51 Commits (3fa2e44e8922b0ae8e9d6e45d8170aa5814c080b)

Author	SHA1	Message	Date
Hui Zhang	6a9daa800d	encoder.forward_one_step for prefix feat or ys	4 years ago
Hui Zhang	0cd30e48a9	fix no_pos call	4 years ago
Hui Zhang	12ea02fc48	fix no pos	4 years ago
Hui Zhang	e76d51fda0	linear embed with relu	4 years ago
Hui Zhang	4f1a61011d	add none pos embed	4 years ago
Hui Zhang	4566351127	add transformer lm and encoder score api	4 years ago
Hui Zhang	12f788dd0e	Merge branch 'develop' into join_ctc	4 years ago
Hui Zhang	dfd80b3aa2	recog into decoders, format code	4 years ago
Hui Zhang	a4e27da64b	decoder with ctc prefix score	4 years ago
Hui Zhang	30499a7654	not change ctc grad manual	4 years ago
Hui Zhang	f2f305cd66	add recog interface	4 years ago
huangyuxin	b464e20a40	fix the bug of import ctc decoder	4 years ago
Hui Zhang	10a2da6808	format code	4 years ago
Hui Zhang	69bd17dcb2	refactor raw ctc decoder into ctcdecoder	4 years ago
Hui Zhang	8b45c3e65e	refactor trainer.py and rm ueseless dir setup code	4 years ago
Hui Zhang	2691013232	ctcloss can work w/ paddle2.1.2, but loss larger than before	4 years ago
Hui Zhang	913ea33f29	remove conv and rnn uesless	4 years ago
Hui Zhang	c8e96d732b	bool logical, sum and multiply op; ctc grad norm; support old and new pd api	4 years ago
Hui Zhang	f7d7e70cb2	more ctc check; valid dataloader with num workers	4 years ago
Hui Zhang	0c08915207	ctc using nn.Dropout; ds2 libri vald batch_size / 4	4 years ago
Hui Zhang	000183ea49	tensor.size to tensor.shape	4 years ago
Hui Zhang	cd001d5daf	change type_as to astype	4 years ago
Junkun	064d42137c	fix norm_by_batchsize	4 years ago
Hui Zhang	8873ebe38c	add timer for u2; refactor grad norm type	5 years ago
Jackwaterveg	59766313b1	Merge pull request #817 from PaddlePaddle/ctc ... export ctc grad norm config	5 years ago
Hui Zhang	1a8c5278a1	export ctc grad norm config	5 years ago
Hui Zhang	6e4a3aff07	float_mul_bool type promote, lhs type promote to rhs type, https://github.com/PaddlePaddle/Paddle/pull/29265	5 years ago
Hui Zhang	244132c1c4	fix activation	5 years ago
huangyuxin	04d9db199f	add blank_id parameter	5 years ago
Hui Zhang	f54dc983b6	using bw rnn in ds2	5 years ago
Hui Zhang	797ca389fc	paddle support some bool op	5 years ago
Hui Zhang	e7b71d7860	import ctcdecoder when needed	5 years ago
Hui Zhang	341038b626	ds2 offline cer 6p4287	5 years ago
Hui Zhang	ab23eb5710	fix for kaldi	5 years ago
Hui Zhang	f05f367cc5	Merge pull request #756 from PaddlePaddle/filter ... test w/ all example & fix ctc api & add new io	5 years ago
Hui Zhang	7d133368e5	fix bugs	5 years ago
Hui Zhang	4b5410eecd	remove fixed hack api	5 years ago
huangyuxin	2537221b61	Complete the modification according to the comments ... , but still need to be unit tested	5 years ago
Hui Zhang	259781768e	comment u2 model for easy understand	5 years ago
Hui Zhang	1cd88d2619	Merge pull request #657 from PaddlePaddle/add_utt ... add utt to datapipeline	5 years ago
Haoxin Ma	8781ab58cf	fix export and run.sh	5 years ago
Haoxin Ma	a58b1cb30a	add result output	5 years ago
Hui Zhang	69dfc2a5fa	fix mask for bool type; fix other	5 years ago
Hui Zhang	4acaaba349	replace list zip by stack	5 years ago
Hui Zhang	34689bd1df	add crf	5 years ago
Hui Zhang	b3bc451328	remove sequnce_mask and change ds2 export audio shape to [B,T,D] (#639) ... * remove sequnce_mask * format * fix ds2 export audio shape from B,D,T to B,T,D	5 years ago
Hui Zhang	71e046b0ba	E2E/Streaming Transformer/Conformer ASR (#578) ... * add cmvn and label smoothing loss layer * add layer for transformer * add glu and conformer conv * add torch compatiable hack, mask funcs * not hack size since it exists * add test; attention * add attention, common utils, hack paddle * add audio utils * conformer batch padding mask bug fix #223 * fix typo, python infer fix rnn mem opt name error and batchnorm1d, will be available at 2.0.2 * fix ci * fix ci * add encoder * refactor egs * add decoder * refactor ctc, add ctc align, refactor ckpt, add warmup lr scheduler, cmvn utils * refactor docs * add fix * fix readme * fix bugs, refactor collator, add pad_sequence, fix ckpt bugs * fix docstring * refactor data feed order * add u2 model * refactor cmvn, test * add utils * add u2 config * fix bugs * fix bugs * fix autograd maybe has problem when using inplace operation * refactor data, build vocab; add format data * fix text featurizer * refactor build vocab * add fbank, refactor feature of speech * refactor audio feat * refactor data preprare * refactor data * model init from config * add u2 bins * flake8 * can train * fix bugs, add coverage, add scripts * test can run * fix data * speed perturb with sox * add spec aug * fix for train * fix train logitc * fix logger * log valid loss, time dataset process * using np for speed perturb, remove some debug log of grad clip * fix logger * fix build vocab * fix logger name * using module logger as default * fix * fix install * reorder imports * fix board logger * fix logger * kaldi fbank and mfcc * fix cmvn and print prarams * fix add_eos_sos and cmvn * fix cmvn compute * fix logger and cmvn * fix subsampling, label smoothing loss, remove useless * add notebook test * fix log * fix tb logger * multi gpu valid * fix log * fix log * fix config * fix compute cmvn, need paddle 2.1 * add cmvn notebook * fix layer tools * fix compute cmvn * add rtf * fix decoding * fix layer tools * fix log, add avg script * more avg and test info * fix dataset pickle problem; using 2.1 paddle; num_workers can > 0; ckpt save in exp dir;fix setup.sh; * add vimrc * refactor tiny script, add transformer and stream conf * spm demo; librisppech scripts and confs * fix log * add librispeech scripts * refactor data pipe; fix conf; fix u2 default params * fix bugs * refactor aishell scripts * fix test * fix cmvn * fix s0 scripts * fix ds2 scripts and bugs * fix dev & test dataset filter * fix dataset filter * filter dev * fix ckpt path * filter test, since librispeech will cause OOM, but all test wer will be worse, since mismatch train with test * add comment * add syllable doc * fix ds2 configs * add doc * add pypinyin tools * fix decoder using blank_id=0 * mmseg with pybind11 * format code	5 years ago
Hui Zhang	e0a87a5ab1	batch average ctc loss (#567) ... * when loss div batchsize, change lr, more epoch, loss can reduce more and cer lower than before * since loss reduce more when loss div batchsize, less lm alpha can be better. * less lm alpha, more cer reduce * alpha 2.2, cer 0.077478 * alpha 1.9, cer 0.077249 * large librispeech lr for batch_average ctc loss * since loss reduce and model more confidence, then less lm alpha	5 years ago
Hui Zhang	258307df9b	fix egs bugs (#552) ... * fix egs * fix log	5 years ago
Hui Zhang	1539f3e0a3	Refactor CTC module, add embedding and fix log (#549) ... * add acts, refactor ctc, add pos embed * fix export, dataloader time log * fix egs * fix libri readme	5 years ago