562687, -0.015166439, -0.047905028, 0.0749506, 0.009047566, -0.0014494036, 0.057557758, -0.024520323, 0.0061114025, -0.013194055, -0.014105536, -0.005905946, -0.102265134, -0.0096602, -0.00...
Compared to ViT, DeiT models use a so-called distillation token to effectively learn from a teacher (which, in the DeiT paper, is a ResNet like-model). The distillation token is learned through backpropagation, by interacting with the class ([CLS]) and patch tokens through the self-attention layers. There are 2 w...
[ 0.03520608, -0.02312798, 0.0055626426, -0.019686889, -0.04299309, -0.03402706, -0.040580217, 0.018466739, -0.0025705374, 0.070741184, 0.027940027, -0.0088357935, 0.044473723, -0.031806115, -0.015149033, 0.041814074, 0.02485538, -0.0398399, -0.05544135, -0.018425612, 0.0205917...
Besides that: [DeiTForMaskedImageModeling] is supported by this example script. If you're interested in submitting a resource to be included here, please feel free to open a Pull Request and we'll review it! The resource should ideally demonstrate something new instead of duplicating an existing resource. DeiTConfig ...
[ 0.011416229, -0.04272323, 0.0052201245, -0.0045121945, -0.018571133, 0.003311119, -0.0012079484, 0.02294243, -0.024963122, 0.05641446, -0.014199838, 0.0022509424, 0.048139237, -0.010275294, -0.004013894, 0.02925194, 0.0014012546, -0.029499372, -0.044840146, -0.0115880575, 0.0...
DeiTModel [[autodoc]] DeiTModel - forward DeiTForMaskedImageModeling [[autodoc]] DeiTForMaskedImageModeling - forward DeiTForImageClassification [[autodoc]] DeiTForImageClassification - forward DeiTForImageClassificationWithTeacher [[autodoc]] DeiTForImageClassificationWithTeacher - forward
[ 0.032722127, -0.009966422, 0.005321666, -0.020624157, -0.032347664, -0.024642412, 0.0276525, -0.0014762403, -0.015223278, 0.04839188, -0.008288549, 0.005069625, 0.046173915, -0.03252049, -0.023331799, 0.012544442, 0.0067438977, -0.040845048, -0.052597363, -0.0067727026, -0.00...
Resources A list of official Hugging Face and community (indicated by 🌎) resources to help you get started with DeiT. [DeiTForImageClassification] is supported by this example script and notebook. See also: Image classification task guide Besides that: [DeiTForMaskedImageModeling] is supported by this example scrip...
[ 0.060596935, 0.0169471, 0.002472066, 0.018509181, -0.0008008426, -0.005183602, 0.01108193, -0.045919273, 0.01808182, 0.06666842, 0.023740677, 0.0063035847, -0.026805893, -0.06100956, 0.019629164, -0.03133003, -0.0058614863, -0.06201165, -0.06407478, -0.007906191, -0.000819263...
PEGASUS-X Overview The PEGASUS-X model was proposed in Investigating Efficiently Extending Transformers for Long Input Summarization by Jason Phang, Yao Zhao and Peter J. Liu. PEGASUS-X (PEGASUS eXtended) extends the PEGASUS models for long input summarization through additional long input pretraining and using stagg...
[ 0.00073781254, -0.033152767, 0.002457429, 0.013079594, 0.027093446, -0.03325954, 0.019566005, 0.024691071, -0.011137675, 0.07474054, 0.0108707445, 0.026306, 0.043189354, -0.02434406, -0.01855167, 0.038971853, 0.007307222, -0.033312924, -0.061394013, -0.015962444, 0.013786959,...
TFDeiTModel [[autodoc]] TFDeiTModel - call TFDeiTForMaskedImageModeling [[autodoc]] TFDeiTForMaskedImageModeling - call TFDeiTForImageClassification [[autodoc]] TFDeiTForImageClassification - call TFDeiTForImageClassificationWithTeacher [[autodoc]] TFDeiTForImageClassificationWithTeacher - call
[ 0.020140726, 0.025946317, -0.0102560865, -0.029578252, 0.062128086, -0.0011831301, -0.04080423, -0.012794313, -0.035218753, 0.06036715, -0.005612989, -0.037227325, 0.038933232, -0.05552457, 0.015490749, -0.024419254, -0.01692151, -0.024488041, -0.022451956, -0.02922056, -0.00...
Wav2Vec2-BERT follows the same architecture as Wav2Vec2-Conformer, but employs a causal depthwise convolutional layer and uses as input a mel-spectrogram representation of the audio instead of the raw waveform. Wav2Vec2-BERT can use either no relative position embeddings, Shaw-like position embeddings, Transformer-XL...
[ 0.021349762, 0.03445245, -0.028804997, 0.023292007, 0.01105586, -0.045956522, -0.020737208, -0.03564768, 0.005232858, 0.039681572, -0.0069173826, -0.015343741, 0.01786866, -0.08312812, 0.009300369, -0.009629057, -0.021484226, -0.042639762, -0.029522134, -0.013334263, 0.001100...
Wav2Vec2-BERT Overview The Wav2Vec2-BERT model was proposed in Seamless: Multilingual Expressive and Streaming Speech Translation by the Seamless Communication team from Meta AI. This model was pre-trained on 4.5M hours of unlabeled audio data covering more than 143 languages. It requires finetuning to be used for dow...
[ 0.012706923, 0.027333809, 0.0053788247, 0.0022845347, 0.0043931953, -0.026161313, -0.02324473, -0.032126386, -0.017250344, 0.06554252, 0.041506354, -0.007328099, 0.019741898, -0.05126738, -0.033826504, 0.008954937, -0.02852096, -0.054726243, -0.05724711, -0.05200019, -0.01090...
Resources [Wav2Vec2BertForCTC] is supported by this example script. You can also adapt these notebooks on how to finetune a speech recognition model in English, and how to finetune a speech recognition model in any language. [Wav2Vec2BertForSequenceClassification] can be used by adapting this example script. See also...
[ -0.0033356028, 0.014156801, 0.007234492, -0.0056939386, 0.008225332, -0.0040515866, -0.027037725, 0.0019935572, -0.015948458, 0.06352237, 0.014034643, -0.020278294, 0.045660093, -0.04606729, -0.012833418, 0.019803233, -0.036511786, -0.05339679, -0.061839294, -0.046990264, 0.0...
Wav2Vec2BertConfig [[autodoc]] Wav2Vec2BertConfig Wav2Vec2BertProcessor [[autodoc]] Wav2Vec2BertProcessor - call - pad - from_pretrained - save_pretrained - batch_decode - decode Wav2Vec2BertModel [[autodoc]] Wav2Vec2BertModel - forward Wav2Vec2BertForCTC [[autodoc]] Wav2Vec2BertForCTC ...
[ 0.026699143, -0.018105665, 0.005671271, 0.011879634, 0.0056324024, 0.031716716, 0.0044416124, -0.0070882053, 0.017964326, 0.068352066, -0.00089441775, 0.020946601, -0.01932119, -0.020861797, 0.025186803, -0.0014195845, -0.0074132876, -0.034967538, -0.07197037, -0.0023815804, ...
Translation task guide Summarization task guide PEGASUS-X uses the same tokenizer as PEGASUS. PegasusXConfig [[autodoc]] PegasusXConfig PegasusXModel [[autodoc]] PegasusXModel - forward PegasusXForConditionalGeneration [[autodoc]] PegasusXForConditionalGeneration - forward
[ 0.020806994, -0.006721601, -0.020831458, -0.015779555, -0.012489089, 0.0072903987, -0.01102734, -0.02957749, -0.019755023, 0.03564467, -0.0059020426, 0.013149628, 0.019669399, -0.034030017, -0.020880386, -0.010599212, -0.05323459, -0.03427466, -0.056708537, -0.038849507, -0.0...
Text classification task guide Token classification task guide Question answering task guide Causal language modeling task guide Masked language modeling task guide Multiple choice task guide
[ 0.038789418, -0.03567922, -0.033772025, -0.023238441, 0.025688453, -0.009242562, 0.02963488, -0.014641391, 0.009697354, 0.055191297, 0.046154123, -0.0107609825, 0.00007249405, -0.01640188, 0.01112775, -0.02401599, -0.0016367036, -0.01440666, -0.03462293, -0.0044378964, 0.0044...
RoFormer Overview The RoFormer model was proposed in RoFormer: Enhanced Transformer with Rotary Position Embedding by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu. The abstract from the paper is the following: Position encoding in transformer architecture provides supervision for dependency modeli...
[ 0.003379352, -0.0061334493, -0.01761299, -0.0049656457, -0.008521988, -0.004423097, 0.023170808, 0.002529138, -0.010500306, 0.084108315, 0.04411849, 0.016316166, 0.035199516, -0.03490839, -0.012068406, 0.022919383, -0.013351996, -0.067964174, -0.045918167, -0.033135183, 0.016...
RoFormerModel [[autodoc]] RoFormerModel - forward RoFormerForCausalLM [[autodoc]] RoFormerForCausalLM - forward RoFormerForMaskedLM [[autodoc]] RoFormerForMaskedLM - forward RoFormerForSequenceClassification [[autodoc]] RoFormerForSequenceClassification - forward RoFormerForMultipleChoice [[autodoc]] ...
[ -0.0011240422, -0.024230687, -0.016053343, 0.00877318, -0.007999277, -0.0015674953, -0.004455077, 0.00798558, -0.00745823, 0.06289843, 0.045776684, 0.03435306, 0.044461735, -0.023600606, -0.035531037, 0.038571857, -0.0066055674, -0.042955022, -0.034709193, -0.022888342, 0.024...
TFRoFormerModel [[autodoc]] TFRoFormerModel - call TFRoFormerForMaskedLM [[autodoc]] TFRoFormerForMaskedLM - call TFRoFormerForCausalLM [[autodoc]] TFRoFormerForCausalLM - call TFRoFormerForSequenceClassification [[autodoc]] TFRoFormerForSequenceClassification - call TFRoFormerForMultipleChoice [[auto...
[ -0.00040348351, -0.039567858, -0.012091697, 0.010875695, 0.0033952391, -0.010944011, -0.0055676457, 0.040087048, 0.012979788, 0.069954224, 0.03525037, 0.042874288, 0.039704487, -0.0046829707, -0.038064934, 0.05331277, -0.0017010353, -0.07481823, -0.026547082, -0.00586823, 0.0...
FlaxRoFormerModel [[autodoc]] FlaxRoFormerModel - call FlaxRoFormerForMaskedLM [[autodoc]] FlaxRoFormerForMaskedLM - call FlaxRoFormerForSequenceClassification [[autodoc]] FlaxRoFormerForSequenceClassification - call FlaxRoFormerForMultipleChoice [[autodoc]] FlaxRoFormerForMultipleChoice - call FlaxRoFo...
[ -0.0033429826, 0.0076302714, -0.027391778, 0.009780808, -0.023752408, -0.00956024, 0.011441961, 0.01309622, 0.017342152, 0.052853588, 0.029114965, 0.00862972, 0.013392609, -0.056327533, -0.007285634, 0.022442786, -0.017355938, -0.03950923, -0.054342423, -0.0054383776, -0.0128...
RoFormerConfig [[autodoc]] RoFormerConfig RoFormerTokenizer [[autodoc]] RoFormerTokenizer - build_inputs_with_special_tokens - get_special_tokens_mask - create_token_type_ids_from_sequences - save_vocabulary RoFormerTokenizerFast [[autodoc]] RoFormerTokenizerFast - build_inputs_with_special_tokens
[ -0.0044704606, -0.015932707, -0.018324828, -0.0110377185, -0.030359263, 0.016715314, -0.030418327, -0.0073424825, -0.028734984, 0.047340367, -0.04270379, -0.046040945, 0.03762422, -0.013289561, 0.00023925824, 0.018841643, -0.0002319905, -0.025766982, -0.047045045, 0.0261952, ...
OWL-ViT architecture. Taken from the original paper. This model was contributed by adirik. The original code can be found here. Usage tips OWL-ViT is a zero-shot text-conditioned object detection model. OWL-ViT uses CLIP as its multi-modal backbone, with a ViT-like Transformer to get visual features and a causal la...
[ 0.018949833, 0.0023152588, -0.0074002733, -0.026649538, -0.017253045, -0.004894304, -0.033650566, 0.0028446137, 0.00809182, 0.056321923, 0.020589584, -0.018921316, 0.016269194, -0.02185861, -0.024582025, 0.006883395, -0.0032937631, -0.03151176, -0.030142924, -0.010843753, -0....
import requests from PIL import Image import torch from transformers import OwlViTProcessor, OwlViTForObjectDetection processor = OwlViTProcessor.from_pretrained("google/owlvit-base-patch32") model = OwlViTForObjectDetection.from_pretrained("google/owlvit-base-patch32") url = "http://images.cocodataset.org/val2017/00...
[ 0.0014540801, -0.0058642365, -0.031934474, -0.023545409, -0.052752327, 0.013313402, -0.032376777, 0.010490017, -0.026582574, 0.051572848, -0.030312685, -0.029722944, 0.041163918, -0.013328145, 0.0067672776, 0.02841077, 0.009384253, -0.014124296, -0.056467697, 0.009841302, -0....
Resources A demo notebook on using OWL-ViT for zero- and one-shot (image-guided) object detection can be found here. OwlViTConfig [[autodoc]] OwlViTConfig - from_text_vision_configs OwlViTTextConfig [[autodoc]] OwlViTTextConfig OwlViTVisionConfig [[autodoc]] OwlViTVisionConfig OwlViTImageProcessor [[autodoc]] Owl...
[ -0.0074474034, -0.009762486, -0.025536064, -0.008551407, -0.015271423, 0.010124333, -0.008514483, -0.013248034, -0.021843748, 0.050569974, -0.010441872, -0.031576697, 0.033998854, -0.0240296, 0.0023575444, 0.00335447, -0.0072258646, -0.012657262, -0.04401242, 0.006158785, -0....
OWL-ViT Overview The OWL-ViT (short for Vision Transformer for Open-World Localization) was proposed in Simple Open-Vocabulary Object Detection with Vision Transformers by Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa D...
[ 0.014335808, -0.0051225172, -0.015291529, 0.029612856, -0.022213263, -0.003862704, -0.043297037, -0.0038047815, -0.018882722, 0.0655103, 0.034492824, 0.015682505, 0.0049306494, -0.049639545, -0.014183762, 0.029511493, -0.012076832, -0.054881528, -0.06452562, -0.003372173, 0.0...
import torch from transformers import AutoTokenizer, RwkvConfig, RwkvModel model = RwkvModel.from_pretrained("sgugger/rwkv-430M-pile") tokenizer = AutoTokenizer.from_pretrained("sgugger/rwkv-430M-pile") inputs = tokenizer("This is an example.", return_tensors="pt") Feed everything to the model outputs = model(inputs[...
[ 0.03574853, 0.026730618, -0.031195512, 0.040477794, -0.02811121, 0.007266469, -0.012036121, -0.002394006, -0.032811098, 0.10445503, 0.03563103, -0.0022214318, 0.0031063328, -0.017830202, -0.018182695, 0.017169282, -0.0057793944, -0.053754978, -0.020400455, 0.0027244668, 0.013...
RWKV Overview The RWKV model was proposed in this repo It suggests a tweak in the traditional Transformer attention to make it linear. This way, the model can be used as recurrent network: passing inputs for timestamp 0 and timestamp 1 together is the same as passing inputs at timestamp 0, then inputs at timestamp 1 a...
[ 0.0064903656, 0.006681467, 0.023229422, 0.019279996, -0.0004109562, -0.031906836, 0.010595504, -0.013419556, 0.0010829072, 0.082669, 0.014849277, 0.0034362841, 0.04764792, -0.019520642, -0.028764281, 0.016222375, 0.00056578364, -0.047591295, -0.044363808, -0.0014208733, -0.01...
If you want to make sure the model stops generating when '\n\n' is detected, we recommend using the following stopping criteria: thon from transformers import StoppingCriteria class RwkvStoppingCriteria(StoppingCriteria): def init(self, eos_sequence = [187,187], eos_token_id = 537): self.eos_sequence = e...
[ -0.006705131, 0.0018915774, 0.014904759, 0.041819107, -0.036430877, -0.07581054, -0.005016282, 0.019113477, -0.004356157, 0.043883257, 0.006182392, 0.0029906693, -0.011011025, -0.031712823, -0.014113949, -0.025051253, -0.0042187707, -0.04535765, -0.046563968, 0.013792263, -0....
output = model.generate(inputs["input_ids"], max_new_tokens=64, stopping_criteria = [RwkvStoppingCriteria()])
[ -0.007251733, 0.015085115, 0.0044303555, 0.01229773, -0.02160412, 0.009782285, -0.010469689, -0.015802734, -0.06532603, 0.055354893, 0.02500337, 0.028780315, -0.0067418455, -0.0026495263, -0.0129926875, 0.036319096, -0.007164863, -0.07124827, -0.03329754, -0.0032425066, 0.031...
RwkvConfig [[autodoc]] RwkvConfig RwkvModel [[autodoc]] RwkvModel - forward RwkvLMHeadModel [[autodoc]] RwkvForCausalLM - forward Rwkv attention and the recurrent formulas In a traditional auto-regressive Transformer, attention is written as $$O = \hbox{softmax}(QK^{T} / \sqrt{d}) V$$ with \(Q\), \(K\) and \(...
[ 0.0057211774, -0.0071559, -0.016224762, 0.01300815, -0.055801846, -0.01715999, 0.0031865009, -0.012901874, -0.029785547, 0.049538665, 0.016295614, 0.013603294, -0.020971745, -0.041376688, -0.037607443, -0.03182604, -0.029813888, -0.014184268, -0.070793815, 0.00679102, 0.02425...
ByT5's architecture is based on the T5v1.1 model, refer to T5v1.1's documentation page for the API reference. They only differ in how inputs should be prepared for the model, see the code examples below. Since ByT5 was pre-trained unsupervisedly, there's no real advantage to using a task prefix during single-task fine...
[ 0.036062505, 0.00618031, -0.014686259, 0.01839369, -0.01976793, -0.015705613, 0.03621352, 0.0067390674, -0.008139737, 0.043945517, 0.032377727, 0.0030297488, 0.030157799, -0.038146522, -0.012700405, -0.011212902, -0.007260071, -0.031864274, -0.043250844, -0.0016724969, 0.0033...
from transformers import T5ForConditionalGeneration import torch model = T5ForConditionalGeneration.from_pretrained("google/byt5-small") num_special_tokens = 3 Model has 3 special tokens which take up the input ids 0,1,2 of ByT5. => Need to shift utf-8 character encodings by 3 before passing ids to model. input_ids =...
[ 0.037462745, -0.0152765075, 0.0033196737, 0.023222694, -0.015682079, -0.033016484, 0.009005179, -0.046745814, -0.022847166, 0.057020277, 0.0080813775, -0.00399187, -0.01111565, -0.039205197, -0.011258351, -0.010109233, -0.016523262, -0.010920375, -0.03962579, 0.013639202, 0.0...
ByT5 Overview The ByT5 model was presented in ByT5: Towards a token-free future with pre-trained byte-to-byte models by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel. The abstract from the paper is the following: Most widely-used pre-trained language mode...
[ 0.03545003, 0.0231768, -0.035745446, 0.0021182736, -0.055592094, 0.0083321, 0.005565923, 0.0031220196, 0.015885374, 0.025553564, -0.011568258, -0.012514935, -0.020692613, -0.037759654, -0.03571859, 0.024788165, -0.039747003, -0.060157627, -0.035611168, -0.011749537, -0.027903...
For batched inference and training it is however recommended to make use of the tokenizer: thon
[ 0.030674303, -0.013267972, -0.0003168201, 0.021270286, -0.020306151, -0.023050226, -0.015248155, -0.011688274, -0.0075944117, 0.0457148, 0.032691568, 0.028019225, -0.0009001677, -0.02113679, -0.021848766, -0.03295856, -0.024919163, -0.018081225, -0.054080516, 0.008172892, 0.0...
Similar to T5, ByT5 was trained on the span-mask denoising task. However, since the model works directly on characters, the pretraining task is a bit different. Let's corrupt some characters of the input sentence "The dog chases a ball in the park." and ask ByT5 to predict them for us. thon
[ 0.028275026, 0.012174582, -0.005135233, 0.02617241, -0.03940565, -0.027775103, 0.029995348, 0.013394981, 0.006826147, 0.026745852, 0.014931507, -0.0063482802, 0.019496975, -0.035465088, -0.00988082, -0.0057160254, -0.012483357, -0.03764122, -0.05087446, 0.015659336, -0.007421...
from transformers import T5ForConditionalGeneration, AutoTokenizer model = T5ForConditionalGeneration.from_pretrained("google/byt5-small") tokenizer = AutoTokenizer.from_pretrained("google/byt5-small") model_inputs = tokenizer( ["Life is like a box of chocolates.", "Today is Monday."], padding="longest", return_...
[ -0.000152483, -0.048048206, 0.0023094683, 0.028122002, -0.0017444231, -0.030811688, -0.010857318, 0.024897195, 0.004330253, 0.050414003, 0.0043795407, 0.004946346, -0.004052131, -0.012230325, 0.00023213502, 0.00782262, -0.005643411, -0.03912014, -0.06421448, 0.02048949, 0.006...
ByT5Tokenizer [[autodoc]] ByT5Tokenizer See [ByT5Tokenizer] for all details.
[ 0.030017095, 0.002239737, -0.009282209, 0.035651073, -0.024029069, -0.01939566, 0.0047757966, -0.01788711, 0.0036328381, 0.04328619, 0.017179014, 0.035589498, -0.021720061, -0.040884823, -0.038606603, -0.004760403, -0.013730897, -0.041716065, -0.031371713, -0.0059687835, 0.01...
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM import torch tokenizer = AutoTokenizer.from_pretrained("google/byt5-base") model = AutoModelForSeq2SeqLM.from_pretrained("google/byt5-base") input_ids_prompt = "The dog chases a ball in the park." input_ids = tokenizer(input_ids_prompt).input_ids Note that...
[ 0.04145663, -0.017457658, -0.036925852, 0.0044883033, -0.013394116, 0.00084288424, 0.011022536, 0.0011043774, -0.026986454, 0.037860323, 0.06496005, 0.0006738649, -0.006297075, -0.040210664, 0.009557112, 0.0025361744, -0.042164564, -0.0495554, -0.01992127, -0.0036228537, 0.05...
FLAN-T5 Overview FLAN-T5 was released in the paper Scaling Instruction-Finetuned Language Models - it is an enhanced version of T5 that has been finetuned in a mixture of tasks. One can directly use FLAN-T5 weights without finetuning the model: thon
[ 0.020765394, 0.017082108, -0.006935697, 0.012106197, -0.023461837, -0.01359341, -0.013468318, 0.01998704, 0.022155313, 0.061323237, 0.016484443, 0.004788272, 0.0022099717, -0.06610456, -0.025101943, 0.04066904, -0.014330068, -0.08100449, -0.044672005, -0.014969431, 0.01972295...
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer model = AutoModelForSeq2SeqLM.from_pretrained("google/flan-t5-small") tokenizer = AutoTokenizer.from_pretrained("google/flan-t5-small") inputs = tokenizer("A step by step recipe to make bolognese pasta:", return_tensors="pt") outputs = model.generate(**input...
[ -0.0011318624, -0.04577171, 0.008053468, 0.025280885, 0.0034104686, -0.03546327, -0.009257986, 0.030589173, 0.0065828348, 0.027437815, 0.04305454, 0.023796247, -0.009832234, -0.022059498, -0.006442774, -0.007822368, 0.010609568, -0.010455502, -0.038096406, 0.010623574, 0.0298...
CTRL
[ 0.041994683, -0.002629896, -0.0036529237, 0.037756175, -0.016145365, -0.0024103024, 0.025012769, 0.010059483, 0.022907458, 0.028205592, 0.057498693, 0.057721775, 0.0018421472, -0.03680809, -0.0369754, 0.0145559255, -0.013872745, -0.0074243587, -0.014026112, 0.011411902, 0.022...
FLAN-T5 includes the same improvements as T5 version 1.1 (see here for the full details of the model's improvements.) Google has released the following variants: google/flan-t5-small google/flan-t5-base google/flan-t5-large google/flan-t5-xl google/flan-t5-xxl. The original checkpoints can be found here. Refer t...
[ 0.01759949, -0.01651668, 0.015998151, 0.019826116, 0.010568845, -0.025972212, -0.027802315, -0.003725022, -0.022708528, 0.046454113, -0.020969931, 0.022800034, 0.0098520545, -0.026155222, -0.0070878365, -0.045600068, 0.020176886, 0.016928453, -0.036846075, -0.0015460558, 0.05...
CTRL makes use of control codes to generate text: it requires generations to be started by certain words, sentences or links to generate coherent text. Refer to the original implementation for more information. CTRL is a model with absolute position embeddings so it's usually advised to pad the inputs on the righ...
[ -0.014637305, -0.04074709, -0.017703228, 0.02353837, 0.020599606, -0.008272338, -0.021136494, -0.016375132, -0.018042315, 0.047641885, -0.004114976, 0.0021104706, 0.01368362, -0.046115987, 0.0067323125, 0.0022340964, -0.01511768, -0.025205553, -0.06651779, -0.013839035, 0.007...
Resources Text classification task guide Causal language modeling task guide CTRLConfig [[autodoc]] CTRLConfig CTRLTokenizer [[autodoc]] CTRLTokenizer - save_vocabulary CTRLModel [[autodoc]] CTRLModel - forward CTRLLMHeadModel [[autodoc]] CTRLLMHeadModel - forward CTRLForSequenceClassification [[autodoc]...
[ 0.0037351185, -0.023847584, 0.008952248, 0.0129469065, 0.010772789, -0.04035282, 0.0057249246, -0.014285982, -0.025036203, 0.03556826, -0.010893155, 0.011043613, 0.016490191, -0.04919975, 0.004461078, -0.012224708, 0.0019898063, 0.0074326224, -0.053713486, -0.0031784242, 0.03...
Overview CTRL model was proposed in CTRL: A Conditional Transformer Language Model for Controllable Generation by Nitish Shirish Keskar, Bryan McCann, Lav R. Varshney, Caiming Xiong and Richard Socher. It's a causal (unidirectional) transformer pre-trained using language modeling on a very large corpus of ~140 GB of ...
[ -0.006049923, -0.038004518, 0.013069209, 0.008985511, 0.04394444, 0.0029459002, 0.0014127946, 0.00093498814, 0.006342107, 0.061159223, 0.022728462, 0.03066211, 0.020418491, -0.03701453, -0.032037094, 0.034154568, 0.009796751, -0.049801867, -0.047134403, -0.01182485, 0.0262346...
CTRLModel [[autodoc]] CTRLModel - forward CTRLLMHeadModel [[autodoc]] CTRLLMHeadModel - forward CTRLForSequenceClassification [[autodoc]] CTRLForSequenceClassification - forward TFCTRLModel [[autodoc]] TFCTRLModel - call TFCTRLLMHeadModel [[autodoc]] TFCTRLLMHeadModel - call TFCTRLForSequenceClassi...
[ 0.023110589, -0.01588761, -0.0117686, -0.020889265, -0.0044904565, -0.061608624, 0.012813063, -0.01499761, -0.0021459307, 0.06460962, -0.02706778, -0.020315547, 0.029877534, -0.028391749, 0.009709096, -0.022845795, -0.017520504, -0.020815711, -0.07490714, -0.0015924387, 0.032...
GIT Overview The GIT model was proposed in GIT: A Generative Image-to-text Transformer for Vision and Language by Jianfeng Wang, Zhengyuan Yang, Xiaowei Hu, Linjie Li, Kevin Lin, Zhe Gan, Zicheng Liu, Ce Liu, Lijuan Wang. GIT is a decoder-only Transformer that leverages CLIP's vision encoder to condition the model on ...
[ 0.050753754, 0.0072821807, -0.014315162, -0.045409795, 0.0024850788, -0.050393797, 0.030568546, -0.01916072, -0.013671394, 0.08345435, -0.033586636, -0.012480772, 0.036826238, -0.06362909, -0.017997785, 0.008424348, -0.020711297, -0.009407304, -0.056568425, 0.042751662, 0.061...
GIT architecture. Taken from the original paper. This model was contributed by nielsr. The original code can be found here. Usage tips GIT is implemented in a very similar way to GPT-2, the only difference being that the model is also conditioned on pixel_values. Resources A list of official Hugging Face and communi...
[ 0.0074810353, -0.00991752, -0.021550873, -0.013822758, -0.029594703, -0.043486092, 0.037034556, 0.013767851, -0.004564118, 0.05644407, -0.021386152, 0.017336786, 0.030253582, -0.05224371, -0.008455629, 0.010116557, 0.003757676, -0.03264202, -0.041976158, 0.0028156834, -0.0042...
Resources A list of official Hugging Face and community (indicated by 🌎) resources to help you get started with GIT. Demo notebooks regarding inference + fine-tuning GIT on custom data can be found here. See also: Causal language modeling task guide
[ 0.013595265, -0.009666596, -0.025833588, 0.016890695, -0.045231394, -0.0366762, -0.028586242, 0.03897654, -0.0023665712, 0.06632215, 0.010584147, 0.010577685, 0.03135182, -0.02190492, -0.011152769, 0.030989967, 0.012380479, -0.028146852, -0.03941593, -0.015288211, 0.012134938...
If you're interested in submitting a resource to be included here, please feel free to open a Pull Request and we will review it. The resource should ideally demonstrate something new instead of duplicating an existing resource. GitVisionConfig [[autodoc]] GitVisionConfig GitVisionModel [[autodoc]] GitVisionModel ...
[ 0.00016472864, 0.0051608067, -0.008678702, 0.01541243, -0.0017101389, -0.038162712, -0.033418156, -0.050863966, 0.0030648045, 0.04838855, -0.015206146, -0.021055797, 0.0019210284, -0.0402845, 0.019950699, -0.016723813, -0.021262081, -0.036453493, -0.05929218, -0.028879888, 0....
CLAP Overview The CLAP model was proposed in Large Scale Contrastive Language-Audio pretraining with feature fusion and keyword-to-caption augmentation by Yusong Wu, Ke Chen, Tianyu Zhang, Yuchen Hui, Taylor Berg-Kirkpatrick, Shlomo Dubnov. CLAP (Contrastive Language-Audio Pretraining) is a neural network trained on a...
[ 0.01960428, 0.03651184, 0.008824957, 0.05332849, -0.035087723, -0.051934678, -0.04223859, 0.009097659, -0.029088268, 0.010483896, -0.017998366, 0.0044995914, 0.022543408, -0.016771205, -0.022922162, -0.007378118, 0.0032515987, -0.024361424, -0.048298646, -0.0053896625, -0.042...
CLVP is an integral part of the Tortoise TTS model. CLVP can be used to compare different generated speech candidates with the provided text, and the best speech tokens are forwarded to the diffusion model. The use of the [ClvpModelForConditionalGeneration.generate()] method is strongly recommended for tortoise usage. ...
[ 0.026741587, 0.030573802, -0.009748371, 0.005549019, 0.002201076, -0.039357126, -0.05863009, -0.014377799, -0.003954594, 0.023566723, 0.03890957, -0.020251997, 0.03409832, -0.0622665, -0.011937209, -0.03160878, 0.0005345345, -0.06573507, -0.03583261, -0.019147089, -0.02836398...
The [ClvpTokenizer] tokenizes the text input, and the [ClvpFeatureExtractor] extracts the log mel-spectrogram from the desired audio. [ClvpConditioningEncoder] takes those text tokens and audio representations and converts them into embeddings conditioned on the text and audio. The [ClvpForCausalLM] uses those embedd...
[ 0.0019744847, 0.019863814, 0.0028072605, 0.00933323, -0.012764112, -0.0543415, -0.038346063, -0.024300167, -0.033525944, 0.016394554, -0.027262853, -0.03521452, 0.041231997, -0.028352754, 0.0014516012, 0.00445938, -0.000797757, -0.016839724, -0.044210035, 0.012672008, -0.0384...
CLVP Overview The CLVP (Contrastive Language-Voice Pretrained Transformer) model was proposed in Better speech synthesis through scaling by James Betker. The abstract from the paper is the following: In recent years, the field of image generation has been revolutionized by the application of autoregressive transformer...
[ 0.020842394, -0.007600846, 0.016809292, -0.00061210425, -0.050512485, -0.014679926, 0.017570786, 0.012339035, 0.0095962435, 0.029782906, 0.033703193, 0.03502876, -0.008891156, 0.005584295, -0.0667295, 0.0273292, -0.033703193, -0.034944147, -0.06616543, -0.024607562, 0.0111685...
Example : thon
[ 0.013549056, -0.009931169, -0.012222978, 0.016374178, -0.030470962, -0.017858809, -0.02608914, 0.012460807, -0.0021710922, 0.022514496, 0.029634956, 0.0065042675, 0.0073402734, -0.055983547, -0.002807105, 0.027472872, -0.0009369028, -0.035746444, -0.06653451, -0.014075163, -0...
ClvpConfig [[autodoc]] ClvpConfig - from_sub_model_configs ClvpEncoderConfig [[autodoc]] ClvpEncoderConfig ClvpDecoderConfig [[autodoc]] ClvpDecoderConfig ClvpTokenizer [[autodoc]] ClvpTokenizer - save_vocabulary ClvpFeatureExtractor [[autodoc]] ClvpFeatureExtractor - call ClvpProcessor [[autodoc]] ClvpPr...
[ -0.00014486792, 0.027019508, 0.011116855, 0.005472163, -0.004954577, -0.040776804, -0.008641442, -0.034385737, -0.005603435, 0.00055556145, 0.000799821, -0.016277716, 0.056169372, -0.04452743, -0.0069199046, 0.016742794, -0.022458747, -0.06409069, -0.071111865, -0.02499417, -...
import datasets from transformers import ClvpProcessor, ClvpModelForConditionalGeneration Define the Text and Load the Audio (We are taking an audio example from HuggingFace Hub using datasets library). text = "This is an example text." ds = datasets.load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", ...
[ 0.010727232, 0.0046743746, -0.039037444, -0.03431541, -0.035576575, -0.011651108, -0.021674434, -0.0053379526, -0.03235034, 0.07074254, 0.02165977, 0.011357814, 0.015163305, 0.008380879, 0.010382611, -0.008036259, 0.024299417, -0.010030658, -0.028420199, -0.020046651, 0.02363...
DETR uses so-called object queries to detect objects in an image. The number of queries determines the maximum number of objects that can be detected in a single image, and is set to 100 by default (see parameter num_queries of [~transformers.DetrConfig]). Note that it's good to have some slack (in COCO, the au...
[ -0.008139617, -0.010228856, -0.014806951, -0.07156693, -0.044252593, 0.013601082, -0.0136922235, 0.021158798, -0.015732385, 0.053983677, 0.027650861, -0.056984328, 0.0392889, -0.0023013172, 0.008854725, 0.04683259, -0.025183035, -0.024285644, -0.04262607, -0.031633034, 0.0257...
There are three ways to instantiate a DETR model (depending on what you prefer): Option 1: Instantiate DETR with pre-trained weights for entire model from transformers import DetrForObjectDetection model = DetrForObjectDetection.from_pretrained("facebook/detr-resnet-50") Option 2: Instantiate DETR with randomly initi...
[ 0.01711811, -0.011238411, -0.00021804657, -0.037034657, -0.0100847995, -0.024903128, -0.008172036, -0.012027333, 0.015748661, 0.049627632, 0.02283407, 0.013389339, 0.013821012, -0.0044358233, -0.017698636, 0.025870673, -0.004956809, -0.010531358, -0.060166433, -0.008566497, 0...
As a summary, consider the following table: | Task | Object detection | Instance segmentation | Panoptic segmentation | |------|------------------|-----------------------|-----------------------| | Description | Predicting bounding boxes and class labels around objects in an image | Predicting masks around objects (i...
[ -0.0015605353, -0.0040799975, -0.0355277, -0.03083151, -0.0025942074, -0.030685665, -0.032727487, -0.011456079, 0.005888468, 0.069888644, 0.0060525434, -0.012972861, 0.020870335, 0.003729971, -0.004069059, -0.00659946, 0.005647825, -0.009953882, -0.05720018, -0.0053707208, 0....
DETR Overview The DETR model was proposed in End-to-End Object Detection with Transformers by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov and Sergey Zagoruyko. DETR consists of a convolutional backbone followed by an encoder-decoder Transformer which can be trained end-to-end...
[ -0.011569209, -0.008598831, -0.009237323, -0.058796845, -0.059407577, 0.009195683, -0.025081657, 0.005090591, -0.036616165, 0.043001093, 0.011291604, -0.06046248, 0.013470807, -0.009514929, 0.008480848, 0.030897493, -0.046026994, -0.018308083, -0.041418742, -0.016892295, 0.02...
Option 2: Instantiate DETR with randomly initialized weights for Transformer, but pre-trained weights for backbone from transformers import DetrConfig, DetrForObjectDetection config = DetrConfig() model = DetrForObjectDetection(config) Option 3: Instantiate DETR with randomly initialized weights for backbone + Transfo...
[ -0.0035394584, -0.026302705, 0.0041131526, -0.01219855, -0.043802053, -0.035454974, -0.025645137, 0.014560426, 0.0031972548, 0.05426946, 0.011480594, 0.008447731, 0.05461837, 0.0008118947, 0.004596263, 0.009360273, 0.016251314, -0.023484558, -0.055987187, -0.013708272, -0.009...
All example notebooks illustrating fine-tuning [DetrForObjectDetection] and [DetrForSegmentation] on a custom dataset an be found here. See also: Object detection task guide
[ 0.012248011, -0.0054443157, -0.0197246, -0.003357668, -0.060519584, -0.038497634, -0.046137348, -0.0048733763, -0.0077552614, 0.06775148, -0.00056966505, -0.008332998, 0.033957306, -0.013301531, -0.0056924024, 0.017413653, 0.03278824, -0.026045715, -0.05173799, -0.023027893, ...
If you're interested in submitting a resource to be included here, please feel free to open a Pull Request and we'll review it! The resource should ideally demonstrate something new instead of duplicating an existing resource. DetrConfig [[autodoc]] DetrConfig DetrImageProcessor [[autodoc]] DetrImageProcessor - p...
[ -0.0034928306, -0.0016713358, 0.00786703, 0.01816272, 0.010667823, -0.03407958, -0.016399981, -0.04345473, -0.00046271845, 0.0655477, 0.037683398, 0.00820652, 0.033139456, -0.017248707, -0.012064955, 0.024586916, -0.029405063, -0.026924174, -0.053639434, 0.046980206, 0.013227...
XLNet
[ 0.018734036, -0.0071976725, -0.029433459, 0.010337867, -0.0077333134, 0.000023342744, -0.008717554, -0.02023383, -0.0031234552, 0.05217141, -0.008315823, -0.009782139, 0.022403175, -0.04716317, 0.009983005, -0.016765555, -0.030799344, -0.018600125, -0.055438817, -0.036316443, ...
Resources Text classification task guide Token classification task guide Question answering task guide Causal language modeling task guide Multiple choice task guide
[ 0.04340444, 0.0475325, 0.01249635, 0.022494927, -0.013947153, -0.006001389, -0.021358214, -0.03649443, -0.03434066, 0.05246822, 0.022509884, 0.03634486, 0.014208896, -0.020475768, -0.021432998, 0.014724904, -0.011875646, -0.023751292, -0.07693744, 0.031050177, 0.008704819, ...
The specific attention pattern can be controlled at training and test time using the perm_mask input. Due to the difficulty of training a fully auto-regressive model over various factorization order, XLNet is pretrained using only a sub-set of the output tokens as target which are selected with the target_mapping i...
[ 0.023134286, 0.007455801, -0.02037919, 0.008052265, -0.005339774, -0.032351077, -0.010580136, -0.04476321, -0.013981402, 0.0846979, 0.045047242, -0.0007904037, 0.021444304, -0.035304993, 0.0033994904, 0.0026024296, -0.028232634, -0.030675296, -0.043229446, 0.0069836006, -0.00...
Overview The XLNet model was proposed in XLNet: Generalized Autoregressive Pretraining for Language Understanding by Zhilin Yang, Zihang Dai, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le. XLnet is an extension of the Transformer-XL model pre-trained using an autoregressive method to learn bidirectio...
[ 0.010784725, -0.025921298, 0.009788273, 0.017773442, 0.014316366, 0.0039620814, -0.0023216645, -0.010337338, 0.023291208, 0.08096339, 0.03988518, 0.029337704, 0.04972768, -0.009354443, -0.033404853, 0.066538565, -0.020810248, -0.054093085, -0.046609532, 0.0033333679, 0.019929...
XLNetModel [[autodoc]] XLNetModel - forward XLNetLMHeadModel [[autodoc]] XLNetLMHeadModel - forward XLNetForSequenceClassification [[autodoc]] XLNetForSequenceClassification - forward XLNetForMultipleChoice [[autodoc]] XLNetForMultipleChoice - forward XLNetForTokenClassification [[autodoc]] XLNetForTo...