improve parameters tuning

Author: Yibing Liu

deep_speech_2/evaluate.py

@@ -12,6 +12,7 @@
 from decoder import *
 from lm.lm_scorer import LmScorer
 from error_rate import wer
+import utils
 
 parser = argparse.ArgumentParser(description=__doc__)
 parser.add_argument(
@@ -67,12 +68,12 @@
  help="Path for language model. (default: %(default)s)")
 parser.add_argument(
  "--alpha",
- default=0.26,
+ default=0.34,
  type=float,
  help="Parameter associated with language model. (default: %(default)f)")
 parser.add_argument(
  "--beta",
- default=0.1,
+ default=0.35,
  type=float,
  help="Parameter associated with word count. (default: %(default)f)")
 parser.add_argument(
@@ -192,11 +193,12 @@ def evaluate():
  else:
  raise ValueError("Decoding method [%s] is not supported." %
  decode_method)
-
+ print("Cur WER = %f" % (wer_sum / wer_counter))
  print("Final WER = %f" % (wer_sum / wer_counter))
 
 
 def main():
+ utils.print_arguments(args)
  paddle.init(use_gpu=args.use_gpu, trainer_count=1)
  evaluate()
 

deep_speech_2/infer.py

@@ -94,12 +94,12 @@
  help="Path for language model. (default: %(default)s)")
 parser.add_argument(
  "--alpha",
- default=0.26,
+ default=0.34,
  type=float,
  help="Parameter associated with language model. (default: %(default)f)")
 parser.add_argument(
  "--beta",
- default=0.1,
+ default=0.35,
  type=float,
  help="Parameter associated with word count. (default: %(default)f)")
 parser.add_argument(

deep_speech_2/tune.py

@@ -16,10 +16,10 @@
 
 parser = argparse.ArgumentParser(description=__doc__)
 parser.add_argument(
- "--num_samples",
+ "--batch_size",
  default=100,
  type=int,
- help="Number of samples for parameters tuning. (default: %(default)s)")
+ help="Minibatch size for evaluation. (default: %(default)s)")
 parser.add_argument(
  "--num_conv_layers",
  default=2,
@@ -57,7 +57,7 @@
  help="Manifest path for normalizer. (default: %(default)s)")
 parser.add_argument(
  "--decode_manifest_path",
- default='datasets/manifest.test',
+ default='datasets/manifest.dev',
  type=str,
  help="Manifest path for decoding. (default: %(default)s)")
 parser.add_argument(
@@ -82,17 +82,17 @@
  help="Path for language model. (default: %(default)s)")
 parser.add_argument(
  "--alpha_from",
- default=0.1,
+ default=0.22,
  type=float,
  help="Where alpha starts from. (default: %(default)f)")
 parser.add_argument(
  "--num_alphas",
- default=14,
+ default=10,
  type=int,
  help="Number of candidate alphas. (default: %(default)d)")
 parser.add_argument(
  "--alpha_to",
- default=0.36,
+ default=0.40,
  type=float,
  help="Where alpha ends with. (default: %(default)f)")
 parser.add_argument(
@@ -102,12 +102,12 @@
  help="Where beta starts from. (default: %(default)f)")
 parser.add_argument(
  "--num_betas",
- default=20,
+ default=7,
  type=float,
  help="Number of candidate betas. (default: %(default)d)")
 parser.add_argument(
  "--beta_to",
- default=1.0,
+ default=0.35,
  type=float,
  help="Where beta ends with. (default: %(default)f)")
 parser.add_argument(
@@ -160,55 +160,77 @@ def tune():
  # prepare infer data
  batch_reader = data_generator.batch_reader_creator(
  manifest_path=args.decode_manifest_path,
- batch_size=args.num_samples,
+ batch_size=args.batch_size,
+ min_batch_size=1,
  sortagrad=False,
  shuffle_method=None)
- # get one batch data for tuning
- infer_data = batch_reader().next()
-
- # run inference
- infer_results = paddle.infer(
- output_layer=output_probs, parameters=parameters, input=infer_data)
- num_steps = len(infer_results) // len(infer_data)
- probs_split = [
- infer_results[i * num_steps:(i + 1) * num_steps]
- for i in xrange(0, len(infer_data))
- ]
+
+ # define inferer
+ inferer = paddle.inference.Inference(
+ output_layer=output_probs, parameters=parameters)
 
  # create grid for search
  cand_alphas = np.linspace(args.alpha_from, args.alpha_to, args.num_alphas)
  cand_betas = np.linspace(args.beta_from, args.beta_to, args.num_betas)
  params_grid = [(alpha, beta) for alpha in cand_alphas
  for beta in cand_betas]
 
+ # external scorer
  ext_scorer = LmScorer(args.alpha_from, args.beta_from,
  args.language_model_path)
- ## tune parameters in loop
- for alpha, beta in params_grid:
- wer_sum, wer_counter = 0, 0
- # reset scorer
- ext_scorer.reset_params(alpha, beta)
- # beam search using multiple processes
- beam_search_results = ctc_beam_search_decoder_batch(
- probs_split=probs_split,
- vocabulary=data_generator.vocab_list,
- beam_size=args.beam_size,
- cutoff_prob=args.cutoff_prob,
- blank_id=len(data_generator.vocab_list),
- num_processes=args.num_processes_beam_search,
- ext_scoring_func=ext_scorer, )
- for i, beam_search_result in enumerate(beam_search_results):
- target_transcription = ''.join([
- data_generator.vocab_list[index] for index in infer_data[i][1]
- ])
- wer_sum += wer(target_transcription, beam_search_result[0][1])
- wer_counter += 1
 
- print("alpha = %f\tbeta = %f\tWER = %f" %
- (alpha, beta, wer_sum / wer_counter))
+ wer_sum = [0.0 for i in xrange(len(params_grid))]
+ wer_counter = [0 for i in xrange(len(params_grid))]
+ ave_wer = [0.0 for i in xrange(len(params_grid))]
+ num_batches = 0
+
+ ## incremental tuning batch by batch
+ for infer_data in batch_reader():
+ # run inference
+ infer_results = inferer.infer(input=infer_data)
+ num_steps = len(infer_results) // len(infer_data)
+ probs_split = [
+ infer_results[i * num_steps:(i + 1) * num_steps]
+ for i in xrange(0, len(infer_data))
+ ]
+ # target transcription
+ target_transcription = [
+ ''.join([
+ data_generator.vocab_list[index] for index in infer_data[i][1]
+ ]) for i, probs in enumerate(probs_split)
+ ]
+
+ # grid search on current batch
+ for index, (alpha, beta) in enumerate(params_grid):
+ # reset scorer
+ ext_scorer.reset_params(alpha, beta)
+ beam_search_results = ctc_beam_search_decoder_batch(
+ probs_split=probs_split,
+ vocabulary=data_generator.vocab_list,
+ beam_size=args.beam_size,
+ blank_id=len(data_generator.vocab_list),
+ num_processes=args.num_processes_beam_search,
+ ext_scoring_func=ext_scorer,
+ cutoff_prob=args.cutoff_prob, )
+ for i, beam_search_result in enumerate(beam_search_results):
+ wer_sum[index] += wer(target_transcription[i],
+ beam_search_result[0][1])
+ wer_counter[index] += 1
+ ave_wer[index] = wer_sum[index] / wer_counter[index]
+ print("alpha = %f, beta = %f, WER = %f" %
+ (alpha, beta, ave_wer[index]))
+
+ # output tuning result til current batch
+ ave_wer_min = min(ave_wer)
+ min_index = ave_wer.index(ave_wer_min)
+ print("Finish batch %d, alpha_opt = %f, beta_opt = %f, WER_opt = %f\n" %
+ (num_batches, params_grid[min_index][0],
+ params_grid[min_index][1], ave_wer_min))
+ num_batches += 1
 
 
 def main():
+ utils.print_arguments(args)
  paddle.init(use_gpu=args.use_gpu, trainer_count=1)
  tune()