Update libri_trans asr recipe

[hirofumi0810]

• Change RNN to Transformer
• Change character unit to BPE1k unit

[codecov]

Codecov Report

Merging #1344 into v.0.6.0 will increase coverage by 0.23%.
The diff coverage is n/a.

@@             Coverage Diff             @@
##           v.0.6.0    #1344      +/-   ##
===========================================
+ Coverage    77.45%   77.69%   +0.23%     
===========================================
  Files          110      110              
  Lines        10541    10776     +235     
===========================================
+ Hits          8165     8372     +207     
- Misses        2376     2404      +28

Impacted Files	Coverage Δ
espnet/nets/pytorch_backend/streaming/segment.py	60.6% <0%> (-2.3%)	⬇️
espnet/nets/pytorch_backend/e2e_tts_fastspeech.py	95.03% <0%> (-0.21%)	⬇️
espnet/nets/pytorch_backend/e2e_tts_transformer.py	89.58% <0%> (+0.18%)	⬆️
espnet/nets/pytorch_backend/rnn/encoders.py	98.34% <0%> (+0.53%)	⬆️
espnet/nets/pytorch_backend/e2e_asr_transducer.py	64.64% <0%> (+3.87%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b38b4b0...d9900d2. Read the comment docs.

[sw005320]

LGTM, thanks!

• please consider uploading the model.
• what is the major difference between bpe and char transformer yaml files? Batchsize (due to the shorter output length with bpe?)? Can you add the comment to the yaml file?

[hirofumi0810]

@sw005320 I will update results with SpecAugment tomorrow.

[hirofumi0810]

@sw005320 Ready to be merged.

[sw005320]

Very cool!
The result is impressive.

[sw005320]

LGTM