mirrors/icefall
1
0
Fork 0
mirror of https://github.com/k2-fsa/icefall.git synced 2025-08-09 10:02:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add results for mixed precision with max-duration 300

Browse Source
This commit is contained in:
Daniel Povey 2022-04-13 20:22:11 +08:00
parent c0003483d3
commit af6ae840ee
1 changed files with 20 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

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

@ -2,9 +2,10 @@
### LibriSpeech BPE training results (Pruned Transducer 2) ### LibriSpeech BPE training results (Pruned Transducer 2)
[pruned_transducer_stateless2](./pruned_transducer_stateless2)
This is with a reworked version of the conformer encoder, with many changes. This is with a reworked version of the conformer encoder, with many changes.
[pruned_transducer_stateless2](./pruned_transducer_stateless2) #### Training on fulll librispeech
using commit `34aad74a2c849542dd5f6359c9e6b527e8782fd6`. using commit `34aad74a2c849542dd5f6359c9e6b527e8782fd6`.
See <https://github.com/k2-fsa/icefall/pull/288> See <https://github.com/k2-fsa/icefall/pull/288>
@ -33,9 +34,9 @@ The Tensorboard log is at <https://tensorboard.dev/experiment/Xoz0oABMTWewo1slNF
only from epoch 3). only from epoch 3).
The WERs for librispeech 100 hours are: #### Training on train-clean-100:
Trained with one job: Trained with 1 job:
`python3 ./pruned_transducer_stateless2/train.py --exp-dir=pruned_transducer_stateless2/exp_100h_ws1 --world-size 1 --num-epochs 40 --full-libri 0 --max-duration 300` `python3 ./pruned_transducer_stateless2/train.py --exp-dir=pruned_transducer_stateless2/exp_100h_ws1 --world-size 1 --num-epochs 40 --full-libri 0 --max-duration 300`
and decoded with: and decoded with:
`python3 ./pruned_transducer_stateless2/decode.py --exp-dir pruned_transducer_stateless2/exp_100h_ws1 --epoch 19 --avg 8 --bpe-model ./data/lang_bpe_500/bpe.model --max-duration 600`. `python3 ./pruned_transducer_stateless2/decode.py --exp-dir pruned_transducer_stateless2/exp_100h_ws1 --epoch 19 --avg 8 --bpe-model ./data/lang_bpe_500/bpe.model --max-duration 600`.
@ -51,7 +52,7 @@ schedule is not visible due to a since-fixed bug).
| fast beam search | 6.58 | 17.27 | --epoch 29 --avg 8 --decoding-method fast_beam_search | | fast beam search | 6.58 | 17.27 | --epoch 29 --avg 8 --decoding-method fast_beam_search |
| fast beam search | 6.53 | 16.82 | --epoch 39 --avg 10 --decoding-method fast_beam_search | | fast beam search | 6.53 | 16.82 | --epoch 39 --avg 10 --decoding-method fast_beam_search |
Trained with two jobs: Trained with 2 jobs:
`python3 ./pruned_transducer_stateless2/train.py --exp-dir=pruned_transducer_stateless2/exp_100h_ws2 --world-size 2 --num-epochs 40 --full-libri 0 --max-duration 300` `python3 ./pruned_transducer_stateless2/train.py --exp-dir=pruned_transducer_stateless2/exp_100h_ws2 --world-size 2 --num-epochs 40 --full-libri 0 --max-duration 300`
and decoded with: and decoded with:
`python3 ./pruned_transducer_stateless2/decode.py --exp-dir pruned_transducer_stateless2/exp_100h_ws2 --epoch 19 --avg 8 --bpe-model ./data/lang_bpe_500/bpe.model --max-duration 600`. `python3 ./pruned_transducer_stateless2/decode.py --exp-dir pruned_transducer_stateless2/exp_100h_ws2 --epoch 19 --avg 8 --bpe-model ./data/lang_bpe_500/bpe.model --max-duration 600`.
@ -82,6 +83,21 @@ The Tensorboard log is at <https://tensorboard.dev/experiment/a3T0TyC0R5aLj5bmFb
| greedy search (max sym per frame 1) | 6.86 | 18.29 | --epoch 30 --avg 10 | | greedy search (max sym per frame 1) | 6.86 | 18.29 | --epoch 30 --avg 10 |
Trained with 1 job, with --use-fp16=True --max-duration=300 i.e. with half-precision
floats (but without increasing max-duration), after merging <https://github.com/k2-fsa/icefall/pull/305>.
Train command was
`python3 ./pruned_transducer_stateless2/train.py --exp-dir=pruned_transducer_stateless2/exp_100h_fp16 --world-size 1 --num-epochs 40 --full-libri 0 --max-duration 300 --use-fp16 True`
The Tensorboard log is at <https://tensorboard.dev/experiment/DAtGG9lpQJCROUDwPNxwpA>
| | test-clean | test-other | comment |
|-------------------------------------|------------|------------|-----------------------|
| greedy search (max sym per frame 1) | 7.10 | 18.57 | --epoch 19 --avg 8 |
| greedy search (max sym per frame 1) | 6.81 | 17.84 | --epoch 29 --avg 8 |
| greedy search (max sym per frame 1) | 6.63 | 17.39 | --epoch 30 --avg 10 |
Trained with 1 job, with --use-fp16=True --max-duration=500, i.e. with half-precision Trained with 1 job, with --use-fp16=True --max-duration=500, i.e. with half-precision
floats and max-duration increased from 300 to 500, after merging <https://github.com/k2-fsa/icefall/pull/305>. floats and max-duration increased from 300 to 500, after merging <https://github.com/k2-fsa/icefall/pull/305>.
Train command was Train command was