Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix CI test for gigaspeech #1787

Merged
merged 3 commits into from
Oct 29, 2024
Merged

Fix CI test for gigaspeech #1787

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
be7539e
Fix CI for gigaspeech
csukuangfj Oct 29, 2024
865ecc6
Disable modified_beam_search test.
csukuangfj Oct 29, 2024
cf75bb4
fix CI
csukuangfj Oct 29, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 11 additions & 3 deletions .github/scripts/run-gigaspeech-pruned-transducer-stateless2-2022-05-12.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,7 @@ repo=$(basename $repo_url)

echo "GITHUB_EVENT_NAME: ${GITHUB_EVENT_NAME}"
echo "GITHUB_EVENT_LABEL_NAME: ${GITHUB_EVENT_LABEL_NAME}"
if [[ x"${GITHUB_EVENT_NAME}" == x"schedule" || x"${GITHUB_EVENT_LABEL_NAME}" == x"run-decode" ]]; then
if [[ x"${GITHUB_EVENT_NAME}" == x"schedule" || x"${GITHUB_EVENT_NAME}" == x"workflow_dispatch" || x"${GITHUB_EVENT_LABEL_NAME}" == x"run-decode" ]]; then
mkdir -p pruned_transducer_stateless2/exp
ln -s $PWD/$repo/exp/pretrained-iter-3488000-avg-20.pt pruned_transducer_stateless2/exp/epoch-999.pt
ln -s $PWD/$repo/data/lang_bpe_500 data/
Expand All @@ -29,8 +29,16 @@ if [[ x"${GITHUB_EVENT_NAME}" == x"schedule" || x"${GITHUB_EVENT_LABEL_NAME}" ==
ls -lh data/fbank
ls -lh pruned_transducer_stateless2/exp

ln -sf data/fbank/cuts_DEV.jsonl.gz data/fbank/gigaspeech_cuts_DEV.jsonl.gz
ln -sf data/fbank/cuts_TEST.jsonl.gz data/fbank/gigaspeech_cuts_TEST.jsonl.gz
pushd data/fbank
curl -SL -O https://huggingface.co/csukuangfj/giga-dev-dataset-fbank/resolve/main/data/fbank/cuts_DEV.jsonl.gz
curl -SL -O https://huggingface.co/csukuangfj/giga-dev-dataset-fbank/resolve/main/data/fbank/cuts_TEST.jsonl.gz
curl -SL -O https://huggingface.co/csukuangfj/giga-dev-dataset-fbank/resolve/main/data/fbank/feats_DEV.lca
curl -SL -O https://huggingface.co/csukuangfj/giga-dev-dataset-fbank/resolve/main/data/fbank/feats_TEST.lca

ln -sf cuts_DEV.jsonl.gz gigaspeech_cuts_DEV.jsonl.gz
ln -sf cuts_TEST.jsonl.gz gigaspeech_cuts_TEST.jsonl.gz
popd


log "Decoding dev and test"

Expand Down
4 changes: 2 additions & 2 deletions .github/scripts/run-librispeech-lstm-transducer-stateless2-2022-09-03.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -162,7 +162,7 @@ if [[ x"${GITHUB_EVENT_LABEL_NAME}" == x"LODR" ]]; then
--ngram-lm-scale -0.16
fi

if [[ x"${GITHUB_EVENT_NAME}" == x"schedule" ]]; then
if [[ x"${GITHUB_EVENT_NAME}" == x"schedule" || x"${GITHUB_EVENT_NAME}" == x"workflow_dispatch" ]]; then
mkdir -p lstm_transducer_stateless2/exp
ln -s $PWD/$repo/exp/pretrained.pt lstm_transducer_stateless2/exp/epoch-999.pt
ln -s $PWD/$repo/data/lang_bpe_500 data/
Expand All @@ -175,7 +175,7 @@ if [[ x"${GITHUB_EVENT_NAME}" == x"schedule" ]]; then
# use a small value for decoding with CPU
max_duration=100

for method in greedy_search fast_beam_search modified_beam_search; do
for method in greedy_search fast_beam_search; do
log "Decoding with $method"

./lstm_transducer_stateless2/decode.py \
Expand Down
6 changes: 3 additions & 3 deletions .github/workflows/run-gigaspeech-2022-05-13.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,7 +41,7 @@ concurrency:

jobs:
run_gigaspeech_2022_05_13:
if: github.event.label.name == 'ready' || github.event.label.name == 'run-decode' || github.event_name == 'push' || github.event_name == 'schedule'
if: github.event_name == 'workflow_dispatch' || github.event.label.name == 'ready' || github.event.label.name == 'run-decode' || github.event_name == 'push' || github.event_name == 'schedule'
runs-on: ${{ matrix.os }}
strategy:
matrix:
Expand Down Expand Up @@ -106,7 +106,7 @@ jobs:
.github/scripts/run-gigaspeech-pruned-transducer-stateless2-2022-05-12.sh

- name: Display decoding results for gigaspeech pruned_transducer_stateless2
if: github.event_name == 'schedule' || github.event.label.name == 'run-decode'
if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch' || github.event.label.name == 'run-decode'
shell: bash
run: |
cd egs/gigaspeech/ASR/
Expand All @@ -122,7 +122,7 @@ jobs:

- name: Upload decoding results for gigaspeech pruned_transducer_stateless2
uses: actions/upload-artifact@v4
if: github.event_name == 'schedule' || github.event.label.name == 'run-decode'
if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch' || github.event.label.name == 'run-decode'
with:
name: torch-${{ matrix.torch }}-python-${{ matrix.python-version }}-ubuntu-latest-cpu-gigaspeech-pruned_transducer_stateless2-2022-05-12
path: egs/gigaspeech/ASR/pruned_transducer_stateless2/exp/
12 changes: 6 additions & 6 deletions .github/workflows/run-librispeech-lstm-transducer-stateless2-2022-09-03.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@ concurrency:

jobs:
run_librispeech_lstm_transducer_stateless2_2022_09_03:
if: github.event.label.name == 'ready' || github.event.label.name == 'LODR' || github.event.label.name == 'shallow-fusion' || github.event_name == 'push' || github.event_name == 'schedule'
if: github.event.label.name == 'ready' || github.event.label.name == 'LODR' || github.event.label.name == 'shallow-fusion' || github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
runs-on: ${{ matrix.os }}
strategy:
matrix:
Expand Down Expand Up @@ -116,7 +116,7 @@ jobs:
.github/scripts/run-librispeech-lstm-transducer-stateless2-2022-09-03.sh

- name: Display decoding results for lstm_transducer_stateless2
if: github.event_name == 'schedule'
if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
shell: bash
run: |
cd egs/librispeech/ASR
Expand All @@ -130,9 +130,9 @@ jobs:
find fast_beam_search -name "log-*" -exec grep -n --color "best for test-clean" {} + | sort -n -k2
find fast_beam_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2

echo "===modified beam search==="
find modified_beam_search -name "log-*" -exec grep -n --color "best for test-clean" {} + | sort -n -k2
find modified_beam_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2
# echo "===modified beam search==="
# find modified_beam_search -name "log-*" -exec grep -n --color "best for test-clean" {} + | sort -n -k2
# find modified_beam_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2

- name: Display decoding results for lstm_transducer_stateless2
if: github.event.label.name == 'shallow-fusion'
Expand All @@ -159,7 +159,7 @@ jobs:

- name: Upload decoding results for lstm_transducer_stateless2
uses: actions/upload-artifact@v4
if: github.event_name == 'schedule' || github.event.label.name == 'shallow-fusion' || github.event.label.name == 'LODR'
if: github.event_name == 'schedule' || github.event.label.name == 'shallow-fusion' || github.event.label.name == 'LODR' || github.event_name == 'workflow_dispatch'
with:
name: torch-${{ matrix.torch }}-python-${{ matrix.python-version }}-ubuntu-latest-cpu-lstm_transducer_stateless2-2022-09-03
path: egs/librispeech/ASR/lstm_transducer_stateless2/exp/
2 changes: 1 addition & 1 deletion egs/gigaspeech/ASR/pruned_transducer_stateless2/decode.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -260,7 +260,7 @@ def decode_one_batch(
Return the decoding result. See above description for the format of
the returned dict.
"""
device = model.device
device = next(model.parameters()).device
feature = batch["inputs"]
assert feature.ndim == 3

Expand Down
Loading