Skip to content
GitLab
Explore
Sign in
Primary navigation
Search or go to…
Project
A
asr-benchmarks
Manage
Activity
Members
Labels
Plan
Issues
8
Issue boards
Milestones
Wiki
Redmine
Code
Merge requests
0
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Snippets
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Package Registry
Container Registry
Operate
Environments
Terraform modules
Monitor
Incidents
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
Marcin Wątroba
asr-benchmarks
Commits
9b9238ce
Commit
9b9238ce
authored
2 years ago
by
Marcin Wątroba
Browse files
Options
Downloads
Patches
Plain Diff
Update docs
parent
aa8fb055
Branches
Branches containing commit
No related merge requests found
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
docs/run_experiment.ipynb
+5
-5
5 additions, 5 deletions
docs/run_experiment.ipynb
with
5 additions
and
5 deletions
docs/run_experiment.ipynb
+
5
−
5
View file @
9b9238ce
...
...
@@ -25,13 +25,13 @@
"outputs": [],
"source": [
"# imports\n",
"from
experiment.sentence_wer_processor
.flair_upos_multi_transformers_wer_processor_base import \\\n",
"from
sziszapangma.integration.task
.flair_upos_multi_transformers_wer_processor_base import \\\n",
" FlairUposMultiTransformersWerProcessorBase\n",
"from
experiment.sentence_wer_processor
.wikineural_multilingual_ner_transformers_wer_processor_base import \\\n",
"from
sziszapangma.integration.task
.wikineural_multilingual_ner_transformers_wer_processor_base import \\\n",
" WikineuralMultilingualNerTransformersWerProcessorBase\n",
"from
experiment.sentence_wer_processor
.spacy_pos_sentence_dep_tag_processor import SpacyDepTagSentenceWerProcessor\n",
"from
experiment.sentence_wer_processor
.spacy_ner_sentence_wer_processor import SpacyNerSentenceWerProcessor\n",
"from
experiment.sentence_wer_processor
.spacy_pos_sentence_wer_processor import SpacyPosSentenceWerProcessor\n",
"from
sziszapangma.integration.task
.spacy_pos_sentence_dep_tag_processor import SpacyDepTagSentenceWerProcessor\n",
"from
sziszapangma.integration.task
.spacy_ner_sentence_wer_processor import SpacyNerSentenceWerProcessor\n",
"from
sziszapangma.integration.task
.spacy_pos_sentence_wer_processor import SpacyPosSentenceWerProcessor\n",
"from sziszapangma.core.transformer.fasttext_embedding_transformer import FasttextEmbeddingTransformer\n",
"from sziszapangma.integration.task.embedding_wer_metrics_task import EmbeddingWerMetricsTask\n",
"from sziszapangma.integration.task.classic_wer_metric_task import ClassicWerMetricTask\n",
...
...
%% Cell type:markdown id: tags:
# Run experiment
%% Cell type:markdown id: tags:
To run any experiment steps in
`import_any_dataset.ipynb`
and
`use_any_asr.ipynb`
must be done.
Experiment is managed by
`ExperimentRepository`
. Below is example of simple experiment.
%% Cell type:code id: tags:
```
python
# imports
from
experiment.sentence_wer_processor
.flair_upos_multi_transformers_wer_processor_base
import
\
from
sziszapangma.integration.task
.flair_upos_multi_transformers_wer_processor_base
import
\
FlairUposMultiTransformersWerProcessorBase
from
experiment.sentence_wer_processor
.wikineural_multilingual_ner_transformers_wer_processor_base
import
\
from
sziszapangma.integration.task
.wikineural_multilingual_ner_transformers_wer_processor_base
import
\
WikineuralMultilingualNerTransformersWerProcessorBase
from
experiment.sentence_wer_processor
.spacy_pos_sentence_dep_tag_processor
import
SpacyDepTagSentenceWerProcessor
from
experiment.sentence_wer_processor
.spacy_ner_sentence_wer_processor
import
SpacyNerSentenceWerProcessor
from
experiment.sentence_wer_processor
.spacy_pos_sentence_wer_processor
import
SpacyPosSentenceWerProcessor
from
sziszapangma.integration.task
.spacy_pos_sentence_dep_tag_processor
import
SpacyDepTagSentenceWerProcessor
from
sziszapangma.integration.task
.spacy_ner_sentence_wer_processor
import
SpacyNerSentenceWerProcessor
from
sziszapangma.integration.task
.spacy_pos_sentence_wer_processor
import
SpacyPosSentenceWerProcessor
from
sziszapangma.core.transformer.fasttext_embedding_transformer
import
FasttextEmbeddingTransformer
from
sziszapangma.integration.task.embedding_wer_metrics_task
import
EmbeddingWerMetricsTask
from
sziszapangma.integration.task.classic_wer_metric_task
import
ClassicWerMetricTask
from
experiment.hf_asr.wav2vec2_hf
import
Wav2Vec2AsrProcessor
from
experiment.utils.property_helper
import
PropertyHelper
from
sziszapangma.integration.task.asr_task
import
AsrTask
from
sziszapangma.integration.experiment_manager
import
ExperimentManager
from
sziszapangma.integration.audio_repository.local_audio_record_repository
import
LocalAudioRecordRepository
from
experiment.utils.loaded_dataset_helper
import
LoadedDatasetHelper
from
pathlib
import
Path
from
sziszapangma.integration.repository.multi_files_experiment_repository
import
MultiFilesExperimentRepository
```
%% Cell type:code id: tags:
```
python
# globals
DATASET_NAME
=
'
dataset_name
'
ASR_NAME
=
'
asr_name
'
REPOSITORY_ROOT_PATH
=
Path
.
home
()
/
'
asr-benchmarks-repository
'
AUDIO_ROOT_PATH
=
Path
.
home
()
/
'
.cache/asr-benchmarks
'
FASTTEXT_LANGUAGE_CODE
=
'
pl
'
WIKINEURAL
=
"
wikineural
"
SPACY_MODEL_NAME
=
'
pl_core_news_lg
'
FLAIR_UPOS_MULTI
=
'
flair_upos_multi
'
```
%% Cell type:code id: tags:
```
python
# data providers
repository
=
MultiFilesExperimentRepository
(
REPOSITORY_ROOT_PATH
,
DATASET_NAME
)
record_provider
=
LoadedDatasetHelper
(
repository
,
LocalAudioRecordRepository
(
AUDIO_ROOT_PATH
,
DATASET_NAME
),
DATASET_NAME
)
```
%% Cell type:code id: tags:
```
python
# experiment definition
experiment_processor
=
ExperimentManager
(
record_id_iterator
=
record_provider
,
processing_tasks
=
[
AsrTask
(
asr_property_name
=
PropertyHelper
.
asr_result
(
ASR_NAME
),
task_name
=
f
"
AsrTask___
{
DATASET_NAME
}
___
{
ASR_NAME
}
"
,
require_update
=
False
,
asr_processor
=
Wav2Vec2AsrProcessor
(
"
facebook/wav2vec2-large-xlsr-53-polish
"
),
record_path_provider
=
record_provider
,
),
ClassicWerMetricTask
(
task_name
=
f
"
ClassicWerMetricTask___
{
DATASET_NAME
}
___
{
ASR_NAME
}
"
,
asr_property_name
=
PropertyHelper
.
asr_result
(
ASR_NAME
),
gold_transcript_property_name
=
PropertyHelper
.
get_gold_transcript_words
(),
metrics_property_name
=
PropertyHelper
.
word_wer_classic_metrics
(
ASR_NAME
),
require_update
=
False
,
alignment_property_name
=
PropertyHelper
.
word_wer_classic_alignment
(
ASR_NAME
),
),
EmbeddingWerMetricsTask
(
task_name
=
f
"
EmbeddingWerMetricsTask___
{
DATASET_NAME
}
___
{
ASR_NAME
}
"
,
asr_property_name
=
PropertyHelper
.
asr_result
(
ASR_NAME
),
gold_transcript_property_name
=
PropertyHelper
.
get_gold_transcript_words
(),
require_update
=
False
,
embedding_transformer
=
FasttextEmbeddingTransformer
(
FASTTEXT_LANGUAGE_CODE
),
embeddings_alignment_property_name
=
PropertyHelper
.
word_wer_embeddings_alignment
(
ASR_NAME
),
embeddings_metrics_property_name
=
PropertyHelper
.
word_wer_embeddings_metrics
(
ASR_NAME
),
soft_alignment_property_name
=
PropertyHelper
.
word_wer_soft_alignment
(
ASR_NAME
),
soft_metrics_property_name
=
PropertyHelper
.
word_wer_soft_metrics
(
ASR_NAME
),
),
SpacyPosSentenceWerProcessor
(
model_name
=
SPACY_MODEL_NAME
,
gold_transcript_property_name
=
PropertyHelper
.
get_gold_transcript_raw
(),
asr_property_name
=
PropertyHelper
.
asr_result
(
ASR_NAME
),
alignment_property_name
=
PropertyHelper
.
pos_alignment
(
ASR_NAME
,
SPACY_MODEL_NAME
),
wer_property_name
=
PropertyHelper
.
pos_metrics
(
ASR_NAME
,
SPACY_MODEL_NAME
),
task_name
=
f
"
SpacyPosSentenceWerProcessor___
{
DATASET_NAME
}
___
{
ASR_NAME
}
"
,
require_update
=
False
,
),
SpacyNerSentenceWerProcessor
(
model_name
=
SPACY_MODEL_NAME
,
gold_transcript_property_name
=
PropertyHelper
.
get_gold_transcript_raw
(),
asr_property_name
=
PropertyHelper
.
asr_result
(
ASR_NAME
),
alignment_property_name
=
PropertyHelper
.
ner_alignment
(
ASR_NAME
,
SPACY_MODEL_NAME
),
wer_property_name
=
PropertyHelper
.
ner_metrics
(
ASR_NAME
,
SPACY_MODEL_NAME
),
task_name
=
f
"
SpacyNerSentenceWerProcessor___
{
DATASET_NAME
}
___
{
ASR_NAME
}
"
,
require_update
=
False
,
),
SpacyDepTagSentenceWerProcessor
(
model_name
=
SPACY_MODEL_NAME
,
gold_transcript_property_name
=
PropertyHelper
.
get_gold_transcript_raw
(),
asr_property_name
=
PropertyHelper
.
asr_result
(
ASR_NAME
),
alignment_property_name
=
PropertyHelper
.
dep_tag_alignment
(
ASR_NAME
,
SPACY_MODEL_NAME
),
wer_property_name
=
PropertyHelper
.
dep_tag_metrics
(
ASR_NAME
,
SPACY_MODEL_NAME
),
task_name
=
f
"
SpacyDepTagSentenceWerProcessor___
{
DATASET_NAME
}
___
{
ASR_NAME
}
"
,
require_update
=
False
,
),
WikineuralMultilingualNerTransformersWerProcessorBase
(
gold_transcript_property_name
=
PropertyHelper
.
get_gold_transcript_raw
(),
asr_property_name
=
PropertyHelper
.
asr_result
(
ASR_NAME
),
alignment_property_name
=
PropertyHelper
.
ner_alignment
(
ASR_NAME
,
WIKINEURAL
),
wer_property_name
=
PropertyHelper
.
ner_metrics
(
ASR_NAME
,
WIKINEURAL
),
task_name
=
f
"
WikineuralMultilingualNerTransformersWerProcessorBase___
{
DATASET_NAME
}
___
{
ASR_NAME
}
"
,
require_update
=
False
,
),
FlairUposMultiTransformersWerProcessorBase
(
gold_transcript_property_name
=
PropertyHelper
.
get_gold_transcript_raw
(),
asr_property_name
=
PropertyHelper
.
asr_result
(
ASR_NAME
),
alignment_property_name
=
PropertyHelper
.
pos_alignment
(
ASR_NAME
,
FLAIR_UPOS_MULTI
),
wer_property_name
=
PropertyHelper
.
pos_metrics
(
ASR_NAME
,
FLAIR_UPOS_MULTI
),
task_name
=
f
"
FlairUposMultiTransformersWerProcessorBase___
{
DATASET_NAME
}
___
{
ASR_NAME
}
"
,
require_update
=
False
,
)
],
experiment_repository
=
repository
,
)
```
%% Cell type:code id: tags:
```
python
# run experiment
experiment_processor
.
process
()
```
...
...
This diff is collapsed.
Click to expand it.
Preview
0%
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment