mirrors/icefall
1
0
Fork 0
mirror of https://github.com/k2-fsa/icefall.git synced 2025-09-05 07:04:18 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add results for standard beam search.

Browse Source
This commit is contained in:
Fangjun Kuang 2022-03-27 23:46:50 +08:00
parent cc6c608fb6
commit ce079e1417
1 changed files with 22 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
egs/librispeech/ASR/RESULTS.md
View File

@ -168,6 +168,7 @@ Using commit `395a3f952be1449cd7c92b896f4eb9a1c899e2c7`.
| greedy search (max sym per frame 2) | 2.60 | 6.32 | --epoch 59, --avg 19, --max-duration 1000| | greedy search (max sym per frame 2) | 2.60 | 6.32 | --epoch 59, --avg 19, --max-duration 1000|
| greedy search (max sym per frame 3) | 2.60 | 6.32 | --epoch 59, --avg 19, --max-duration 1000| | greedy search (max sym per frame 3) | 2.60 | 6.32 | --epoch 59, --avg 19, --max-duration 1000|
| modified beam search (beam size 4) | 2.56 | 6.20 | --epoch 59, --avg 19, --max-duration 1000| | modified beam search (beam size 4) | 2.56 | 6.20 | --epoch 59, --avg 19, --max-duration 1000|
| beam search (beam size 4) | 2.57 | 6.21 | --epoch 59, --avg 19, --max-duration 1000|
The training command for reproducing is given below: The training command for reproducing is given below:
```bash ```bash
@ -203,7 +204,7 @@ for sym in 1 2 3; do
--max-sym-per-frame $sym --max-sym-per-frame $sym
done done
## modified_beam_search ## modified beam search
./transducer_stateless/decode.py \ ./transducer_stateless/decode.py \
--epoch $epoch \ --epoch $epoch \
--avg $avg \ --avg $avg \
@ -211,6 +212,15 @@ done
--max-duration 1000 \ --max-duration 1000 \
--decoding-method modified_beam_search \ --decoding-method modified_beam_search \
--beam-size 4 --beam-size 4
## beam search
./transducer_stateless/decode.py \
--epoch $epoch \
--avg $avg \
--exp-dir ./transducer_stateless/exp-2 \
--max-duration 1000 \
--decoding-method beam_search \
--beam-size 4
``` ```
You can find a pretrained model by visiting You can find a pretrained model by visiting
@ -228,6 +238,7 @@ Using commit `395a3f952be1449cd7c92b896f4eb9a1c899e2c7`.
| greedy search (max sym per frame 2) | 2.63 | 6.37 | --epoch 49, --avg 11, --max-duration 1000| | greedy search (max sym per frame 2) | 2.63 | 6.37 | --epoch 49, --avg 11, --max-duration 1000|
| greedy search (max sym per frame 3) | 2.63 | 6.37 | --epoch 49, --avg 11, --max-duration 1000| | greedy search (max sym per frame 3) | 2.63 | 6.37 | --epoch 49, --avg 11, --max-duration 1000|
| modified beam search (beam size 4) | 2.62 | 6.32 | --epoch 49, --avg 11, --max-duration 1000| | modified beam search (beam size 4) | 2.62 | 6.32 | --epoch 49, --avg 11, --max-duration 1000|
| beam search (beam size 4) | 2.58 | 6.24 | --epoch 49, --avg 11, --max-duration 1000|
The training command for reproducing is given below: The training command for reproducing is given below:
```bash ```bash
@ -263,7 +274,7 @@ for sym in 1 2 3; do
--max-sym-per-frame $sym --max-sym-per-frame $sym
done done
## modified_beam_search ## modified beam search
./transducer_stateless/decode.py \ ./transducer_stateless/decode.py \
--epoch $epoch \ --epoch $epoch \
--avg $avg \ --avg $avg \
@ -271,6 +282,15 @@ done
--max-duration 1000 \ --max-duration 1000 \
--decoding-method modified_beam_search \ --decoding-method modified_beam_search \
--beam-size 4 --beam-size 4
## beam search
./transducer_stateless/decode.py \
--epoch $epoch \
--avg $avg \
--exp-dir ./transducer_stateless/exp \
--max-duration 1000 \
--decoding-method beam_search \
--beam-size 4
``` ```
You can find a pretrained model by visiting You can find a pretrained model by visiting