A freely available set of articulatory feature MLPstrained on 2000 hours of conversational telephone speech. Project Details Articulatory feature MLPs. Background. This page gives information on a set of articulatory feature (AF)classification multi-layer perceptrons (MLPs) which were trained aspart of the Johns Hopkins 2006 summer workshop. The resources requiredto generate compatible front-end parameters are given along with theMLP weights. As a starting point, the work is summarised in a paperwhich was presented at Interspeech 2007:J. Frankel, M. Magimai-Doss, S. King, K. Livescu andO. Cetin. Articulatory feature classifiers trained on 2000 hours oftelephone speech. Proc Interspeech 2007. pdfA number of related papers came out of the workshop, and theseinclude:K. Livescu, A. Bezman, N. Borges, L. Yung, O. Cetin, J. Frankel,S. King, M. Magimai-Doss, X. Chi and L. Lavoie. Manualtranscription of conversational speech at the articulatory featurelevel. Proc. ICASSP 2007. pdfK. Livescu, O. Cetin, M. Hasegawa-Johnson, S. King, C. Bartels,N. Borges, A. Kantor, P. Lal, L. Yung, A. Bezman, S. Dawson-Haggerty,B. Woods, J. Frankel, M. Magimai-Doss and K. Saenko. Articulatoryfeature-based methods for acoustic and audio-visual speechrecognition: Summary from the 2006 JHU Summer WorkshopProc. ICASSP 2007. pdfO. Cetin, A. Kantor, S. King, C. Bartels, M. Magimai-Doss, J. Frankeland K. Livescu. An articulatory feature-based tandem approach andfactored observation modeling. Proc ICASSP 2007. pdf Acoustic parameterization. The input to the MLPs are PLP cepstra computed using HTK. In order to computecompatible front-end features, this HCopy config file should be used. Oncethe base parameters (12 PLP cepstra plus energy) have been computed,these should be mean and variance normalized on a per-speaker basis,then expanded to include first a second order derivatives, and scaledagainst this global variance. Ascript which takes care of this process is available on request.MLP weights. Links to the MLP weight files are given in the table below. Feature abbreviation units place pl1 351, 1900, 10 degree dg1 351, 1600, 6 nasality nas 351, 1200, 3 rounding rou 351, 1200, 3 glottal state glo 351, 1400, 4 vowel vow 351, 2400, 23 height ht 351, 1800, 8 frontness frt 351, 1700, 7Feature groups, abbreviated names which link to weightfiles, and units (INPUT, HIDDEN, OUTPUT)They are all in matlab binary format, as used by quicknet. Therefore,if you wish to inspect or manipulate the weights, fire up matlab, andload (e.g. the place MLP weights) using:>> weights = load('pl1_win9_idim39_size351,1900,10_lr0.0001.wts', '-mat')weights = weights12: [1900x351 double] bias2: [1x1900 double] weights23: [10x1900 double]] bias3: [1x10 double]>>The weights for the input to hidden layer are in weights12, the biason layer 2 is in bias2 and so on. Should you wish to manipulateweights and then save them again, this can be achieved in matlab as:>> save -v4 my_new_weights_file weights12 bias2 weights23 bias3 MLP forward passThe QuicknetMLP toolkit was used to train the MLPs, and supports a forward pass inorder to generate activations given acoustic input. The HTK featuresare not normalized to have zero mean and unit variance, so the norm file WS06_AFMLP_PLP.norm is supplied to quicknet.Below is an example call to qnmultifwd. Here we assume the input PLPs(input.spknorm-plp.pfile) are in pfile format, and that pfile is alsothe format we would like the outputs written to. A log file is writtento out.log. qnmultifwd \ftr1_ftr_count=39 ftr1_window_len=9 window_extent=9 \mlp_size=351,1900,10 mlp_output_type=softmax mlp_bunch_size=256 \init_weight_format=matlab activation_format=pfile \ftr1_norm_file=WS06_AFMLP_PLP.norm \init_weight_file=pl1_win9_idim39_size351,1900,10_lr0.0001.wts \log_file=out.log ftr1_file=input.spknorm-plp.pfile \activation_file=out.act.pfile \Note that because of the 9-frame input windows on the MLPs, 4 framesat either end of each utterance are lost. If a set of activations arerequired with numbers of frames matching those of the inputs, then thePLPs for each utterance should be padded prior to runningqnmultifwd. This can be done using feacat, e.g.: feacat -ippfile -op pfile -i plp.pfile -o plp.pad4.pfile -pad 4Personnel Joe Frankel
Matlab R2007b Plp Serial Number
Hai Ilham MutaqinCoba "run as administrator" software MATLAB, setelah jendela aktivasi muncul, pilih file .datnyaKalo gagal, berarti file .dat tidak cocok dengan serial number yang dimasukkan di awal atau dapat juga file.dat berasal dari versi software yang berbeda. Namun, jika file .dat berasal dari web yang sama, kemungkinan itu kecil terjadi.Semoga membantu ? 2ff7e9595c
Comments