url
stringlengths 52
53
| repository_url
stringclasses 1
value | labels_url
stringlengths 66
67
| comments_url
stringlengths 61
62
| events_url
stringlengths 59
60
| html_url
stringlengths 40
43
| id
int64 719M
2.8B
| node_id
stringlengths 18
32
| number
int64 1.28k
11.9k
| title
stringlengths 1
350
| user
dict | labels
listlengths 0
10
| state
stringclasses 2
values | locked
bool 2
classes | assignee
dict | assignees
listlengths 0
3
| milestone
null | comments
sequencelengths 0
30
⌀ | created_at
timestamp[s] | updated_at
timestamp[s] | closed_at
timestamp[s] | author_association
stringclasses 4
values | sub_issues_summary
dict | active_lock_reason
stringclasses 1
value | draft
bool 2
classes | pull_request
dict | body
stringlengths 0
73.4k
⌀ | closed_by
dict | reactions
dict | timeline_url
stringlengths 61
62
| performed_via_github_app
null | state_reason
stringclasses 3
values | is_pull_request
bool 2
classes |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
https://api.github.com/repos/NVIDIA/NeMo/issues/5580 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5580/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5580/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5580/events | https://github.com/NVIDIA/NeMo/pull/5580 | 1,485,697,101 | PR_kwDOC_bI7s5E2Qbu | 5,580 | remove useless files. | {
"login": "XuesongYang",
"id": 1646669,
"node_id": "MDQ6VXNlcjE2NDY2Njk=",
"avatar_url": "https://avatars.githubusercontent.com/u/1646669?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/XuesongYang",
"html_url": "https://github.com/XuesongYang",
"followers_url": "https://api.github.com/users/XuesongYang/followers",
"following_url": "https://api.github.com/users/XuesongYang/following{/other_user}",
"gists_url": "https://api.github.com/users/XuesongYang/gists{/gist_id}",
"starred_url": "https://api.github.com/users/XuesongYang/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/XuesongYang/subscriptions",
"organizations_url": "https://api.github.com/users/XuesongYang/orgs",
"repos_url": "https://api.github.com/users/XuesongYang/repos",
"events_url": "https://api.github.com/users/XuesongYang/events{/privacy}",
"received_events_url": "https://api.github.com/users/XuesongYang/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4847373924,
"node_id": "LA_kwDOC_bI7s8AAAABIO0OZA",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/common",
"name": "common",
"color": "c5def5",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-09T00:52:51 | 2022-12-09T18:04:22 | 2022-12-09T18:04:18 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5580",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5580",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5580.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5580.patch",
"merged_at": "2022-12-09T18:04:18"
} | Signed-off-by: Xuesong Yang <[email protected]>
# What does this PR do ?
remove `vocab.py` because nowhere else uses it.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "XuesongYang",
"id": 1646669,
"node_id": "MDQ6VXNlcjE2NDY2Njk=",
"avatar_url": "https://avatars.githubusercontent.com/u/1646669?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/XuesongYang",
"html_url": "https://github.com/XuesongYang",
"followers_url": "https://api.github.com/users/XuesongYang/followers",
"following_url": "https://api.github.com/users/XuesongYang/following{/other_user}",
"gists_url": "https://api.github.com/users/XuesongYang/gists{/gist_id}",
"starred_url": "https://api.github.com/users/XuesongYang/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/XuesongYang/subscriptions",
"organizations_url": "https://api.github.com/users/XuesongYang/orgs",
"repos_url": "https://api.github.com/users/XuesongYang/repos",
"events_url": "https://api.github.com/users/XuesongYang/events{/privacy}",
"received_events_url": "https://api.github.com/users/XuesongYang/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5580/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5580/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5579 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5579/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5579/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5579/events | https://github.com/NVIDIA/NeMo/issues/5579 | 1,485,629,853 | I_kwDOC_bI7s5YjOmd | 5,579 | omegaconf.errors.ConfigKeyError: Key 'use_tarred_dataset' not in 'MTEncDecConfig' | {
"login": "felicitywang1",
"id": 96318866,
"node_id": "U_kgDOBb21kg",
"avatar_url": "https://avatars.githubusercontent.com/u/96318866?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/felicitywang1",
"html_url": "https://github.com/felicitywang1",
"followers_url": "https://api.github.com/users/felicitywang1/followers",
"following_url": "https://api.github.com/users/felicitywang1/following{/other_user}",
"gists_url": "https://api.github.com/users/felicitywang1/gists{/gist_id}",
"starred_url": "https://api.github.com/users/felicitywang1/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/felicitywang1/subscriptions",
"organizations_url": "https://api.github.com/users/felicitywang1/orgs",
"repos_url": "https://api.github.com/users/felicitywang1/repos",
"events_url": "https://api.github.com/users/felicitywang1/events{/privacy}",
"received_events_url": "https://api.github.com/users/felicitywang1/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
}
] | closed | false | null | [] | null | null | 2022-12-09T00:03:18 | 2023-01-12T21:38:35 | 2023-01-12T21:38:35 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | Hi, I'm running nmt traininig following https://github.com/NVIDIA/NeMo/pull/1738 with script
```
# in /group-volume/Bixby-Compression/sharing/felicity.wang/NeMo/examples/nlp/machine_translation
python3 enc_dec_nmt.py \
--config-path=conf \
--config-name=aayn_base \
trainer.devices=[0] \
~trainer.max_epochs \
+trainer.max_steps=100 \
model.beam_size=4 \
model.max_generation_delta=5 \
model.label_smoothing=0.1 \
model.encoder_tokenizer.vocab_size=32000 \
model.decoder_tokenizer.vocab_size=32000 \
+model.encoder_tokenizer.bpe_dropout=0.1 \
+model.decoder_tokenizer.bpe_dropout=0.1 \
model.preproc_out_dir=tmp_data/en_es_preproc \
model.train_ds.src_file_name=tmp_data/WikiMatrix.en-fr.langidfilter.lengthratio.bicleaner.60.dedup.moses.norm.tok.shuf.en \
model.train_ds.tgt_file_name=tmp_data/WikiMatrix.en-fr.langidfilter.lengthratio.bicleaner.60.dedup.moses.norm.tok.shuf.fr \
model.validation_ds.src_file_name=data/wmt14.en-fr.en \
model.validation_ds.tgt_file_name=data/wmt14.en-fr.fr \
model.test_ds.src_file_name=data/wmt13.en-fr.en \
model.test_ds.tgt_file_name=data/wmt13.en-fr.fr \
+use_tarred_dataset=True \
model.encoder.num_layers=6 \
model.encoder.hidden_size=256 \
model.encoder.inner_size=1024 \
model.encoder.num_attention_heads=8 \
model.encoder.ffn_dropout=0.1 \
model.decoder.num_layers=6 \
model.decoder.hidden_size=256 \
model.decoder.inner_size=1024 \
model.decoder.num_attention_heads=8 \
model.decoder.ffn_dropout=0.1 \
model.train_ds.tokens_in_batch=12500 \
model.validation_ds.tokens_in_batch=8192 \
model.optim.lr=0.001 \
model.optim.sched.warmup_ratio=0.05 \
+exp_manager.create_wandb_logger=True \
+exp_manager.wandb_logger_kwargs.name=TEST-nmt-base \
+exp_manager.wandb_logger_kwargs.project=nmt-de-en \
+exp_manager.create_checkpoint_callback=True \
+exp_manager.checkpoint_callback_params.monitor=val_sacreBLEU \
+exp_manager.exp_dir=nmt_base \
+exp_manager.checkpoint_callback_params.mode=max
```
trying to do BPE training and model training together within this script and got
```
...
Error executing job with overrides: ['trainer.devices=[0]', '~trainer.max_epochs', '+trainer.max_steps=100', 'model.beam_size=4', 'model.max_generation_delta=5', 'model.label_smoothing=0.1', 'model.encoder_tokenizer.vocab_size=32000', 'model.decoder_tokenizer.vocab_size=32000', '+model.encoder_tokenizer.bpe_dropout=0.1', '+model.decoder_tokenizer.bpe_dropout=0.1', 'model.preproc_out_dir=tmp_data/en_es_preproc', 'model.train_ds.src_file_name=tmp_data/WikiMatrix.en-fr.langidfilter.lengthratio.bicleaner.60.dedup.moses.norm.tok.shuf.en', 'model.train_ds.tgt_file_name=tmp_data/WikiMatrix.en-fr.langidfilter.lengthratio.bicleaner.60.dedup.moses.norm.tok.shuf.fr', 'model.validation_ds.src_file_name=data/wmt14.en-fr.en', 'model.validation_ds.tgt_file_name=data/wmt14.en-fr.fr', 'model.test_ds.src_file_name=data/wmt13.en-fr.en', 'model.test_ds.tgt_file_name=data/wmt13.en-fr.fr', '+use_tarred_dataset=True', 'model.encoder.num_layers=6', 'model.encoder.hidden_size=256', 'model.encoder.inner_size=1024', 'model.encoder.num_attention_heads=8', 'model.encoder.ffn_dropout=0.1', 'model.decoder.num_layers=6', 'model.decoder.hidden_size=256', 'model.decoder.inner_size=1024', 'model.decoder.num_attention_heads=8', 'model.decoder.ffn_dropout=0.1', 'model.train_ds.tokens_in_batch=12500', 'model.validation_ds.tokens_in_batch=8192', 'model.optim.lr=0.001', 'model.optim.sched.warmup_ratio=0.05', '+exp_manager.create_wandb_logger=True', '+exp_manager.wandb_logger_kwargs.name=TEST-nmt-base', '+exp_manager.wandb_logger_kwargs.project=nmt-de-en', '+exp_manager.create_checkpoint_callback=True', '+exp_manager.checkpoint_callback_params.monitor=val_sacreBLEU', '+exp_manager.exp_dir=nmt_base', '+exp_manager.checkpoint_callback_params.mode=max']
Traceback (most recent call last):
File "enc_dec_nmt.py", line 107, in main
cfg = update_model_config(default_cfg, cfg)
File "/home/user/.local/lib/python3.8/site-packages/nemo/utils/config_utils.py", line 105, in update_model_config
model_cfg = OmegaConf.merge(model_cls, update_cfg)
omegaconf.errors.ConfigKeyError: Key 'use_tarred_dataset' not in 'MTEncDecConfig'
full_key: use_tarred_dataset
object_type=MTEncDecConfig
```
My environments are
```
nemo-toolkit 1.13.0
hydra-core 1.1.2
omegaconf 2.1.2
```
Any guidance how I can solve this? Or could you provide the exact environment versions for training nmt?
Thank you. | {
"login": "felicitywang1",
"id": 96318866,
"node_id": "U_kgDOBb21kg",
"avatar_url": "https://avatars.githubusercontent.com/u/96318866?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/felicitywang1",
"html_url": "https://github.com/felicitywang1",
"followers_url": "https://api.github.com/users/felicitywang1/followers",
"following_url": "https://api.github.com/users/felicitywang1/following{/other_user}",
"gists_url": "https://api.github.com/users/felicitywang1/gists{/gist_id}",
"starred_url": "https://api.github.com/users/felicitywang1/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/felicitywang1/subscriptions",
"organizations_url": "https://api.github.com/users/felicitywang1/orgs",
"repos_url": "https://api.github.com/users/felicitywang1/repos",
"events_url": "https://api.github.com/users/felicitywang1/events{/privacy}",
"received_events_url": "https://api.github.com/users/felicitywang1/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5579/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5579/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5578 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5578/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5578/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5578/events | https://github.com/NVIDIA/NeMo/pull/5578 | 1,485,618,777 | PR_kwDOC_bI7s5E1_Gw | 5,578 | Add RETRO model documentation | {
"login": "yidong72",
"id": 43824965,
"node_id": "MDQ6VXNlcjQzODI0OTY1",
"avatar_url": "https://avatars.githubusercontent.com/u/43824965?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/yidong72",
"html_url": "https://github.com/yidong72",
"followers_url": "https://api.github.com/users/yidong72/followers",
"following_url": "https://api.github.com/users/yidong72/following{/other_user}",
"gists_url": "https://api.github.com/users/yidong72/gists{/gist_id}",
"starred_url": "https://api.github.com/users/yidong72/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/yidong72/subscriptions",
"organizations_url": "https://api.github.com/users/yidong72/orgs",
"repos_url": "https://api.github.com/users/yidong72/repos",
"events_url": "https://api.github.com/users/yidong72/events{/privacy}",
"received_events_url": "https://api.github.com/users/yidong72/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-08T23:54:46 | 2022-12-12T17:25:18 | 2022-12-12T17:25:15 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5578",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5578",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5578.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5578.patch",
"merged_at": "2022-12-12T17:25:15"
} | # What does this PR do ?
Add RETRO model documentation | {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5578/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5578/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5577 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5577/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5577/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5577/events | https://github.com/NVIDIA/NeMo/issues/5577 | 1,485,543,393 | I_kwDOC_bI7s5Yi5fh | 5,577 | AttributeError: 'EncDecSpeakerLabelModel' object has no attribute 'get_label' | {
"login": "egorsmkv",
"id": 7875085,
"node_id": "MDQ6VXNlcjc4NzUwODU=",
"avatar_url": "https://avatars.githubusercontent.com/u/7875085?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/egorsmkv",
"html_url": "https://github.com/egorsmkv",
"followers_url": "https://api.github.com/users/egorsmkv/followers",
"following_url": "https://api.github.com/users/egorsmkv/following{/other_user}",
"gists_url": "https://api.github.com/users/egorsmkv/gists{/gist_id}",
"starred_url": "https://api.github.com/users/egorsmkv/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/egorsmkv/subscriptions",
"organizations_url": "https://api.github.com/users/egorsmkv/orgs",
"repos_url": "https://api.github.com/users/egorsmkv/repos",
"events_url": "https://api.github.com/users/egorsmkv/events{/privacy}",
"received_events_url": "https://api.github.com/users/egorsmkv/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
},
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | null | [] | null | null | 2022-12-08T22:58:38 | 2023-03-09T02:04:52 | 2023-03-09T02:04:52 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
I cannot get the language of a file with AmberNet
**Steps/Code to reproduce bug**
Code:
```python
import nemo
from nemo.collections.asr.models import EncDecSpeakerLabelModel
model = EncDecSpeakerLabelModel.from_pretrained('langid_ambernet')
print('NeMO:', nemo.__version__)
print('AmberNet method:', model.get_label)
```
Result:
```
...
NeMO: 1.13.0
Traceback (most recent call last):
File "/Users/yehorsmoliakov/Work/egorsmkv/asr-corpus-creator/rep.py", line 8, in <module>
print('AmberNet method:', model.get_label)
File "/Users/yehorsmoliakov/.local/share/virtualenvs/asr-corpus-creator-Sp_hMA78/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1265, in __getattr__
raise AttributeError("'{}' object has no attribute '{}'".format(
AttributeError: 'EncDecSpeakerLabelModel' object has no attribute 'get_label'
```
**Expected behavior**
Info about the method `get_label`
| {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5577/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5577/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5576 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5576/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5576/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5576/events | https://github.com/NVIDIA/NeMo/pull/5576 | 1,485,493,834 | PR_kwDOC_bI7s5E1jZC | 5,576 | MegatronNMT: updated decode function to support beam search | {
"login": "Davood-M",
"id": 3964285,
"node_id": "MDQ6VXNlcjM5NjQyODU=",
"avatar_url": "https://avatars.githubusercontent.com/u/3964285?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Davood-M",
"html_url": "https://github.com/Davood-M",
"followers_url": "https://api.github.com/users/Davood-M/followers",
"following_url": "https://api.github.com/users/Davood-M/following{/other_user}",
"gists_url": "https://api.github.com/users/Davood-M/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Davood-M/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Davood-M/subscriptions",
"organizations_url": "https://api.github.com/users/Davood-M/orgs",
"repos_url": "https://api.github.com/users/Davood-M/repos",
"events_url": "https://api.github.com/users/Davood-M/events{/privacy}",
"received_events_url": "https://api.github.com/users/Davood-M/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
},
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-08T22:11:57 | 2023-02-20T17:51:29 | 2023-01-20T02:00:36 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | true | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5576",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5576",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5576.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5576.patch",
"merged_at": null
} | Signed-off-by: David Mosallanezhad <[email protected]>
# What does this PR do ?
Adds simple beam search to the lm_encoder_decoder's decode function
**Collection**: [Note which collection this PR will affect]
# Changelog
- Updated decode function
# Usage
* You can specify beam search size and the penalty function parameter
```bash
python NeMo/examples/nlp/machine_translation/nmt_transformer_infer_megatron.py --config-path=conf --config-name="nmt_megatron_infer" model_file=./megatron_nmt.nemo srctext=./valid.de-en.en tgtout=./output_cache.txt source_lang=en target_lang=de batch_size=32 beam_size=6
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5576/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5576/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5575 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5575/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5575/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5575/events | https://github.com/NVIDIA/NeMo/pull/5575 | 1,485,461,498 | PR_kwDOC_bI7s5E1cHs | 5,575 | [Add] ASR+VAD Inference Pipeline | {
"login": "stevehuang52",
"id": 105218074,
"node_id": "U_kgDOBkWAGg",
"avatar_url": "https://avatars.githubusercontent.com/u/105218074?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stevehuang52",
"html_url": "https://github.com/stevehuang52",
"followers_url": "https://api.github.com/users/stevehuang52/followers",
"following_url": "https://api.github.com/users/stevehuang52/following{/other_user}",
"gists_url": "https://api.github.com/users/stevehuang52/gists{/gist_id}",
"starred_url": "https://api.github.com/users/stevehuang52/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/stevehuang52/subscriptions",
"organizations_url": "https://api.github.com/users/stevehuang52/orgs",
"repos_url": "https://api.github.com/users/stevehuang52/repos",
"events_url": "https://api.github.com/users/stevehuang52/events{/privacy}",
"received_events_url": "https://api.github.com/users/stevehuang52/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1719393562,
"node_id": "MDU6TGFiZWwxNzE5MzkzNTYy",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/feature",
"name": "feature",
"color": "68F3F0",
"default": false,
"description": "request/PR for a new feature"
},
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
},
{
"id": 4847373924,
"node_id": "LA_kwDOC_bI7s8AAAABIO0OZA",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/common",
"name": "common",
"color": "c5def5",
"default": false,
"description": ""
}
] | closed | false | {
"login": "stevehuang52",
"id": 105218074,
"node_id": "U_kgDOBkWAGg",
"avatar_url": "https://avatars.githubusercontent.com/u/105218074?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stevehuang52",
"html_url": "https://github.com/stevehuang52",
"followers_url": "https://api.github.com/users/stevehuang52/followers",
"following_url": "https://api.github.com/users/stevehuang52/following{/other_user}",
"gists_url": "https://api.github.com/users/stevehuang52/gists{/gist_id}",
"starred_url": "https://api.github.com/users/stevehuang52/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/stevehuang52/subscriptions",
"organizations_url": "https://api.github.com/users/stevehuang52/orgs",
"repos_url": "https://api.github.com/users/stevehuang52/repos",
"events_url": "https://api.github.com/users/stevehuang52/events{/privacy}",
"received_events_url": "https://api.github.com/users/stevehuang52/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "stevehuang52",
"id": 105218074,
"node_id": "U_kgDOBkWAGg",
"avatar_url": "https://avatars.githubusercontent.com/u/105218074?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stevehuang52",
"html_url": "https://github.com/stevehuang52",
"followers_url": "https://api.github.com/users/stevehuang52/followers",
"following_url": "https://api.github.com/users/stevehuang52/following{/other_user}",
"gists_url": "https://api.github.com/users/stevehuang52/gists{/gist_id}",
"starred_url": "https://api.github.com/users/stevehuang52/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/stevehuang52/subscriptions",
"organizations_url": "https://api.github.com/users/stevehuang52/orgs",
"repos_url": "https://api.github.com/users/stevehuang52/repos",
"events_url": "https://api.github.com/users/stevehuang52/events{/privacy}",
"received_events_url": "https://api.github.com/users/stevehuang52/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-12-08T21:48:03 | 2022-12-16T15:17:08 | 2022-12-16T15:16:07 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5575",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5575",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5575.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5575.patch",
"merged_at": "2022-12-16T15:16:07"
} | # What does this PR do ?
This PR aims to provide a pipeline for integrating Voice Activity Detection (VAD) results in ASR inference.
**Collection**: [ASR]
# Changelog
- Added example folder `examples/asr/asr_vad`
- Added `feature_to_label.py` and `feature_to_label_dataset.py`
- Added `feature_to_text.py` and `feature_to_text_dataset.py`
- Added classes to `collections.py`
- Updated `manifest.py`
# Usage
To run the code with ASR+VAD default settings:
```bash
python speech_to_text_with_vad.py \
manifest_filepath=/PATH/TO/MANIFEST.json \
vad_model=vad_multilingual_marblenet \
asr_model=stt_en_conformer_ctc_large \
vad_config=../conf/vad/vad_inference_postprocess.yaml
```
To use only ASR and disable VAD, set `vad_model=None` and `use_rttm=False`.
To use only VAD, set `asr_model=None` and specify both `vad_model` and `vad_config`.
To enable profiling, set `profiling=True`, but this will significantly slow down the program.
To use or disable feature masking, set `use_rttm` to `True` or `False`.
To normalize feature before masking, set `normalize=pre_norm`,
and set `normalize=post_norm` for masking before normalization.
To use a specific value for feature masking, set `feat_mask_val` to the desired value.
Default is `feat_mask_val=None`, where -16.530 (zero mel-spectralgram value) will be used for `post_norm` and 0 (same as SpectralAugmentation) will be used for `pre_norm`.
See more options in the `InferenceConfig` class.
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [x] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "stevehuang52",
"id": 105218074,
"node_id": "U_kgDOBkWAGg",
"avatar_url": "https://avatars.githubusercontent.com/u/105218074?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stevehuang52",
"html_url": "https://github.com/stevehuang52",
"followers_url": "https://api.github.com/users/stevehuang52/followers",
"following_url": "https://api.github.com/users/stevehuang52/following{/other_user}",
"gists_url": "https://api.github.com/users/stevehuang52/gists{/gist_id}",
"starred_url": "https://api.github.com/users/stevehuang52/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/stevehuang52/subscriptions",
"organizations_url": "https://api.github.com/users/stevehuang52/orgs",
"repos_url": "https://api.github.com/users/stevehuang52/repos",
"events_url": "https://api.github.com/users/stevehuang52/events{/privacy}",
"received_events_url": "https://api.github.com/users/stevehuang52/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5575/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5575/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5574 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5574/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5574/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5574/events | https://github.com/NVIDIA/NeMo/pull/5574 | 1,485,291,480 | PR_kwDOC_bI7s5E01_S | 5,574 | Remove SDP (moved to separate repo) | {
"login": "erastorgueva-nv",
"id": 80532067,
"node_id": "MDQ6VXNlcjgwNTMyMDY3",
"avatar_url": "https://avatars.githubusercontent.com/u/80532067?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/erastorgueva-nv",
"html_url": "https://github.com/erastorgueva-nv",
"followers_url": "https://api.github.com/users/erastorgueva-nv/followers",
"following_url": "https://api.github.com/users/erastorgueva-nv/following{/other_user}",
"gists_url": "https://api.github.com/users/erastorgueva-nv/gists{/gist_id}",
"starred_url": "https://api.github.com/users/erastorgueva-nv/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/erastorgueva-nv/subscriptions",
"organizations_url": "https://api.github.com/users/erastorgueva-nv/orgs",
"repos_url": "https://api.github.com/users/erastorgueva-nv/repos",
"events_url": "https://api.github.com/users/erastorgueva-nv/events{/privacy}",
"received_events_url": "https://api.github.com/users/erastorgueva-nv/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-12-08T19:44:29 | 2022-12-14T23:06:40 | 2022-12-14T18:42:33 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5574",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5574",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5574.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5574.patch",
"merged_at": null
} | # What does this PR do ?
Remove SDP code from tools/ folder and add docs page briefly describing SDP and its new location
**Collection**: tools
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
N/A
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ N/A ] Did you write any new necessary tests?
- [x] Did you add or update any necessary documentation?
- [ N/A ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [x] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5574/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5574/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5573 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5573/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5573/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5573/events | https://github.com/NVIDIA/NeMo/pull/5573 | 1,484,623,124 | PR_kwDOC_bI7s5EygiH | 5,573 | optimized loop and bugfix | {
"login": "Jorjeous",
"id": 37293288,
"node_id": "MDQ6VXNlcjM3MjkzMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/37293288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Jorjeous",
"html_url": "https://github.com/Jorjeous",
"followers_url": "https://api.github.com/users/Jorjeous/followers",
"following_url": "https://api.github.com/users/Jorjeous/following{/other_user}",
"gists_url": "https://api.github.com/users/Jorjeous/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Jorjeous/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Jorjeous/subscriptions",
"organizations_url": "https://api.github.com/users/Jorjeous/orgs",
"repos_url": "https://api.github.com/users/Jorjeous/repos",
"events_url": "https://api.github.com/users/Jorjeous/events{/privacy}",
"received_events_url": "https://api.github.com/users/Jorjeous/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
},
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | {
"login": "Jorjeous",
"id": 37293288,
"node_id": "MDQ6VXNlcjM3MjkzMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/37293288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Jorjeous",
"html_url": "https://github.com/Jorjeous",
"followers_url": "https://api.github.com/users/Jorjeous/followers",
"following_url": "https://api.github.com/users/Jorjeous/following{/other_user}",
"gists_url": "https://api.github.com/users/Jorjeous/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Jorjeous/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Jorjeous/subscriptions",
"organizations_url": "https://api.github.com/users/Jorjeous/orgs",
"repos_url": "https://api.github.com/users/Jorjeous/repos",
"events_url": "https://api.github.com/users/Jorjeous/events{/privacy}",
"received_events_url": "https://api.github.com/users/Jorjeous/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "Jorjeous",
"id": 37293288,
"node_id": "MDQ6VXNlcjM3MjkzMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/37293288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Jorjeous",
"html_url": "https://github.com/Jorjeous",
"followers_url": "https://api.github.com/users/Jorjeous/followers",
"following_url": "https://api.github.com/users/Jorjeous/following{/other_user}",
"gists_url": "https://api.github.com/users/Jorjeous/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Jorjeous/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Jorjeous/subscriptions",
"organizations_url": "https://api.github.com/users/Jorjeous/orgs",
"repos_url": "https://api.github.com/users/Jorjeous/repos",
"events_url": "https://api.github.com/users/Jorjeous/events{/privacy}",
"received_events_url": "https://api.github.com/users/Jorjeous/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-12-08T13:17:00 | 2022-12-08T20:04:09 | 2022-12-08T20:04:01 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5573",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5573",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5573.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5573.patch",
"merged_at": "2022-12-08T20:04:00"
} | Signed-off-by: George Zelenfroynd <[email protected]>
# What does this PR do ?
Bugfix and code optimization
Add a one line overview of what this PR aims to accomplish.
**Collection**: [Note which collection this PR will affect]
ASR
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [x] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "vsl9",
"id": 4344862,
"node_id": "MDQ6VXNlcjQzNDQ4NjI=",
"avatar_url": "https://avatars.githubusercontent.com/u/4344862?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/vsl9",
"html_url": "https://github.com/vsl9",
"followers_url": "https://api.github.com/users/vsl9/followers",
"following_url": "https://api.github.com/users/vsl9/following{/other_user}",
"gists_url": "https://api.github.com/users/vsl9/gists{/gist_id}",
"starred_url": "https://api.github.com/users/vsl9/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/vsl9/subscriptions",
"organizations_url": "https://api.github.com/users/vsl9/orgs",
"repos_url": "https://api.github.com/users/vsl9/repos",
"events_url": "https://api.github.com/users/vsl9/events{/privacy}",
"received_events_url": "https://api.github.com/users/vsl9/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5573/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5573/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5572 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5572/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5572/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5572/events | https://github.com/NVIDIA/NeMo/pull/5572 | 1,483,886,774 | PR_kwDOC_bI7s5Ev4ve | 5,572 | Fixed RadTTS unit test | {
"login": "borisfom",
"id": 14189615,
"node_id": "MDQ6VXNlcjE0MTg5NjE1",
"avatar_url": "https://avatars.githubusercontent.com/u/14189615?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/borisfom",
"html_url": "https://github.com/borisfom",
"followers_url": "https://api.github.com/users/borisfom/followers",
"following_url": "https://api.github.com/users/borisfom/following{/other_user}",
"gists_url": "https://api.github.com/users/borisfom/gists{/gist_id}",
"starred_url": "https://api.github.com/users/borisfom/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/borisfom/subscriptions",
"organizations_url": "https://api.github.com/users/borisfom/orgs",
"repos_url": "https://api.github.com/users/borisfom/repos",
"events_url": "https://api.github.com/users/borisfom/events{/privacy}",
"received_events_url": "https://api.github.com/users/borisfom/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4418253970,
"node_id": "LA_kwDOC_bI7s8AAAABB1k0kg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/TTS",
"name": "TTS",
"color": "49AF28",
"default": false,
"description": ""
}
] | closed | false | {
"login": "XuesongYang",
"id": 1646669,
"node_id": "MDQ6VXNlcjE2NDY2Njk=",
"avatar_url": "https://avatars.githubusercontent.com/u/1646669?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/XuesongYang",
"html_url": "https://github.com/XuesongYang",
"followers_url": "https://api.github.com/users/XuesongYang/followers",
"following_url": "https://api.github.com/users/XuesongYang/following{/other_user}",
"gists_url": "https://api.github.com/users/XuesongYang/gists{/gist_id}",
"starred_url": "https://api.github.com/users/XuesongYang/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/XuesongYang/subscriptions",
"organizations_url": "https://api.github.com/users/XuesongYang/orgs",
"repos_url": "https://api.github.com/users/XuesongYang/repos",
"events_url": "https://api.github.com/users/XuesongYang/events{/privacy}",
"received_events_url": "https://api.github.com/users/XuesongYang/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "XuesongYang",
"id": 1646669,
"node_id": "MDQ6VXNlcjE2NDY2Njk=",
"avatar_url": "https://avatars.githubusercontent.com/u/1646669?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/XuesongYang",
"html_url": "https://github.com/XuesongYang",
"followers_url": "https://api.github.com/users/XuesongYang/followers",
"following_url": "https://api.github.com/users/XuesongYang/following{/other_user}",
"gists_url": "https://api.github.com/users/XuesongYang/gists{/gist_id}",
"starred_url": "https://api.github.com/users/XuesongYang/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/XuesongYang/subscriptions",
"organizations_url": "https://api.github.com/users/XuesongYang/orgs",
"repos_url": "https://api.github.com/users/XuesongYang/repos",
"events_url": "https://api.github.com/users/XuesongYang/events{/privacy}",
"received_events_url": "https://api.github.com/users/XuesongYang/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-12-08T07:01:48 | 2022-12-15T02:13:12 | 2022-12-15T02:13:12 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5572",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5572",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5572.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5572.patch",
"merged_at": "2022-12-15T02:13:12"
} | Signed-off-by: Boris Fomitchev <[email protected]>
# What does this PR do ?
It looks like (due to updates to internals?) RadTTS now requires a bit higher minimum length of the input.
I will investigate some more but that fix would unblock the test.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "borisfom",
"id": 14189615,
"node_id": "MDQ6VXNlcjE0MTg5NjE1",
"avatar_url": "https://avatars.githubusercontent.com/u/14189615?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/borisfom",
"html_url": "https://github.com/borisfom",
"followers_url": "https://api.github.com/users/borisfom/followers",
"following_url": "https://api.github.com/users/borisfom/following{/other_user}",
"gists_url": "https://api.github.com/users/borisfom/gists{/gist_id}",
"starred_url": "https://api.github.com/users/borisfom/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/borisfom/subscriptions",
"organizations_url": "https://api.github.com/users/borisfom/orgs",
"repos_url": "https://api.github.com/users/borisfom/repos",
"events_url": "https://api.github.com/users/borisfom/events{/privacy}",
"received_events_url": "https://api.github.com/users/borisfom/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5572/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5572/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5571 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5571/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5571/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5571/events | https://github.com/NVIDIA/NeMo/pull/5571 | 1,483,254,941 | PR_kwDOC_bI7s5EtncI | 5,571 | NeMo Forced Aligner | {
"login": "erastorgueva-nv",
"id": 80532067,
"node_id": "MDQ6VXNlcjgwNTMyMDY3",
"avatar_url": "https://avatars.githubusercontent.com/u/80532067?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/erastorgueva-nv",
"html_url": "https://github.com/erastorgueva-nv",
"followers_url": "https://api.github.com/users/erastorgueva-nv/followers",
"following_url": "https://api.github.com/users/erastorgueva-nv/following{/other_user}",
"gists_url": "https://api.github.com/users/erastorgueva-nv/gists{/gist_id}",
"starred_url": "https://api.github.com/users/erastorgueva-nv/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/erastorgueva-nv/subscriptions",
"organizations_url": "https://api.github.com/users/erastorgueva-nv/orgs",
"repos_url": "https://api.github.com/users/erastorgueva-nv/repos",
"events_url": "https://api.github.com/users/erastorgueva-nv/events{/privacy}",
"received_events_url": "https://api.github.com/users/erastorgueva-nv/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-12-08T01:01:48 | 2023-01-17T22:09:15 | 2023-01-17T21:36:22 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5571",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5571",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5571.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5571.patch",
"merged_at": "2023-01-17T21:36:22"
} | Signed-off-by: Elena Rastorgueva <[email protected]>
# What does this PR do ?
Add NFA - a tool to do forced alignment using NeMo CTC-based models.
**Collection**: [Note which collection this PR will affect]
Tools
# Changelog
- Add NeMo/tools/nemo_forced_aligner/
# Usage
* You can potentially add a usage example below
```bash
python $NEMO_PATH/tools/nemo_forced_aligner/align.py \
pretrained_name="stt_en_citrinet_1024_gamma_0_25" \
model_downsample_factor=8 \
manifest_filepath="manifest.json" \
output_dir="nfa_ctms/" \
batch_size=1 \
additional_ctm_grouping_separator="|" \
audio_filepath_parts_in_utt_id=1
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ N/A ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ N/A ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5571/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5571/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5570 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5570/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5570/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5570/events | https://github.com/NVIDIA/NeMo/pull/5570 | 1,482,973,520 | PR_kwDOC_bI7s5Esmqr | 5,570 | Merge r1.13.0 main | {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-07T22:17:13 | 2022-12-08T03:35:50 | 2022-12-08T03:35:47 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5570",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5570",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5570.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5570.patch",
"merged_at": "2022-12-08T03:35:47"
} | # What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5570/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5570/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5569 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5569/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5569/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5569/events | https://github.com/NVIDIA/NeMo/pull/5569 | 1,482,899,455 | PR_kwDOC_bI7s5EsVu1 | 5,569 | Add an option to defer data setup from ``__init__`` to ``setup`` | {
"login": "anteju",
"id": 108555623,
"node_id": "U_kgDOBnhtZw",
"avatar_url": "https://avatars.githubusercontent.com/u/108555623?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/anteju",
"html_url": "https://github.com/anteju",
"followers_url": "https://api.github.com/users/anteju/followers",
"following_url": "https://api.github.com/users/anteju/following{/other_user}",
"gists_url": "https://api.github.com/users/anteju/gists{/gist_id}",
"starred_url": "https://api.github.com/users/anteju/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/anteju/subscriptions",
"organizations_url": "https://api.github.com/users/anteju/orgs",
"repos_url": "https://api.github.com/users/anteju/repos",
"events_url": "https://api.github.com/users/anteju/events{/privacy}",
"received_events_url": "https://api.github.com/users/anteju/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1923495043,
"node_id": "MDU6TGFiZWwxOTIzNDk1MDQz",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/core",
"name": "core",
"color": "f25e3a",
"default": false,
"description": "Changes to NeMo Core"
},
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-07T21:32:13 | 2022-12-09T21:14:12 | 2022-12-09T21:14:12 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5569",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5569",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5569.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5569.patch",
"merged_at": "2022-12-09T21:14:12"
} | Signed-off-by: Ante Jukić <[email protected]>
# What does this PR do ?
This PR adds an option to defer dataloader setup in `ModelPT` from `__init__` to `setup`.
Default behavior remains unchanged.
**Collection**: Core
# Changelog
- In `ModelPT::__init__()`
- Nothing is changed if `self._cfg.train_ds.defer_setup` is not set or is set to `False` (default)
- If `self._cfg.train_ds.defer_setup` is set to `True`, `self.setup_training_data` will be skipped
- Analogous behavior for validation and test data
- In `ModelPT::setup()`
- If dataloader is not initialized and `defer_setup==True`, run data setup
# Usage
To use deferred setup, set the following
```python
cfg.model.train_ds.defer_setup = True
cfg.model.validation_ds.defer_setup = True
cfg.model.test_ds.defer_setup = True
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5569/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5569/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5568 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5568/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5568/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5568/events | https://github.com/NVIDIA/NeMo/pull/5568 | 1,482,578,272 | PR_kwDOC_bI7s5ErMqS | 5,568 | Add support for Time Stamp calculation using transcribe_speech.py | {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-07T18:31:25 | 2022-12-07T19:16:52 | 2022-12-07T19:16:48 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5568",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5568",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5568.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5568.patch",
"merged_at": "2022-12-07T19:16:48"
} | Signed-off-by: smajumdar <[email protected]>
# What does this PR do ?
Adds flag to compute word and char level timestamp using the transcribe speech script.
**Collection**: [ASR]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
```python
python transcribe_speech.py \
model_path=null \
pretrained_name=null \
audio_dir="<remove or path to folder of audio files>" \
dataset_manifest="<remove or path to manifest>" \
output_filename="<remove or specify output filename>" \
batch_size=32 \
compute_timestamps=True \
compute_langs=False \
cuda=0 \
amp=True
(Optionally)
ctc_decoding.ctc_timestamp_type="all" # (default all, can be [all, char, word])
rnnt_decoding.rnnt_timestamp_type="all" # (default all, can be [all, char, word])
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [x] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
| {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5568/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5568/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5567 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5567/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5567/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5567/events | https://github.com/NVIDIA/NeMo/pull/5567 | 1,482,538,833 | PR_kwDOC_bI7s5ErEAI | 5,567 | Vits exp | {
"login": "treacker",
"id": 36159472,
"node_id": "MDQ6VXNlcjM2MTU5NDcy",
"avatar_url": "https://avatars.githubusercontent.com/u/36159472?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/treacker",
"html_url": "https://github.com/treacker",
"followers_url": "https://api.github.com/users/treacker/followers",
"following_url": "https://api.github.com/users/treacker/following{/other_user}",
"gists_url": "https://api.github.com/users/treacker/gists{/gist_id}",
"starred_url": "https://api.github.com/users/treacker/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/treacker/subscriptions",
"organizations_url": "https://api.github.com/users/treacker/orgs",
"repos_url": "https://api.github.com/users/treacker/repos",
"events_url": "https://api.github.com/users/treacker/events{/privacy}",
"received_events_url": "https://api.github.com/users/treacker/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4418253970,
"node_id": "LA_kwDOC_bI7s8AAAABB1k0kg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/TTS",
"name": "TTS",
"color": "49AF28",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-07T18:12:15 | 2022-12-12T20:09:34 | 2022-12-12T20:09:33 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5567",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5567",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5567.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5567.patch",
"merged_at": null
} | # What does this PR do ?
Implementation of VITS model
**Collection**: [tts]
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "treacker",
"id": 36159472,
"node_id": "MDQ6VXNlcjM2MTU5NDcy",
"avatar_url": "https://avatars.githubusercontent.com/u/36159472?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/treacker",
"html_url": "https://github.com/treacker",
"followers_url": "https://api.github.com/users/treacker/followers",
"following_url": "https://api.github.com/users/treacker/following{/other_user}",
"gists_url": "https://api.github.com/users/treacker/gists{/gist_id}",
"starred_url": "https://api.github.com/users/treacker/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/treacker/subscriptions",
"organizations_url": "https://api.github.com/users/treacker/orgs",
"repos_url": "https://api.github.com/users/treacker/repos",
"events_url": "https://api.github.com/users/treacker/events{/privacy}",
"received_events_url": "https://api.github.com/users/treacker/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5567/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5567/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5566 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5566/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5566/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5566/events | https://github.com/NVIDIA/NeMo/pull/5566 | 1,482,527,365 | PR_kwDOC_bI7s5ErBXf | 5,566 | Update torchmetrics | {
"login": "nithinraok",
"id": 19668129,
"node_id": "MDQ6VXNlcjE5NjY4MTI5",
"avatar_url": "https://avatars.githubusercontent.com/u/19668129?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/nithinraok",
"html_url": "https://github.com/nithinraok",
"followers_url": "https://api.github.com/users/nithinraok/followers",
"following_url": "https://api.github.com/users/nithinraok/following{/other_user}",
"gists_url": "https://api.github.com/users/nithinraok/gists{/gist_id}",
"starred_url": "https://api.github.com/users/nithinraok/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/nithinraok/subscriptions",
"organizations_url": "https://api.github.com/users/nithinraok/orgs",
"repos_url": "https://api.github.com/users/nithinraok/repos",
"events_url": "https://api.github.com/users/nithinraok/events{/privacy}",
"received_events_url": "https://api.github.com/users/nithinraok/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-07T18:06:31 | 2022-12-08T21:57:13 | 2022-12-08T21:56:19 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5566",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5566",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5566.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5566.patch",
"merged_at": "2022-12-08T21:56:19"
} | Signed-off-by: nithinraok <[email protected]>
# What does this PR do ?
Update torch metrics version
**Collection**: Speaker Tasks
# Change Log
Torch metrics updated their API to use `final_state` instead of `get_final_stat`.
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [x] Bugfix
- [ ] Documentation
| {
"login": "fayejf",
"id": 36722593,
"node_id": "MDQ6VXNlcjM2NzIyNTkz",
"avatar_url": "https://avatars.githubusercontent.com/u/36722593?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/fayejf",
"html_url": "https://github.com/fayejf",
"followers_url": "https://api.github.com/users/fayejf/followers",
"following_url": "https://api.github.com/users/fayejf/following{/other_user}",
"gists_url": "https://api.github.com/users/fayejf/gists{/gist_id}",
"starred_url": "https://api.github.com/users/fayejf/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/fayejf/subscriptions",
"organizations_url": "https://api.github.com/users/fayejf/orgs",
"repos_url": "https://api.github.com/users/fayejf/repos",
"events_url": "https://api.github.com/users/fayejf/events{/privacy}",
"received_events_url": "https://api.github.com/users/fayejf/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5566/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5566/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5565 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5565/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5565/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5565/events | https://github.com/NVIDIA/NeMo/pull/5565 | 1,482,257,406 | PR_kwDOC_bI7s5EqEZ4 | 5,565 | [TTS] GAN-based spectrogram enhancer | {
"login": "racoiaws",
"id": 8864149,
"node_id": "MDQ6VXNlcjg4NjQxNDk=",
"avatar_url": "https://avatars.githubusercontent.com/u/8864149?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/racoiaws",
"html_url": "https://github.com/racoiaws",
"followers_url": "https://api.github.com/users/racoiaws/followers",
"following_url": "https://api.github.com/users/racoiaws/following{/other_user}",
"gists_url": "https://api.github.com/users/racoiaws/gists{/gist_id}",
"starred_url": "https://api.github.com/users/racoiaws/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/racoiaws/subscriptions",
"organizations_url": "https://api.github.com/users/racoiaws/orgs",
"repos_url": "https://api.github.com/users/racoiaws/repos",
"events_url": "https://api.github.com/users/racoiaws/events{/privacy}",
"received_events_url": "https://api.github.com/users/racoiaws/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4418253970,
"node_id": "LA_kwDOC_bI7s8AAAABB1k0kg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/TTS",
"name": "TTS",
"color": "49AF28",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-07T15:58:07 | 2023-01-30T18:41:28 | 2023-01-30T18:41:28 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5565",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5565",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5565.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5565.patch",
"merged_at": "2023-01-30T18:41:28"
} | # What does this PR do ?
Add a GAN-based spectrogram enhancer to NeMo.
The intended purpose is to add details to blurry spectrograms from TTS models like FastPitch.
**Collection**: TTS
# Changelog
- Add the spectrogram enhancer NeMo model
- Add example config and training script
# Usage
Given a trained model,
```python
model = SpectrogramEnhancer.restore_from("checkpoint.nemo")
with torch.no_grad():
# spectrograms :: B x 80 x L
# lengths :: B - the usual LongTensor with lengths or every spectrogram in the batch
enhanced_spectrograms = model.forward(condition=spectrograms, lengths=lengths)
# enhanced_spectrograms = B x 80 x L
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [x] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Based on https://github.com/lucidrains/stylegan2-pytorch (MIT License)
| {
"login": "racoiaws",
"id": 8864149,
"node_id": "MDQ6VXNlcjg4NjQxNDk=",
"avatar_url": "https://avatars.githubusercontent.com/u/8864149?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/racoiaws",
"html_url": "https://github.com/racoiaws",
"followers_url": "https://api.github.com/users/racoiaws/followers",
"following_url": "https://api.github.com/users/racoiaws/following{/other_user}",
"gists_url": "https://api.github.com/users/racoiaws/gists{/gist_id}",
"starred_url": "https://api.github.com/users/racoiaws/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/racoiaws/subscriptions",
"organizations_url": "https://api.github.com/users/racoiaws/orgs",
"repos_url": "https://api.github.com/users/racoiaws/repos",
"events_url": "https://api.github.com/users/racoiaws/events{/privacy}",
"received_events_url": "https://api.github.com/users/racoiaws/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5565/reactions",
"total_count": 1,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 1,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5565/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5564 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5564/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5564/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5564/events | https://github.com/NVIDIA/NeMo/pull/5564 | 1,482,218,019 | PR_kwDOC_bI7s5Ep7lj | 5,564 | Fix all gather while writing to a file during T5 finetuning | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4556669155,
"node_id": "LA_kwDOC_bI7s8AAAABD5lA4w",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/cherry-pick",
"name": "cherry-pick",
"color": "ededed",
"default": false,
"description": null
},
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-07T15:40:08 | 2022-12-07T17:44:35 | 2022-12-07T17:44:32 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5564",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5564",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5564.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5564.patch",
"merged_at": "2022-12-07T17:44:32"
} | # What does this PR do ?
Fixes the all-gather logic to only gather across data parallel ranks to write to a file. Previously, examples would be duplicated in the file since all model parallel ranks have the same data.
**Collection**: NLP
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5564/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5564/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5563 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5563/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5563/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5563/events | https://github.com/NVIDIA/NeMo/pull/5563 | 1,481,520,610 | PR_kwDOC_bI7s5EneVG | 5,563 | get nemo working on my laptop | {
"login": "coopie",
"id": 6509132,
"node_id": "MDQ6VXNlcjY1MDkxMzI=",
"avatar_url": "https://avatars.githubusercontent.com/u/6509132?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/coopie",
"html_url": "https://github.com/coopie",
"followers_url": "https://api.github.com/users/coopie/followers",
"following_url": "https://api.github.com/users/coopie/following{/other_user}",
"gists_url": "https://api.github.com/users/coopie/gists{/gist_id}",
"starred_url": "https://api.github.com/users/coopie/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/coopie/subscriptions",
"organizations_url": "https://api.github.com/users/coopie/orgs",
"repos_url": "https://api.github.com/users/coopie/repos",
"events_url": "https://api.github.com/users/coopie/events{/privacy}",
"received_events_url": "https://api.github.com/users/coopie/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4418253970,
"node_id": "LA_kwDOC_bI7s8AAAABB1k0kg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/TTS",
"name": "TTS",
"color": "49AF28",
"default": false,
"description": ""
},
{
"id": 4847373924,
"node_id": "LA_kwDOC_bI7s8AAAABIO0OZA",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/common",
"name": "common",
"color": "c5def5",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-07T10:06:19 | 2022-12-07T10:07:52 | 2022-12-07T10:06:30 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5563",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5563",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5563.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5563.patch",
"merged_at": null
} | null | {
"login": "coopie",
"id": 6509132,
"node_id": "MDQ6VXNlcjY1MDkxMzI=",
"avatar_url": "https://avatars.githubusercontent.com/u/6509132?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/coopie",
"html_url": "https://github.com/coopie",
"followers_url": "https://api.github.com/users/coopie/followers",
"following_url": "https://api.github.com/users/coopie/following{/other_user}",
"gists_url": "https://api.github.com/users/coopie/gists{/gist_id}",
"starred_url": "https://api.github.com/users/coopie/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/coopie/subscriptions",
"organizations_url": "https://api.github.com/users/coopie/orgs",
"repos_url": "https://api.github.com/users/coopie/repos",
"events_url": "https://api.github.com/users/coopie/events{/privacy}",
"received_events_url": "https://api.github.com/users/coopie/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5563/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5563/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5562 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5562/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5562/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5562/events | https://github.com/NVIDIA/NeMo/pull/5562 | 1,481,378,844 | PR_kwDOC_bI7s5Em-jv | 5,562 | Fixed Docker build | {
"login": "borisfom",
"id": 14189615,
"node_id": "MDQ6VXNlcjE0MTg5NjE1",
"avatar_url": "https://avatars.githubusercontent.com/u/14189615?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/borisfom",
"html_url": "https://github.com/borisfom",
"followers_url": "https://api.github.com/users/borisfom/followers",
"following_url": "https://api.github.com/users/borisfom/following{/other_user}",
"gists_url": "https://api.github.com/users/borisfom/gists{/gist_id}",
"starred_url": "https://api.github.com/users/borisfom/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/borisfom/subscriptions",
"organizations_url": "https://api.github.com/users/borisfom/orgs",
"repos_url": "https://api.github.com/users/borisfom/repos",
"events_url": "https://api.github.com/users/borisfom/events{/privacy}",
"received_events_url": "https://api.github.com/users/borisfom/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4840216855,
"node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI",
"name": "CI",
"color": "1D76DB",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-07T08:58:12 | 2022-12-07T18:18:01 | 2022-12-07T18:18:00 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5562",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5562",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5562.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5562.patch",
"merged_at": "2022-12-07T18:18:00"
} | Signed-off-by: Boris Fomitchev <[email protected]>
# What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "Davood-M",
"id": 3964285,
"node_id": "MDQ6VXNlcjM5NjQyODU=",
"avatar_url": "https://avatars.githubusercontent.com/u/3964285?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Davood-M",
"html_url": "https://github.com/Davood-M",
"followers_url": "https://api.github.com/users/Davood-M/followers",
"following_url": "https://api.github.com/users/Davood-M/following{/other_user}",
"gists_url": "https://api.github.com/users/Davood-M/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Davood-M/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Davood-M/subscriptions",
"organizations_url": "https://api.github.com/users/Davood-M/orgs",
"repos_url": "https://api.github.com/users/Davood-M/repos",
"events_url": "https://api.github.com/users/Davood-M/events{/privacy}",
"received_events_url": "https://api.github.com/users/Davood-M/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5562/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5562/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5561 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5561/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5561/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5561/events | https://github.com/NVIDIA/NeMo/pull/5561 | 1,481,350,860 | PR_kwDOC_bI7s5Em4Vh | 5,561 | Fix all gather while writing to a file during T5 finetuning | {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-12-07T08:42:27 | 2022-12-07T21:40:46 | 2022-12-07T15:39:23 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5561",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5561",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5561.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5561.patch",
"merged_at": "2022-12-07T15:39:23"
} | # What does this PR do ?
Fixes the all-gather logic to only gather across data parallel ranks to write to a file. Previously, examples would be duplicated in the file since all model parallel ranks have the same data.
**Collection**: NLP
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "Davood-M",
"id": 3964285,
"node_id": "MDQ6VXNlcjM5NjQyODU=",
"avatar_url": "https://avatars.githubusercontent.com/u/3964285?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Davood-M",
"html_url": "https://github.com/Davood-M",
"followers_url": "https://api.github.com/users/Davood-M/followers",
"following_url": "https://api.github.com/users/Davood-M/following{/other_user}",
"gists_url": "https://api.github.com/users/Davood-M/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Davood-M/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Davood-M/subscriptions",
"organizations_url": "https://api.github.com/users/Davood-M/orgs",
"repos_url": "https://api.github.com/users/Davood-M/repos",
"events_url": "https://api.github.com/users/Davood-M/events{/privacy}",
"received_events_url": "https://api.github.com/users/Davood-M/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5561/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5561/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5560 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5560/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5560/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5560/events | https://github.com/NVIDIA/NeMo/pull/5560 | 1,480,773,949 | PR_kwDOC_bI7s5Ek0rp | 5,560 | Optimizing distributed Adam when running with one work queue | {
"login": "timmoon10",
"id": 4406448,
"node_id": "MDQ6VXNlcjQ0MDY0NDg=",
"avatar_url": "https://avatars.githubusercontent.com/u/4406448?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/timmoon10",
"html_url": "https://github.com/timmoon10",
"followers_url": "https://api.github.com/users/timmoon10/followers",
"following_url": "https://api.github.com/users/timmoon10/following{/other_user}",
"gists_url": "https://api.github.com/users/timmoon10/gists{/gist_id}",
"starred_url": "https://api.github.com/users/timmoon10/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/timmoon10/subscriptions",
"organizations_url": "https://api.github.com/users/timmoon10/orgs",
"repos_url": "https://api.github.com/users/timmoon10/repos",
"events_url": "https://api.github.com/users/timmoon10/events{/privacy}",
"received_events_url": "https://api.github.com/users/timmoon10/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
},
{
"id": 4840216855,
"node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI",
"name": "CI",
"color": "1D76DB",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-07T02:53:16 | 2023-01-30T19:41:02 | 2023-01-30T19:06:58 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5560",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5560",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5560.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5560.patch",
"merged_at": "2023-01-30T19:06:58"
} | # What does this PR do ?
When running GPT with `CUDA_DEVICE_MAX_CONNECTIONS=1`, @erhoo82 observed poor overlapping between the backward compute and the gradient reduce-scatters in the distributed Adam optimizer. This PR integrates performance optimizations from https://github.com/NVIDIA/apex/pull/1551.
I haven't fully tested, but this also seems to fix a hang that occurs when running with 32+ Selene nodes.
**Collection**: NLP
# Changelog
- Manually configure dist Adam for GPT, T5, and BERT so each Transformer layer occupies one effective param bucket.
- Make sure dist Adam reduce-scatters are synchronized before all-reducing embedding gradients.
# Usage
Set the optimizer to `distributed_fused_adam` in the config file: https://github.com/NVIDIA/NeMo/blob/f1c8714f2fe0bc18a338fb02f0e608144981b291/examples/nlp/language_modeling/conf/megatron_gpt_config.yaml#L192
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [x] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [x] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
Pinging @erhoo82 and @markelsanz14.
# Additional Information
* Depends on https://github.com/NVIDIA/apex/pull/1551 | {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5560/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5560/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5559 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5559/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5559/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5559/events | https://github.com/NVIDIA/NeMo/pull/5559 | 1,480,563,487 | PR_kwDOC_bI7s5EkEfA | 5,559 | Remove broadcast from T5 prompt learning inference | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4556669155,
"node_id": "LA_kwDOC_bI7s8AAAABD5lA4w",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/cherry-pick",
"name": "cherry-pick",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | null | [] | null | null | 2022-12-07T00:37:57 | 2022-12-07T05:39:03 | 2022-12-07T05:39:00 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5559",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5559",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5559.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5559.patch",
"merged_at": "2022-12-07T05:39:00"
} | Signed-off-by: MaximumEntropy <[email protected]>
# What does this PR do ?
Removes a broadcast call that causes hanging during T5 prompt learning inference.
**Collection**: NLP
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5559/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5559/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5558 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5558/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5558/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5558/events | https://github.com/NVIDIA/NeMo/pull/5558 | 1,480,286,908 | PR_kwDOC_bI7s5EjFlM | 5,558 | Remove broadcast from T5 prompt learning inference | {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-12-06T22:00:18 | 2022-12-07T00:37:16 | 2022-12-07T00:37:12 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5558",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5558",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5558.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5558.patch",
"merged_at": "2022-12-07T00:37:12"
} | Signed-off-by: MaximumEntropy <[email protected]>
# What does this PR do ?
Removes a broadcast call that causes hanging during T5 prompt learning inference.
**Collection**: NLP
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5558/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5558/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5557 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5557/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5557/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5557/events | https://github.com/NVIDIA/NeMo/pull/5557 | 1,480,278,032 | PR_kwDOC_bI7s5EjDi0 | 5,557 | Add wer details - insertion, deletion, substitution rate | {
"login": "fayejf",
"id": 36722593,
"node_id": "MDQ6VXNlcjM2NzIyNTkz",
"avatar_url": "https://avatars.githubusercontent.com/u/36722593?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/fayejf",
"html_url": "https://github.com/fayejf",
"followers_url": "https://api.github.com/users/fayejf/followers",
"following_url": "https://api.github.com/users/fayejf/following{/other_user}",
"gists_url": "https://api.github.com/users/fayejf/gists{/gist_id}",
"starred_url": "https://api.github.com/users/fayejf/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/fayejf/subscriptions",
"organizations_url": "https://api.github.com/users/fayejf/orgs",
"repos_url": "https://api.github.com/users/fayejf/repos",
"events_url": "https://api.github.com/users/fayejf/events{/privacy}",
"received_events_url": "https://api.github.com/users/fayejf/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-06T21:55:01 | 2022-12-07T02:11:46 | 2022-12-07T02:11:43 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5557",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5557",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5557.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5557.patch",
"merged_at": "2022-12-07T02:11:43"
} | Signed-off-by: fayejf <[email protected]>
# What does this PR do ?
Add wer details - insertion, deletion, substitution rate
**Collection**: ASR
# Changelog
- Add insertion, deletion, and substitution rate for evaluating ASR result
- (Note the original wer has not been deprecated at this time, and might be deprecated once we confirm the new package is reliable and faster. )
# Usage
* You can potentially add a usage example below
```python
wer, words, ins_rate, del_rate, sub_rate = word_error_rate_detail(hypotheses=["GPU"], references=["GPU GPU"], use_cer=False)
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [x] Did you write any new necessary tests?
- [x] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5557/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5557/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5556 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5556/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5556/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5556/events | https://github.com/NVIDIA/NeMo/pull/5556 | 1,479,862,669 | PR_kwDOC_bI7s5EhkZU | 5,556 | Bert interleaved | {
"login": "shanmugamr1992",
"id": 111910568,
"node_id": "U_kgDOBqueqA",
"avatar_url": "https://avatars.githubusercontent.com/u/111910568?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/shanmugamr1992",
"html_url": "https://github.com/shanmugamr1992",
"followers_url": "https://api.github.com/users/shanmugamr1992/followers",
"following_url": "https://api.github.com/users/shanmugamr1992/following{/other_user}",
"gists_url": "https://api.github.com/users/shanmugamr1992/gists{/gist_id}",
"starred_url": "https://api.github.com/users/shanmugamr1992/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/shanmugamr1992/subscriptions",
"organizations_url": "https://api.github.com/users/shanmugamr1992/orgs",
"repos_url": "https://api.github.com/users/shanmugamr1992/repos",
"events_url": "https://api.github.com/users/shanmugamr1992/events{/privacy}",
"received_events_url": "https://api.github.com/users/shanmugamr1992/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-06T18:35:37 | 2022-12-16T19:45:59 | 2022-12-16T19:42:26 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5556",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5556",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5556.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5556.patch",
"merged_at": "2022-12-16T19:42:26"
} | # What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "shanmugamr1992",
"id": 111910568,
"node_id": "U_kgDOBqueqA",
"avatar_url": "https://avatars.githubusercontent.com/u/111910568?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/shanmugamr1992",
"html_url": "https://github.com/shanmugamr1992",
"followers_url": "https://api.github.com/users/shanmugamr1992/followers",
"following_url": "https://api.github.com/users/shanmugamr1992/following{/other_user}",
"gists_url": "https://api.github.com/users/shanmugamr1992/gists{/gist_id}",
"starred_url": "https://api.github.com/users/shanmugamr1992/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/shanmugamr1992/subscriptions",
"organizations_url": "https://api.github.com/users/shanmugamr1992/orgs",
"repos_url": "https://api.github.com/users/shanmugamr1992/repos",
"events_url": "https://api.github.com/users/shanmugamr1992/events{/privacy}",
"received_events_url": "https://api.github.com/users/shanmugamr1992/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5556/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5556/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5555 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5555/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5555/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5555/events | https://github.com/NVIDIA/NeMo/pull/5555 | 1,479,275,104 | PR_kwDOC_bI7s5EfdyR | 5,555 | Fix log-rank-0-only logic | {
"login": "mikolajblaz",
"id": 9979479,
"node_id": "MDQ6VXNlcjk5Nzk0Nzk=",
"avatar_url": "https://avatars.githubusercontent.com/u/9979479?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/mikolajblaz",
"html_url": "https://github.com/mikolajblaz",
"followers_url": "https://api.github.com/users/mikolajblaz/followers",
"following_url": "https://api.github.com/users/mikolajblaz/following{/other_user}",
"gists_url": "https://api.github.com/users/mikolajblaz/gists{/gist_id}",
"starred_url": "https://api.github.com/users/mikolajblaz/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/mikolajblaz/subscriptions",
"organizations_url": "https://api.github.com/users/mikolajblaz/orgs",
"repos_url": "https://api.github.com/users/mikolajblaz/repos",
"events_url": "https://api.github.com/users/mikolajblaz/events{/privacy}",
"received_events_url": "https://api.github.com/users/mikolajblaz/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-12-06T14:46:28 | 2022-12-07T10:24:41 | 2022-12-07T10:24:40 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5555",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5555",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5555.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5555.patch",
"merged_at": "2022-12-07T10:24:40"
} | # What does this PR do ?
Fix logic behind `trainer.log_global_rank_0_only` and `trainer.log_local_rank_0_only` flags (currently they do nothing).
**Collection**: all
# Changelog
- Fix `trainer.log_*_rank_0_only` flags logic
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [x] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "SeanNaren",
"id": 6707363,
"node_id": "MDQ6VXNlcjY3MDczNjM=",
"avatar_url": "https://avatars.githubusercontent.com/u/6707363?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SeanNaren",
"html_url": "https://github.com/SeanNaren",
"followers_url": "https://api.github.com/users/SeanNaren/followers",
"following_url": "https://api.github.com/users/SeanNaren/following{/other_user}",
"gists_url": "https://api.github.com/users/SeanNaren/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SeanNaren/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SeanNaren/subscriptions",
"organizations_url": "https://api.github.com/users/SeanNaren/orgs",
"repos_url": "https://api.github.com/users/SeanNaren/repos",
"events_url": "https://api.github.com/users/SeanNaren/events{/privacy}",
"received_events_url": "https://api.github.com/users/SeanNaren/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5555/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5555/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5553 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5553/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5553/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5553/events | https://github.com/NVIDIA/NeMo/pull/5553 | 1,478,438,025 | PR_kwDOC_bI7s5EcivM | 5,553 | Dual streaming/non-streaming Conformer with adaptive look-ahead | {
"login": "VahidooX",
"id": 23551647,
"node_id": "MDQ6VXNlcjIzNTUxNjQ3",
"avatar_url": "https://avatars.githubusercontent.com/u/23551647?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/VahidooX",
"html_url": "https://github.com/VahidooX",
"followers_url": "https://api.github.com/users/VahidooX/followers",
"following_url": "https://api.github.com/users/VahidooX/following{/other_user}",
"gists_url": "https://api.github.com/users/VahidooX/gists{/gist_id}",
"starred_url": "https://api.github.com/users/VahidooX/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/VahidooX/subscriptions",
"organizations_url": "https://api.github.com/users/VahidooX/orgs",
"repos_url": "https://api.github.com/users/VahidooX/repos",
"events_url": "https://api.github.com/users/VahidooX/events{/privacy}",
"received_events_url": "https://api.github.com/users/VahidooX/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
},
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-06T07:23:27 | 2023-02-15T02:02:00 | 2023-02-15T02:02:00 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | true | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5553",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5553",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5553.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5553.patch",
"merged_at": null
} | # What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5553/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5553/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5552 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5552/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5552/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5552/events | https://github.com/NVIDIA/NeMo/pull/5552 | 1,478,274,756 | PR_kwDOC_bI7s5Eb-dz | 5,552 | Fix ASR Buffered inference scripts | {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-06T05:46:15 | 2022-12-06T19:06:01 | 2022-12-06T19:04:51 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5552",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5552",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5552.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5552.patch",
"merged_at": "2022-12-06T19:04:51"
} | # What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
**Collection**: [ASR]
# Changelog
- Deepcopy the alignments before compute of alignments
- Do proper class level check of log level instead of incorrect config checks (older model doesn't have in config because the log=True is set by default inside of the function)
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [x] Bugfix
- [ ] Documentation
| {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5552/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5552/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5551 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5551/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5551/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5551/events | https://github.com/NVIDIA/NeMo/pull/5551 | 1,477,695,785 | PR_kwDOC_bI7s5EZ8Ge | 5,551 | Add kwargs to TE full activation recompute path | {
"login": "ksivaman",
"id": 36168853,
"node_id": "MDQ6VXNlcjM2MTY4ODUz",
"avatar_url": "https://avatars.githubusercontent.com/u/36168853?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ksivaman",
"html_url": "https://github.com/ksivaman",
"followers_url": "https://api.github.com/users/ksivaman/followers",
"following_url": "https://api.github.com/users/ksivaman/following{/other_user}",
"gists_url": "https://api.github.com/users/ksivaman/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ksivaman/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ksivaman/subscriptions",
"organizations_url": "https://api.github.com/users/ksivaman/orgs",
"repos_url": "https://api.github.com/users/ksivaman/repos",
"events_url": "https://api.github.com/users/ksivaman/events{/privacy}",
"received_events_url": "https://api.github.com/users/ksivaman/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-05T23:29:55 | 2023-02-15T23:18:34 | 2022-12-12T23:33:09 | MEMBER | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5551",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5551",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5551.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5551.patch",
"merged_at": null
} | Signed-off-by: Kirthi Shankar Sivamani <[email protected]>
# What does this PR do ?
`te_checkpoint` now has support for kwargs. This is necessary for the `is_first_microbatch` keyword argument during forward pass without which the numerics of the forward pass in the normal vs recompute phase differ. | {
"login": "ksivaman",
"id": 36168853,
"node_id": "MDQ6VXNlcjM2MTY4ODUz",
"avatar_url": "https://avatars.githubusercontent.com/u/36168853?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ksivaman",
"html_url": "https://github.com/ksivaman",
"followers_url": "https://api.github.com/users/ksivaman/followers",
"following_url": "https://api.github.com/users/ksivaman/following{/other_user}",
"gists_url": "https://api.github.com/users/ksivaman/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ksivaman/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ksivaman/subscriptions",
"organizations_url": "https://api.github.com/users/ksivaman/orgs",
"repos_url": "https://api.github.com/users/ksivaman/repos",
"events_url": "https://api.github.com/users/ksivaman/events{/privacy}",
"received_events_url": "https://api.github.com/users/ksivaman/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5551/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5551/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5550 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5550/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5550/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5550/events | https://github.com/NVIDIA/NeMo/pull/5550 | 1,477,693,320 | PR_kwDOC_bI7s5EZ7jv | 5,550 | Upgrade to 22.11 | {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
},
{
"id": 4840216855,
"node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI",
"name": "CI",
"color": "1D76DB",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-05T23:28:49 | 2022-12-06T23:41:08 | 2022-12-06T23:41:05 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5550",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5550",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5550.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5550.patch",
"merged_at": "2022-12-06T23:41:05"
} | # What does this PR do ?
Upgrade to the latest NGC PyTorch 22.11 container
**Collection**: [Core]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5550/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5550/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5549 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5549/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5549/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5549/events | https://github.com/NVIDIA/NeMo/pull/5549 | 1,477,679,132 | PR_kwDOC_bI7s5EZ4Qr | 5,549 | Switch order of args in optimizer_step override | {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-05T23:22:33 | 2022-12-06T03:52:18 | 2022-12-06T03:52:13 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5549",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5549",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5549.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5549.patch",
"merged_at": "2022-12-06T03:52:13"
} | # What does this PR do ?
PTL 1.8 swapped the order of the positional args for the optimizer_step method that we are overriding.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5549/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5549/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5548 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5548/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5548/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5548/events | https://github.com/NVIDIA/NeMo/pull/5548 | 1,477,259,889 | PR_kwDOC_bI7s5EYZLB | 5,548 | Remove notebook | {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-12-05T19:21:25 | 2022-12-05T19:25:17 | 2022-12-05T19:25:11 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5548",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5548",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5548.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5548.patch",
"merged_at": "2022-12-05T19:25:11"
} | # What does this PR do ?
Remove notebook which is no longer supported.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5548/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5548/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5547 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5547/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5547/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5547/events | https://github.com/NVIDIA/NeMo/pull/5547 | 1,476,307,306 | PR_kwDOC_bI7s5EVDdU | 5,547 | Data parallel collect results | {
"login": "michalivne",
"id": 7453913,
"node_id": "MDQ6VXNlcjc0NTM5MTM=",
"avatar_url": "https://avatars.githubusercontent.com/u/7453913?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/michalivne",
"html_url": "https://github.com/michalivne",
"followers_url": "https://api.github.com/users/michalivne/followers",
"following_url": "https://api.github.com/users/michalivne/following{/other_user}",
"gists_url": "https://api.github.com/users/michalivne/gists{/gist_id}",
"starred_url": "https://api.github.com/users/michalivne/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/michalivne/subscriptions",
"organizations_url": "https://api.github.com/users/michalivne/orgs",
"repos_url": "https://api.github.com/users/michalivne/repos",
"events_url": "https://api.github.com/users/michalivne/events{/privacy}",
"received_events_url": "https://api.github.com/users/michalivne/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-12-05T11:13:13 | 2022-12-05T20:37:00 | 2022-12-05T20:35:31 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5547",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5547",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5547.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5547.patch",
"merged_at": "2022-12-05T20:35:30"
} | # What does this PR do ?
Added gather_objects to easily collect results generated using data parallel.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "michalivne",
"id": 7453913,
"node_id": "MDQ6VXNlcjc0NTM5MTM=",
"avatar_url": "https://avatars.githubusercontent.com/u/7453913?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/michalivne",
"html_url": "https://github.com/michalivne",
"followers_url": "https://api.github.com/users/michalivne/followers",
"following_url": "https://api.github.com/users/michalivne/following{/other_user}",
"gists_url": "https://api.github.com/users/michalivne/gists{/gist_id}",
"starred_url": "https://api.github.com/users/michalivne/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/michalivne/subscriptions",
"organizations_url": "https://api.github.com/users/michalivne/orgs",
"repos_url": "https://api.github.com/users/michalivne/repos",
"events_url": "https://api.github.com/users/michalivne/events{/privacy}",
"received_events_url": "https://api.github.com/users/michalivne/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5547/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5547/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5546 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5546/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5546/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5546/events | https://github.com/NVIDIA/NeMo/issues/5546 | 1,476,004,916 | I_kwDOC_bI7s5X-gw0 | 5,546 | [SD + ASR] Can't use local ASR model produce word timestamp. | {
"login": "triumph9989",
"id": 48488138,
"node_id": "MDQ6VXNlcjQ4NDg4MTM4",
"avatar_url": "https://avatars.githubusercontent.com/u/48488138?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/triumph9989",
"html_url": "https://github.com/triumph9989",
"followers_url": "https://api.github.com/users/triumph9989/followers",
"following_url": "https://api.github.com/users/triumph9989/following{/other_user}",
"gists_url": "https://api.github.com/users/triumph9989/gists{/gist_id}",
"starred_url": "https://api.github.com/users/triumph9989/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/triumph9989/subscriptions",
"organizations_url": "https://api.github.com/users/triumph9989/orgs",
"repos_url": "https://api.github.com/users/triumph9989/repos",
"events_url": "https://api.github.com/users/triumph9989/events{/privacy}",
"received_events_url": "https://api.github.com/users/triumph9989/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
},
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
},
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
},
{
"id": 4811888028,
"node_id": "LA_kwDOC_bI7s8AAAABHs-VnA",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/Speaker%20Tasks",
"name": "Speaker Tasks",
"color": "871AD4",
"default": false,
"description": ""
}
] | closed | false | {
"login": "tango4j",
"id": 13211483,
"node_id": "MDQ6VXNlcjEzMjExNDgz",
"avatar_url": "https://avatars.githubusercontent.com/u/13211483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tango4j",
"html_url": "https://github.com/tango4j",
"followers_url": "https://api.github.com/users/tango4j/followers",
"following_url": "https://api.github.com/users/tango4j/following{/other_user}",
"gists_url": "https://api.github.com/users/tango4j/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tango4j/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tango4j/subscriptions",
"organizations_url": "https://api.github.com/users/tango4j/orgs",
"repos_url": "https://api.github.com/users/tango4j/repos",
"events_url": "https://api.github.com/users/tango4j/events{/privacy}",
"received_events_url": "https://api.github.com/users/tango4j/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "tango4j",
"id": 13211483,
"node_id": "MDQ6VXNlcjEzMjExNDgz",
"avatar_url": "https://avatars.githubusercontent.com/u/13211483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tango4j",
"html_url": "https://github.com/tango4j",
"followers_url": "https://api.github.com/users/tango4j/followers",
"following_url": "https://api.github.com/users/tango4j/following{/other_user}",
"gists_url": "https://api.github.com/users/tango4j/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tango4j/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tango4j/subscriptions",
"organizations_url": "https://api.github.com/users/tango4j/orgs",
"repos_url": "https://api.github.com/users/tango4j/repos",
"events_url": "https://api.github.com/users/tango4j/events{/privacy}",
"received_events_url": "https://api.github.com/users/tango4j/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-12-05T08:28:58 | 2023-01-10T02:03:52 | 2023-01-10T02:03:52 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | Hi,
I want to join my ASR model into `offline_diar_with_asr_infer.py` and use `asr_based_vad`.
But the following error is appearance.
```Error executing job with overrides: []
Traceback (most recent call last):
File "offline_diar_with_asr_infer.py", line 55, in main
word_hyp, word_ts_hyp = asr_decoder_ts.run_ASR(asr_model)
TypeError: 'NoneType' object is not callable
```
If `cfg.diarizer.asr.model_path=???` it will show another error.
```ValueError: `cfg` must have `tokenizer` config to create a tokenizer !```
I think that is about the inconsistency between my ASR class and this task, but I'm not very sure.
Is something that I have not noticed, please?
**Steps/Code to reproduce bug**
`offline_diar_with_asr_infer.py`
```
from omegaconf import OmegaConf
from nemo.collections.asr.parts.utils.decoder_timestamps_utils import ASRDecoderTimeStamps
from nemo.collections.asr.parts.utils.diarization_utils import OfflineDiarWithASR
from nemo.core.config import hydra_runner
from nemo.utils import logging
import nemo.collections.asr as nemo_asr
@hydra_runner(config_path="../conf/inference", config_name="diar_infer_meeting.yaml")
def main(cfg):
logging.info(f'Hydra config: {OmegaConf.to_yaml(cfg)}')
# ASR inference for words and word timestamps
asr_decoder_ts = ASRDecoderTimeStamps(cfg.diarizer)
asr_model = nemo_asr.models.EncDecCTCModel.restore_from('/home/face/NeMo/examples/asr/exp/Conformer-CTC-Char-Aishell-100ep-lr0.9/2022-12-01_09-43-05/checkpoints/Conformer-CTC-Char-Aishell-100ep-lr0.9.nemo')
#asr_model = asr_decoder_ts.set_asr_model()
word_hyp, word_ts_hyp = asr_decoder_ts.run_ASR(asr_model)
# Create a class instance for matching ASR and diarization results
asr_diar_offline = OfflineDiarWithASR(cfg.diarizer)
asr_diar_offline.word_ts_anchor_offset = asr_decoder_ts.word_ts_anchor_offset
# Diarization inference for speaker labels
diar_hyp, diar_score = asr_diar_offline.run_diarization(cfg, word_ts_hyp)
trans_info_dict = asr_diar_offline.get_transcript_with_speaker_labels(diar_hyp, word_hyp, word_ts_hyp)
# If RTTM is provided and DER evaluation
if diar_score is not None:
metric, mapping_dict, _ = diar_score
# Get session-level diarization error rate and speaker counting error
der_results = OfflineDiarWithASR.gather_eval_results(
diar_score=diar_score,
audio_rttm_map_dict=asr_diar_offline.AUDIO_RTTM_MAP,
trans_info_dict=trans_info_dict,
root_path=asr_diar_offline.root_path,
)
# Calculate WER and cpWER if reference CTM files exist
wer_results = OfflineDiarWithASR.evaluate(
hyp_trans_info_dict=trans_info_dict,
audio_file_list=asr_diar_offline.audio_file_list,
ref_ctm_file_list=asr_diar_offline.ctm_file_list,
)
# Print average DER, WER and cpWER
OfflineDiarWithASR.print_errors(der_results=der_results, wer_results=wer_results)
# Save detailed session-level evaluation results in `root_path`.
OfflineDiarWithASR.write_session_level_result_in_csv(
der_results=der_results,
wer_results=wer_results,
root_path=asr_diar_offline.root_path,
csv_columns=asr_diar_offline.csv_columns,
)
if __name__ == '__main__':
main()
```
`diar_infer_meeting.yaml`
```
name: &name "ClusterDiarizer"
num_workers: 4
sample_rate: 16000
batch_size: 64
diarizer:
manifest_filepath: /home/face/NeMo/examples/speaker_tasks/diarization/manifest_Ali_far.json
out_dir: output_Ali_far
oracle_vad: False # If True, uses RTTM files provided in the manifest file to get speech activity (VAD) timestamps
collar: 0.25 # Collar value for scoring
ignore_overlap: True # Consider or ignore overlap segments while scoring
vad:
model_path: vad_multilingual_marblenet # .nemo local model path or pretrained VAD model name
external_vad_manifest: null # This option is provided to use external vad and provide its speech activity labels for speaker embeddings extraction. Only one of model_path or external_vad_manifest should be set
parameters: # Tuned parameters for CH109 (using the 11 multi-speaker sessions as dev set)
window_length_in_sec: 0.63 # Window length in sec for VAD context input
shift_length_in_sec: 0.01 # Shift length in sec for generate frame level VAD prediction
smoothing: False # False or type of smoothing method (eg: median)
overlap: 0.5 # Overlap ratio for overlapped mean/median smoothing filter
onset: 0.9 # Onset threshold for detecting the beginning and end of a speech
offset: 0.5 # Offset threshold for detecting the end of a speech
pad_onset: 0 # Adding durations before each speech segment
pad_offset: 0 # Adding durations after each speech segment
min_duration_on: 0 # Threshold for small non_speech deletion
min_duration_off: 0.6 # Threshold for short speech segment deletion
filter_speech_first: True
speaker_embeddings:
model_path: /home/face/NeMo/examples/speaker_tasks/diarization/spk_model/TitaNet-Finetune-encoder-cn2-300-adjust.nemo # .nemo local model path or pretrained model name (titanet_large, ecapa_tdnn or speakerverification_speakernet)
parameters:
window_length_in_sec: [3.0,2.5,2.0,1.5,1.0,0.5] # Window length(s) in sec (floating-point number). either a number or a list. ex) 1.5 or [1.5,1.0,0.5]
shift_length_in_sec: [1.5,1.25,1.0,0.75,0.5,0.25] # Shift length(s) in sec (floating-point number). either a number or a list. ex) 0.75 or [0.75,0.5,0.25]
multiscale_weights: [1,1,1,1,1,1] # Weight for each scale. should be null (for single scale) or a list matched with window/shift scale count. ex) [0.33,0.33,0.33]
save_embeddings: True # If True, save speaker embeddings in pickle format. This should be True if clustering result is used for other models, such as `msdd_model`.
clustering:
parameters:
oracle_num_speakers: False # If True, use num of speakers value provided in manifest file.
max_num_speakers: 8 # Max number of speakers for each recording. If an oracle number of speakers is passed, this value is ignored.
enhanced_count_thres: 80 # If the number of segments is lower than this number, enhanced speaker counting is activated.
max_rp_threshold: 0.25 # Determines the range of p-value search: 0 < p <= max_rp_threshold.
sparse_search_volume: 30 # The higher the number, the more values will be examined with more time.
maj_vote_spk_count: False # If True, take a majority vote on multiple p-values to estimate the number of speakers.
msdd_model:
model_path: ??? # .nemo local model path or pretrained model name for multiscale diarization decoder (MSDD)
parameters:
use_speaker_model_from_ckpt: True # If True, use speaker embedding model in checkpoint. If False, the provided speaker embedding model in config will be used.
infer_batch_size: 25 # Batch size for MSDD inference.
sigmoid_threshold: [0.7] # Sigmoid threshold for generating binarized speaker labels. The smaller the more generous on detecting overlaps.
seq_eval_mode: False # If True, use oracle number of speaker and evaluate F1 score for the given speaker sequences. Default is False.
split_infer: True # If True, break the input audio clip to short sequences and calculate cluster average embeddings for inference.
diar_window_length: 50 # The length of split short sequence when split_infer is True.
overlap_infer_spk_limit: 5 # If the estimated number of speakers are larger than this number, overlap speech is not estimated.
asr:
model_path: stt_en_conformer_ctc_large # Provide NGC cloud ASR model name. stt_en_conformer_ctc_* models are recommended for diarization purposes.
parameters:
asr_based_vad: True # if True, speech segmentation for diarization is based on word-timestamps from ASR inference.
asr_based_vad_threshold: 100 # Threshold (in sec) that caps the gap between two words when generating VAD timestamps using ASR based VAD.
asr_batch_size: null # Batch size can be dependent on each ASR model. Default batch sizes are applied if set to null.
decoder_delay_in_sec: null # Native decoder delay. null is recommended to use the default values for each ASR model.
word_ts_anchor_offset: null # Offset to set a reference point from the start of the word. Recommended range of values is [-0.05 0.2].
word_ts_anchor_pos: "start" # Select which part of the word timestamp we want to use. The options are: 'start', 'end', 'mid'.
fix_word_ts_with_VAD: False # Fix the word timestamp using VAD output. You must provide a VAD model to use this feature.
colored_text: False # If True, use colored text to distinguish speakers in the output transcript.
print_time: True # If True, the start and end time of each speaker turn is printed in the output transcript.
break_lines: False # If True, the output transcript breaks the line to fix the line width (default is 90 chars)
ctc_decoder_parameters: # Optional beam search decoder (pyctcdecode)
pretrained_language_model: null # KenLM model file: .arpa model file or .bin binary file.
beam_width: 32
alpha: 0.5
beta: 2.5
realigning_lm_parameters: # Experimental feature
arpa_language_model: null # Provide a KenLM language model in .arpa format.
min_number_of_words: 3 # Min number of words for the left context.
max_number_of_words: 10 # Max number of words for the right context.
logprob_diff_threshold: 1.2 # The threshold for the difference between two log probability values from two hypotheses.
```
**Environment overview**
- Environment location: Bare-metal
- Method of NeMo install: From source
**Environment details**
- NeMo toolkit: 1.13.0rc0
- OS version: Ubuntu 20.04.4 LTS
- PyTorch version: pytorch-cuda 11.6
- Python version: 3.8
**Additional context**
GPU: GeForce RTX 2080 Ti
| {
"login": "tango4j",
"id": 13211483,
"node_id": "MDQ6VXNlcjEzMjExNDgz",
"avatar_url": "https://avatars.githubusercontent.com/u/13211483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tango4j",
"html_url": "https://github.com/tango4j",
"followers_url": "https://api.github.com/users/tango4j/followers",
"following_url": "https://api.github.com/users/tango4j/following{/other_user}",
"gists_url": "https://api.github.com/users/tango4j/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tango4j/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tango4j/subscriptions",
"organizations_url": "https://api.github.com/users/tango4j/orgs",
"repos_url": "https://api.github.com/users/tango4j/repos",
"events_url": "https://api.github.com/users/tango4j/events{/privacy}",
"received_events_url": "https://api.github.com/users/tango4j/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5546/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5546/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5545 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5545/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5545/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5545/events | https://github.com/NVIDIA/NeMo/pull/5545 | 1,475,699,209 | PR_kwDOC_bI7s5ES9Ak | 5,545 | added set_start_method + function param bugfix | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4556669155,
"node_id": "LA_kwDOC_bI7s8AAAABD5lA4w",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/cherry-pick",
"name": "cherry-pick",
"color": "ededed",
"default": false,
"description": null
},
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-05T05:14:37 | 2022-12-07T21:24:37 | 2022-12-07T21:24:34 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5545",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5545",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5545.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5545.patch",
"merged_at": "2022-12-07T21:24:34"
} | Signed-off-by: David Mosallanezhad <[email protected]>
# What does this PR do ?
- adds set_start_method for t5, gpt prompt learning and t5 fine-tuning.
- includes a bugfix for function parameters
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [x] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5545/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5545/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5544 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5544/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5544/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5544/events | https://github.com/NVIDIA/NeMo/pull/5544 | 1,473,577,148 | PR_kwDOC_bI7s5ELgry | 5,544 | [BugFix] Adding task variable for torchmetrics Accuracy | {
"login": "tango4j",
"id": 13211483,
"node_id": "MDQ6VXNlcjEzMjExNDgz",
"avatar_url": "https://avatars.githubusercontent.com/u/13211483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tango4j",
"html_url": "https://github.com/tango4j",
"followers_url": "https://api.github.com/users/tango4j/followers",
"following_url": "https://api.github.com/users/tango4j/following{/other_user}",
"gists_url": "https://api.github.com/users/tango4j/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tango4j/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tango4j/subscriptions",
"organizations_url": "https://api.github.com/users/tango4j/orgs",
"repos_url": "https://api.github.com/users/tango4j/repos",
"events_url": "https://api.github.com/users/tango4j/events{/privacy}",
"received_events_url": "https://api.github.com/users/tango4j/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-12-02T23:33:53 | 2023-01-18T21:45:10 | 2022-12-05T21:12:35 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | true | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5544",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5544",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5544.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5544.patch",
"merged_at": null
} | Signed-off-by: Taejin Park <[email protected]>
# What does this PR do ?
In the latest `torchmetrics==0.11.0`, Accuracy class' `__init__()` requires `task` variable to be specified.
Adding task variable for torchmetrics Accuracy
**Collection**: [Note which collection this PR will affect]
ASR
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
This is bugfix for label_models.py
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [x] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [x] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anybody in ASR
# Additional Information
* Related to # (issue)
| {
"login": "tango4j",
"id": 13211483,
"node_id": "MDQ6VXNlcjEzMjExNDgz",
"avatar_url": "https://avatars.githubusercontent.com/u/13211483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tango4j",
"html_url": "https://github.com/tango4j",
"followers_url": "https://api.github.com/users/tango4j/followers",
"following_url": "https://api.github.com/users/tango4j/following{/other_user}",
"gists_url": "https://api.github.com/users/tango4j/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tango4j/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tango4j/subscriptions",
"organizations_url": "https://api.github.com/users/tango4j/orgs",
"repos_url": "https://api.github.com/users/tango4j/repos",
"events_url": "https://api.github.com/users/tango4j/events{/privacy}",
"received_events_url": "https://api.github.com/users/tango4j/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5544/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5544/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5543 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5543/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5543/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5543/events | https://github.com/NVIDIA/NeMo/pull/5543 | 1,473,423,399 | PR_kwDOC_bI7s5EK_ve | 5,543 | Support non-standard padding token id | {
"login": "Numeri",
"id": 6404504,
"node_id": "MDQ6VXNlcjY0MDQ1MDQ=",
"avatar_url": "https://avatars.githubusercontent.com/u/6404504?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Numeri",
"html_url": "https://github.com/Numeri",
"followers_url": "https://api.github.com/users/Numeri/followers",
"following_url": "https://api.github.com/users/Numeri/following{/other_user}",
"gists_url": "https://api.github.com/users/Numeri/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Numeri/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Numeri/subscriptions",
"organizations_url": "https://api.github.com/users/Numeri/orgs",
"repos_url": "https://api.github.com/users/Numeri/repos",
"events_url": "https://api.github.com/users/Numeri/events{/privacy}",
"received_events_url": "https://api.github.com/users/Numeri/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-02T20:44:33 | 2023-01-06T04:47:59 | 2023-01-06T04:47:59 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5543",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5543",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5543.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5543.patch",
"merged_at": "2023-01-06T04:47:59"
} | # What does this PR do ?
Read the id of the padding token from the tokenizer when creating the embedding, rather than always defaulting to 0. This allows use of (admittedly bizarre) non-standard tokenizer models that don't give the padding token <PAD> the id 0.
Opening this as a draft PR because I'd like to get feedback on how the `TabularTokenizer` plays into this, as it does not provide a `TabularTokenizer.pad_id` property. In this case, I just default to 0, which is identical to the previous implementation's behavior. It could also be that `TabularTokenizer` is not meant to be used with this class, in which case this doesn't matter. I also need to test this out one last time before converting it to a normal PR.
**Collection**: NLP
# Changelog
- Remove hard-coded `token_idx = 0` when creating an `nn.Embedding` in the class `nemo.modules.common.transformer.transformer_modules.TransformerEmbedding`
- Get the padding ids for the `encoder_tokenizer` and `decoder_tokenizer` in `MTEncDecModel.__init__` and pass these down to all necessary submodules, until they get to the `TransformerEmbedding` objects
# Usage
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [x] Did you write any new necessary tests?
- [x] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [x] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
| {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5543/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5543/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5542 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5542/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5542/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5542/events | https://github.com/NVIDIA/NeMo/issues/5542 | 1,472,789,456 | I_kwDOC_bI7s5XyPvQ | 5,542 | Unable to process dataset for Neo-megatron-GPT-20 Billion model | {
"login": "mayank-nference",
"id": 86949680,
"node_id": "MDQ6VXNlcjg2OTQ5Njgw",
"avatar_url": "https://avatars.githubusercontent.com/u/86949680?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/mayank-nference",
"html_url": "https://github.com/mayank-nference",
"followers_url": "https://api.github.com/users/mayank-nference/followers",
"following_url": "https://api.github.com/users/mayank-nference/following{/other_user}",
"gists_url": "https://api.github.com/users/mayank-nference/gists{/gist_id}",
"starred_url": "https://api.github.com/users/mayank-nference/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/mayank-nference/subscriptions",
"organizations_url": "https://api.github.com/users/mayank-nference/orgs",
"repos_url": "https://api.github.com/users/mayank-nference/repos",
"events_url": "https://api.github.com/users/mayank-nference/events{/privacy}",
"received_events_url": "https://api.github.com/users/mayank-nference/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
},
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-12-02T12:28:14 | 2023-01-19T02:00:26 | 2023-01-19T02:00:26 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
```
home/mayanksharma/nemo_gpt_fintuning.py:261 in <module> │
│ │
│ 258 │
│ 259 │
│ 260 if __name__ == '__main__': │
│ ❱ 261 │ main() │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/nemo/core/config/hydra_runner.py:104 in │
│ wrapper │
│ │
│ 101 │ │ │ │ │
│ 102 │ │ │ │ # no return value from run_hydra() as it may sometime actually run the t │
│ 103 │ │ │ │ # multiple times (--multirun) │
│ ❱ 104 │ │ │ │ _run_hydra( │
│ 105 │ │ │ │ │ args_parser=_argparse_wrapper(args), │
│ 106 │ │ │ │ │ task_function=task_function, │
│ 107 │ │ │ │ │ config_path=config_path, │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/hydra/_internal/utils.py:377 in _run_hydra │
│ │
│ 374 │ │ if num_commands == 0: │
│ 375 │ │ │ args.run = True │
│ 376 │ │ if args.run: │
│ ❱ 377 │ │ │ run_and_report( │
│ 378 │ │ │ │ lambda: hydra.run( │
│ 379 │ │ │ │ │ config_name=config_name, │
│ 380 │ │ │ │ │ task_function=task_function, │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/hydra/_internal/utils.py:214 in │
│ run_and_report │
│ │
│ 211 │ │ return func() │
│ 212 │ except Exception as ex: │
│ 213 │ │ if _is_env_set("HYDRA_FULL_ERROR") or is_under_debugger(): │
│ ❱ 214 │ │ │ raise ex │
│ 215 │ │ else: │
│ 216 │ │ │ try: │
│ 217 │ │ │ │ if isinstance(ex, CompactHydraException): │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/hydra/_internal/utils.py:211 in │
│ run_and_report │
│ │
│ 208 │
│ 209 def run_and_report(func: Any) -> Any: │
│ 210 │ try: │
│ ❱ 211 │ │ return func() │
│ 212 │ except Exception as ex: │
│ 213 │ │ if _is_env_set("HYDRA_FULL_ERROR") or is_under_debugger(): │
│ 214 │ │ │ raise ex │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/hydra/_internal/utils.py:378 in <lambda> │
│ │
│ 375 │ │ │ args.run = True │
│ 376 │ │ if args.run: │
│ 377 │ │ │ run_and_report( │
│ ❱ 378 │ │ │ │ lambda: hydra.run( │
│ 379 │ │ │ │ │ config_name=config_name, │
│ 380 │ │ │ │ │ task_function=task_function, │
│ 381 │ │ │ │ │ overrides=args.overrides, │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/hydra/_internal/hydra.py:111 in run │
│ │
│ 108 │ │ callbacks.on_run_end(config=cfg, config_name=config_name, job_return=ret) │
│ 109 │ │ │
│ 110 │ │ # access the result to trigger an exception in case the job failed. │
│ ❱ 111 │ │ _ = ret.return_value │
│ 112 │ │ │
│ 113 │ │ return ret │
│ 114 │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/hydra/core/utils.py:233 in return_value │
│ │
│ 230 │ │ │ sys.stderr.write( │
│ 231 │ │ │ │ f"Error executing job with overrides: {self.overrides}" + os.linesep │
│ 232 │ │ │ ) │
│ ❱ 233 │ │ │ raise self._return_value │
│ 234 │ │
│ 235 │ @return_value.setter │
│ 236 │ def return_value(self, value: Any) -> None: │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/hydra/core/utils.py:160 in run_job │
│ │
│ 157 │ │ with env_override(hydra_cfg.hydra.job.env_set): │
│ 158 │ │ │ callbacks.on_job_start(config=config) │
│ 159 │ │ │ try: │
│ ❱ 160 │ │ │ │ ret.return_value = task_function(task_cfg) │
│ 161 │ │ │ │ ret.status = JobStatus.COMPLETED │
│ 162 │ │ │ except Exception as e: │
│ 163 │ │ │ │ ret.return_value = e │
│ │
│ /home/mayanksharma/nemo_gpt_fintuning.py:256 in main │
│ │
│ 253 │ │
│ 254 │ model = MegatronGPTModel(cfg.model, trainer) │
│ 255 │ │
│ ❱ 256 │ trainer.fit(model) │
│ 257 │ # trainer.fit(model, train_dataloaders=train_dataloader, val_dataloaders=valid_datal │
│ 258 │
│ 259 │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py:696 in │
│ fit │
│ │
│ 693 │ │ │ datamodule: An instance of :class:`~pytorch_lightning.core.datamodule.Lightn │
│ 694 │ │ """ │
│ 695 │ │ self.strategy.model = model │
│ ❱ 696 │ │ self._call_and_handle_interrupt( │
│ 697 │ │ │ self._fit_impl, model, train_dataloaders, val_dataloaders, datamodule, ckpt_ │
│ 698 │ │ ) │
│ 699 │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py:648 in │
│ _call_and_handle_interrupt │
│ │
│ 645 │ │ """ │
│ 646 │ │ try: │
│ 647 │ │ │ if self.strategy.launcher is not None: │
│ ❱ 648 │ │ │ │ return self.strategy.launcher.launch(trainer_fn, *args, trainer=self, ** │
│ 649 │ │ │ else: │
│ 650 │ │ │ │ return trainer_fn(*args, **kwargs) │
│ 651 │ │ # TODO(awaelchli): Unify both exceptions below, where `KeyboardError` doesn't re │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/pytorch_lightning/strategies/launchers/subpr │
│ ocess_script.py:93 in launch │
│ │
│ 90 │ │ """ │
│ 91 │ │ if not self.cluster_environment.creates_processes_externally: │
│ 92 │ │ │ self._call_children_scripts() │
│ ❱ 93 │ │ return function(*args, **kwargs) │
│ 94 │ │
│ 95 │ def _call_children_scripts(self) -> None: │
│ 96 │ │ # bookkeeping of spawned processes │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py:735 in │
│ _fit_impl │
│ │
│ 732 │ │ self._ckpt_path = self.__set_ckpt_path( │
│ 733 │ │ │ ckpt_path, model_provided=True, model_connected=self.lightning_module is not │
│ 734 │ │ ) │
│ ❱ 735 │ │ results = self._run(model, ckpt_path=self.ckpt_path) │
│ 736 │ │ │
│ 737 │ │ assert self.state.stopped │
│ 738 │ │ self.training = False │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py:1105 in │
│ _run │
│ │
│ 1102 │ │ self.strategy.setup_environment() │
│ 1103 │ │ self.__setup_profiler() │
│ 1104 │ │ │
│ ❱ 1105 │ │ self._call_setup_hook() # allow user to setup lightning_module in accelerator e │
│ 1106 │ │ │
│ 1107 │ │ # check if we should delay restoring checkpoint till later │
│ 1108 │ │ if not self.strategy.restore_checkpoint_after_setup: │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py:1449 in │
│ _call_setup_hook │
│ │
│ 1446 │ │ if self.datamodule is not None: │
│ 1447 │ │ │ self._call_lightning_datamodule_hook("setup", stage=fn) │
│ 1448 │ │ self._call_callback_hooks("setup", stage=fn) │
│ ❱ 1449 │ │ self._call_lightning_module_hook("setup", stage=fn) │
│ 1450 │ │ │
│ 1451 │ │ self.strategy.barrier("post_setup") │
│ 1452 │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py:1550 in │
│ _call_lightning_module_hook │
│ │
│ 1547 │ │ pl_module._current_fx_name = hook_name │
│ 1548 │ │ │
│ 1549 │ │ with self.profiler.profile(f"[LightningModule]{pl_module.__class__.__name__}.{ho │
│ ❱ 1550 │ │ │ output = fn(*args, **kwargs) │
│ 1551 │ │ │
│ 1552 │ │ # restore current_fx when nested context │
│ 1553 │ │ pl_module._current_fx_name = prev_fx_name │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/nemo/collections/nlp/models/language_modelin │
│ g/megatron_gpt_model.py:604 in setup │
│ │
│ 601 │ │ else: │
│ 602 │ │ │ # TODO: consider adding a ModelPT guard to check if model is being restored. │
│ 603 │ │ │ # allowing restored models to optionally setup datasets │
│ ❱ 604 │ │ │ self.build_train_valid_test_datasets() │
│ 605 │ │ │ self.setup_training_data(self.cfg.data) │
│ 606 │ │ │ self.setup_validation_data(self.cfg.data) │
│ 607 │ │ │ self.setup_test_data(self.cfg.data) │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/nemo/collections/nlp/models/language_modelin │
│ g/megatron_gpt_model.py:519 in build_train_valid_test_datasets │
│ │
│ 516 │ │ │ eval_iters * global_batch_size, │
│ 517 │ │ │ test_iters * global_batch_size, │
│ 518 │ │ ] │
│ ❱ 519 │ │ self._train_ds, self._validation_ds, self._test_ds = build_train_valid_test_data │
│ 520 │ │ │ cfg=self.cfg, │
│ 521 │ │ │ trainer=self.trainer, │
│ 522 │ │ │ data_prefix=self.cfg.data.data_prefix, │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/nemo/collections/nlp/data/language_modeling/ │
│ megatron/gpt_dataset.py:81 in build_train_valid_test_datasets │
│ │
│ 78 │ valid_datasets = [] │
│ 79 │ test_datasets = [] │
│ 80 │ for i in range(len(prefixes)): │
│ ❱ 81 │ │ train_ds, valid_ds, test_ds = _build_train_valid_test_datasets( │
│ 82 │ │ │ cfg, │
│ 83 │ │ │ trainer, │
│ 84 │ │ │ prefixes[i], │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/nemo/collections/nlp/data/language_modeling/ │
│ megatron/gpt_dataset.py:131 in _build_train_valid_test_datasets │
│ │
│ 128 │ """Build train, valid, and test datasets.""" │
│ 129 │ │
│ 130 │ # Indexed dataset. │
│ ❱ 131 │ indexed_dataset = get_indexed_dataset_(data_prefix, data_impl, skip_warmup) │
│ 132 │ │
│ 133 │ total_num_of_documents = indexed_dataset.sizes.shape[0] │
│ 134 │ splits = get_train_valid_test_split_(splits_string, total_num_of_documents) │
│ │
│ /opt/conda/envs/llm_env/lib/python3.8/site-packages/nemo/collections/nlp/data/language_modeling/ │
│ megatron/gpt_dataset.py:183 in get_indexed_dataset_ │
│ │
│ 180 │ indexed_dataset = make_indexed_dataset(data_prefix, data_impl, skip_warmup) │
│ 181 │ logging.info(f"indexed_dataset: {indexed_dataset.__dict__}") │
│ 182 │ logging.info(' > finished creating indexed dataset in {:4f} ' 'seconds'.format(time. │
│ ❱ 183 │ logging.info(' number of documents: {}'.format(indexed_dataset.sizes.shape[0])) │
│ 184 │ │
│ 185 │ return indexed_dataset │
│ 186 │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
AttributeError: 'CSVMemMapDataset' object has no attribute 'sizes'
```
**Steps/Code to reproduce bug**
I am using scrip borrowed from - [https://github.com/NVIDIA/NeMo/blob/v1.12.0/examples/nlp/language_modeling/megatron_gpt_pretraining.py]()
```
@hydra_runner(config_path="conf", config_name="megatron_gpt_config.yaml")
def main(cfg) -> None:
logging.info("\n\n************** Experiment configuration ***********")
logging.info(f'\n{OmegaConf.to_yaml(cfg)}')
cfg.model.data.data_prefix=[0.5,"/home/mayanksharma/nemo_data/small_train", 0.5, "/home/mayanksharma/nemo_data/small_valid", 0.5, "/home/mayanksharma/nemo_data/small_test"]
cfg.model.data.seq_length=128
cfg.model.data.data_impl="csv_mmap"
cfg.trainer.devices=1
cfg.model.sequence_parallel=True
cfg.model.global_batch_size=16
megatron_amp_o2 = cfg.model.get('megatron_amp_O2', False)
with_distributed_adam = cfg.model.optim.get('name') == 'distributed_fused_adam'
plugins = []
strategy = NLPDDPStrategy(
no_ddp_communication_hook=True, # we don't use DDP for async grad allreduce
gradient_as_bucket_view=cfg.model.gradient_as_bucket_view,
find_unused_parameters=False,
)
if cfg.trainer.precision in [16, 'bf16']:
scaler = None
if cfg.trainer.precision == 16:
scaler = GradScaler(
init_scale=cfg.model.get('native_amp_init_scale', 2 ** 32),
growth_interval=cfg.model.get('native_amp_growth_interval', 1000),
hysteresis=cfg.model.get('hysteresis', 2),
)
if megatron_amp_o2 and not with_distributed_adam:
plugins.append(MegatronHalfPrecisionPlugin(precision=cfg.trainer.precision, device='cuda', scaler=scaler))
else:
plugins.append(PipelineMixedPrecisionPlugin(precision=cfg.trainer.precision, device='cuda', scaler=scaler))
if cfg.get('cluster_type', None) == 'BCP':
plugins.append(TorchElasticEnvironment())
trainer = Trainer(plugins=plugins, strategy=strategy, **cfg.trainer)
exp_manager(trainer, cfg.exp_manager)
# update resume from checkpoint found by exp_manager
if cfg.model.resume_from_checkpoint is not None:
resume_from_checkpoint = cfg.model.resume_from_checkpoint
else:
resume_from_checkpoint = trainer._checkpoint_connector.resume_from_checkpoint_fit_path
logging.info(f'Resuming training from checkpoint: {resume_from_checkpoint}')
trainer._checkpoint_connector = CheckpointConnector(trainer, resume_from_checkpoint=resume_from_checkpoint)
# Override timer callback to a stateless one
for idx, callback in enumerate(trainer.callbacks):
if isinstance(callback, Timer):
trainer.callbacks[idx] = StatelessTimer(cfg.trainer.max_time,)
# hydra interpolation does not work here as the interpolation key is lost when PTL saves hparams
with open_dict(cfg):
cfg.model.precision = cfg.trainer.precision
model = MegatronGPTModel(cfg.model, trainer)
trainer.fit(model)
if __name__ == '__main__':
main()
```
**Config File That I am Using**
I have borrowed config from here [https://github.com/NVIDIA/NeMo/blob/v1.12.0/examples/nlp/language_modeling/conf/megatron_gpt_config.yaml]()
```
name: megatron_gpt_20B
restore_from_path: "nemo_model/nemo-megatron-gpt-20B/nemo_gpt20B_bf16_tp4.nemo" # used when starting from a .nemo file
trainer:
devices: 2
num_nodes: 1
accelerator: gpu
precision: 16
logger: False # logger provided by exp_manager
enable_checkpointing: False
replace_sampler_ddp: False
max_epochs: -1 # PTL default. In practice, max_steps will be reached first.
max_steps: 100000 # consumed_samples = global_step * micro_batch_size * data_parallel_size * accumulate_grad_batches
log_every_n_steps: 10
val_check_interval: 100
limit_val_batches: 50
limit_test_batches: 500
accumulate_grad_batches: 1 # do not modify, grad acc is automatic for training megatron models
gradient_clip_val: 1.0
benchmark: False
exp_manager:
explicit_log_dir: null
exp_dir: null
name: megatron_gpt
create_wandb_logger: False
wandb_logger_kwargs:
project: null
name: null
resume_if_exists: True
resume_ignore_no_checkpoint: True
create_checkpoint_callback: True
checkpoint_callback_params:
monitor: val_loss
save_top_k: 10
mode: min
always_save_nemo: False # saves nemo file during validation, not implemented for model parallel
save_nemo_on_train_end: False # not recommended when training large models on clusters with short time limits
filename: 'megatron_gpt--{val_loss:.2f}-{step}-{consumed_samples}'
model_parallel_size: ${multiply:${model.tensor_model_parallel_size}, ${model.pipeline_model_parallel_size}}
model:
# specify micro_batch_size, global_batch_size, and model parallelism
# gradient accumulation will be done automatically based on data_parallel_size
micro_batch_size: 4 # limited by GPU memory
global_batch_size: 8 # will use more micro batches to reach global batch size
tensor_model_parallel_size: 1 # intra-layer model parallelism
pipeline_model_parallel_size: 1 # inter-layer model parallelism
resume_from_checkpoint: null # manually set the checkpoint file to load from
# model architecture
encoder_seq_length: 512
max_position_embeddings: 1024 #${.encoder_seq_length} mayank changed
num_layers: 12
hidden_size: 768
ffn_hidden_size: 3072 # Transformer FFN hidden size. Usually 4 * hidden_size.
num_attention_heads: 12
init_method_std: 0.02 # Standard deviation of the zero mean normal distribution used for weight initialization.')
use_scaled_init_method: True # use scaled residuals initialization
hidden_dropout: 0.1 # Dropout probability for hidden state transformer.
kv_channels: null # Projection weights dimension in multi-head attention. Set to hidden_size // num_attention_heads if null
apply_query_key_layer_scaling: True # scale Q * K^T by 1 / layer-number.
normalization: layernorm # Type of normalization layers
layernorm_epsilon: 1e-5
do_layer_norm_weight_decay: False # True means weight decay on all params
make_vocab_size_divisible_by: 128 # Pad the vocab size to be divisible by this value for computation efficiency.
pre_process: True # add embedding
post_process: True # add pooler
persist_layer_norm: True # Use of persistent fused layer norm kernel.
bert_binary_head: 2 #used for classification
tokenizer:
library: 'huggingface'
type: 'gpt2'
model: null
vocab_file: null
merge_file: null
delimiter: null # only used for tabular tokenizer
sentencepiece_legacy: True # Legacy=True allows you to add special tokens to sentencepiece tokenizers.
# precision
native_amp_init_scale: 4294967296 # 2 ** 32
native_amp_growth_interval: 1000
hysteresis: 2 # Gradient scale hysteresis
fp32_residual_connection: False # Move residual connections to fp32
fp16_lm_cross_entropy: False # Move the cross entropy unreduced loss calculation for lm head to fp16
# Megatron O2-style half-precision
megatron_amp_O2: False # Enable O2-level automatic mixed precision using main parameters
grad_allreduce_chunk_size_mb: 125
grad_div_ar_fusion: True # Fuse grad division into torch.distributed.all_reduce
# miscellaneous
seed: 1234
use_cpu_initialization: False # Init weights on the CPU (slow for large models)
onnx_safe: False # Use work-arounds for known problems with Torch ONNX exporter.
apex_transformer_log_level: 30 # Python logging level displays logs with severity greater than or equal to this
gradient_as_bucket_view: True # PyTorch DDP argument. Allocate gradients in a contiguous bucket to save memory (less fragmentation and buffer memory)
gradient_accumulation_fusion: False # Fuse weight gradient accumulation to GEMMs. Only used with pipeline parallelism.
## Activation Checkpointing
# NeMo Megatron supports 'selective' activation checkpointing where only the memory intensive part of attention is checkpointed.
# These memory intensive activations are also less compute intensive which makes activation checkpointing more efficient for LLMs (20B+).
# See Reducing Activation Recomputation in Large Transformer Models: https://arxiv.org/abs/2205.05198 for more details.
# 'full' will checkpoint the entire transformer layer.
activations_checkpoint_granularity: null # 'selective' or 'full'
activations_checkpoint_method: null # 'uniform', 'block', not used with 'selective'
# 'uniform' divides the total number of transformer layers and checkpoints the input activation
# of each chunk at the specified granularity
# 'block' checkpoints the specified number of layers per pipeline stage at the specified granularity
activations_checkpoint_num_layers: null # not used with 'selective'
# when using 'uniform' this creates groups of transformer layers to checkpoint. Usually set to 1. Increase to save more memory.
# when using 'block' this this will checkpoint the first activations_checkpoint_num_layers per pipeline stage.
## Sequence Parallelism
# Makes tensor parallelism more memory efficient for LLMs (20B+) by parallelizing layer norms and dropout sequentially
# See Reducing Activation Recomputation in Large Transformer Models: https://arxiv.org/abs/2205.05198 for more details.
sequence_parallel: True
data:
# Path to data must be specified by the user.
# can override from the CLI: "model.data.data_prefix=[.5,/raid/data/pile/my-gpt3_00_text_document,.5,/raid/data/pile/my-gpt3_01_text_document]",
# Or see example below:
# data_prefix:
# - .5
# - /raid/data/pile/my-gpt3_00_text_document
# - .5
# - /raid/data/pile/my-gpt3_01_text_document
data_prefix: ???
index_mapping_dir: null # path to save index mapping .npy files, by default will save in the same location as data_prefix
data_impl: csv_map
splits_string: 900,50,50
seq_length: ${model.encoder_seq_length}
skip_warmup: True
num_workers: 2
dataloader_type: single # cyclic
reset_position_ids: False # Reset position ids after end-of-document token
reset_attention_mask: False # Reset attention mask after end-of-document token
eod_mask_loss: False # Mask loss for the end of document tokens
# masked_lm_prob: 0.15
# short_seq_prob: 0.1
# Nsys profiling options
nsys_profile:
enabled: False
start_step: 10 # Global batch to start profiling
end_step: 10 # Global batch to end profiling
ranks: [0] # Global rank IDs to profile
gen_shape: False # Generate model and kernel details including input shapes
optim:
name: fused_adam
lr: 2e-4
weight_decay: 0.01
betas:
- 0.9
- 0.98
sched:
name: CosineAnnealing
warmup_steps: 500
constant_steps: 50000
min_lr: 2e-5
```
**Inpu train,valid, test files I am Using**
Input small_train, small_valid, small_test files look like
```
0,|<startoftext>|sentence[LABEL]:Target|<endoftext>|
1,|<startoftext>|sentence[LABEL]:Target|<endoftext>|
2,|<startoftext>|sentence[LABEL]:Target|<endoftext>|
```
**Expected behaviour**
The train, valid and test files for gpt model should be processed correctly, as a similar file properly formatted for megatron-bert got correctly processed. for the megatron-bert model, the input train file looked something like this. It was a binary classification problem.
```
0,sentence,label1
1,sentence,label2
2,sentence,label1
```
**Environment overview (please complete the following information)**
- I am using GCP machine with 4 A100 Nvidia gpus `Linux nemo-training 4.19.0-22-cloud-amd64 #1 SMP Debian 4.19.260-1 (2022-09-29) x86_64 GNU/Linux`
- Environment location: `Cloud(specify cloud provider - GCP)`
- Method of NeMo install: [pip install or from source]. `pip install nemo-toolkit["all"]` and `nemo-version: 1.12.0`.
**Environment details**
If an NVIDIA docker image is used, you don't need to specify these.
Otherwise, please provide:
- OS version - `Debian 4.19.260-1 (2022-09-29) x86_64 GNU/Linux`
- PyTorch version - `1.13.0`
- PyTorch Lightning version - `1.7.6`
- Python version - `python3.8`
**Additional context**
*Please correct me if my input is formatted incorrectly for a gpt model. I could not find any suitable sample example for fine-tuning/pretraining the neo-megatron-20 billion models. Neither could I find any sample input files, or how they should be formatted to process correctly for GPT* | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5542/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5542/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5541 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5541/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5541/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5541/events | https://github.com/NVIDIA/NeMo/pull/5541 | 1,472,214,855 | PR_kwDOC_bI7s5EG5MZ | 5,541 | [TTS][ZH] Disambiguate polyphones with augmented dict and Jieba segmenter for Chinese FastPitch | {
"login": "yuekaizhang",
"id": 28823245,
"node_id": "MDQ6VXNlcjI4ODIzMjQ1",
"avatar_url": "https://avatars.githubusercontent.com/u/28823245?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/yuekaizhang",
"html_url": "https://github.com/yuekaizhang",
"followers_url": "https://api.github.com/users/yuekaizhang/followers",
"following_url": "https://api.github.com/users/yuekaizhang/following{/other_user}",
"gists_url": "https://api.github.com/users/yuekaizhang/gists{/gist_id}",
"starred_url": "https://api.github.com/users/yuekaizhang/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/yuekaizhang/subscriptions",
"organizations_url": "https://api.github.com/users/yuekaizhang/orgs",
"repos_url": "https://api.github.com/users/yuekaizhang/repos",
"events_url": "https://api.github.com/users/yuekaizhang/events{/privacy}",
"received_events_url": "https://api.github.com/users/yuekaizhang/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4418253970,
"node_id": "LA_kwDOC_bI7s8AAAABB1k0kg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/TTS",
"name": "TTS",
"color": "49AF28",
"default": false,
"description": ""
}
] | closed | false | {
"login": "XuesongYang",
"id": 1646669,
"node_id": "MDQ6VXNlcjE2NDY2Njk=",
"avatar_url": "https://avatars.githubusercontent.com/u/1646669?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/XuesongYang",
"html_url": "https://github.com/XuesongYang",
"followers_url": "https://api.github.com/users/XuesongYang/followers",
"following_url": "https://api.github.com/users/XuesongYang/following{/other_user}",
"gists_url": "https://api.github.com/users/XuesongYang/gists{/gist_id}",
"starred_url": "https://api.github.com/users/XuesongYang/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/XuesongYang/subscriptions",
"organizations_url": "https://api.github.com/users/XuesongYang/orgs",
"repos_url": "https://api.github.com/users/XuesongYang/repos",
"events_url": "https://api.github.com/users/XuesongYang/events{/privacy}",
"received_events_url": "https://api.github.com/users/XuesongYang/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "XuesongYang",
"id": 1646669,
"node_id": "MDQ6VXNlcjE2NDY2Njk=",
"avatar_url": "https://avatars.githubusercontent.com/u/1646669?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/XuesongYang",
"html_url": "https://github.com/XuesongYang",
"followers_url": "https://api.github.com/users/XuesongYang/followers",
"following_url": "https://api.github.com/users/XuesongYang/following{/other_user}",
"gists_url": "https://api.github.com/users/XuesongYang/gists{/gist_id}",
"starred_url": "https://api.github.com/users/XuesongYang/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/XuesongYang/subscriptions",
"organizations_url": "https://api.github.com/users/XuesongYang/orgs",
"repos_url": "https://api.github.com/users/XuesongYang/repos",
"events_url": "https://api.github.com/users/XuesongYang/events{/privacy}",
"received_events_url": "https://api.github.com/users/XuesongYang/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-12-02T02:21:12 | 2022-12-15T22:12:52 | 2022-12-15T22:12:52 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5541",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5541",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5541.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5541.patch",
"merged_at": "2022-12-15T22:12:52"
} | Signed-off-by: Yuekai Zhang <[email protected]>
To improve the G2P accuracy for Chinese polyphone, this PR first splits Chinese sentencs into words using jieba, then call pypinyin with cc_cedict.txt loaded.
Below is the test result for [g2pM test set](https://github.com/kakaobrain/g2pM/blob/master/data/test.sent)
@XuesongYang Would you mind having a look at this PR? Thx. | {
"login": "XuesongYang",
"id": 1646669,
"node_id": "MDQ6VXNlcjE2NDY2Njk=",
"avatar_url": "https://avatars.githubusercontent.com/u/1646669?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/XuesongYang",
"html_url": "https://github.com/XuesongYang",
"followers_url": "https://api.github.com/users/XuesongYang/followers",
"following_url": "https://api.github.com/users/XuesongYang/following{/other_user}",
"gists_url": "https://api.github.com/users/XuesongYang/gists{/gist_id}",
"starred_url": "https://api.github.com/users/XuesongYang/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/XuesongYang/subscriptions",
"organizations_url": "https://api.github.com/users/XuesongYang/orgs",
"repos_url": "https://api.github.com/users/XuesongYang/repos",
"events_url": "https://api.github.com/users/XuesongYang/events{/privacy}",
"received_events_url": "https://api.github.com/users/XuesongYang/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5541/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5541/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5540 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5540/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5540/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5540/events | https://github.com/NVIDIA/NeMo/issues/5540 | 1,472,127,801 | I_kwDOC_bI7s5XvuM5 | 5,540 | Can't reproduce machine translation bleu score | {
"login": "felicitywang1",
"id": 96318866,
"node_id": "U_kgDOBb21kg",
"avatar_url": "https://avatars.githubusercontent.com/u/96318866?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/felicitywang1",
"html_url": "https://github.com/felicitywang1",
"followers_url": "https://api.github.com/users/felicitywang1/followers",
"following_url": "https://api.github.com/users/felicitywang1/following{/other_user}",
"gists_url": "https://api.github.com/users/felicitywang1/gists{/gist_id}",
"starred_url": "https://api.github.com/users/felicitywang1/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/felicitywang1/subscriptions",
"organizations_url": "https://api.github.com/users/felicitywang1/orgs",
"repos_url": "https://api.github.com/users/felicitywang1/repos",
"events_url": "https://api.github.com/users/felicitywang1/events{/privacy}",
"received_events_url": "https://api.github.com/users/felicitywang1/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
}
] | closed | false | {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-12-02T00:10:24 | 2022-12-05T19:19:26 | 2022-12-05T19:19:26 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | Hi, I'm trying to evaluate the [NMT En Fr Transformer12x2 model ](https://catalog.ngc.nvidia.com/orgs/nvidia/teams/nemo/models/nmt_en_fr_transformer12x2) with sacrebleu, and cannot reproduce the bleu score reported in the doc - reported WMT13 and WMT14 bleu scores are 35.3/41.3 and I got 31.1/36.1. Is there anything I'm missing? I did get the warning from sacrebleu about
```
sacreBLEU: That's 100 lines that end in a tokenized period ('.')
sacreBLEU: It looks like you forgot to detokenize your test data, which may hurt your score.
sacreBLEU: If you insist your data is detokenized, or don't care, you can suppress this message with the `force` parameter.
```
Thank you. | {
"login": "felicitywang1",
"id": 96318866,
"node_id": "U_kgDOBb21kg",
"avatar_url": "https://avatars.githubusercontent.com/u/96318866?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/felicitywang1",
"html_url": "https://github.com/felicitywang1",
"followers_url": "https://api.github.com/users/felicitywang1/followers",
"following_url": "https://api.github.com/users/felicitywang1/following{/other_user}",
"gists_url": "https://api.github.com/users/felicitywang1/gists{/gist_id}",
"starred_url": "https://api.github.com/users/felicitywang1/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/felicitywang1/subscriptions",
"organizations_url": "https://api.github.com/users/felicitywang1/orgs",
"repos_url": "https://api.github.com/users/felicitywang1/repos",
"events_url": "https://api.github.com/users/felicitywang1/events{/privacy}",
"received_events_url": "https://api.github.com/users/felicitywang1/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5540/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5540/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5539 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5539/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5539/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5539/events | https://github.com/NVIDIA/NeMo/pull/5539 | 1,472,009,358 | PR_kwDOC_bI7s5EGNA4 | 5,539 | added set_start_method + function param bugfix | {
"login": "Davood-M",
"id": 3964285,
"node_id": "MDQ6VXNlcjM5NjQyODU=",
"avatar_url": "https://avatars.githubusercontent.com/u/3964285?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Davood-M",
"html_url": "https://github.com/Davood-M",
"followers_url": "https://api.github.com/users/Davood-M/followers",
"following_url": "https://api.github.com/users/Davood-M/following{/other_user}",
"gists_url": "https://api.github.com/users/Davood-M/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Davood-M/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Davood-M/subscriptions",
"organizations_url": "https://api.github.com/users/Davood-M/orgs",
"repos_url": "https://api.github.com/users/Davood-M/repos",
"events_url": "https://api.github.com/users/Davood-M/events{/privacy}",
"received_events_url": "https://api.github.com/users/Davood-M/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-12-01T21:59:10 | 2022-12-05T05:14:10 | 2022-12-05T05:14:06 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5539",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5539",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5539.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5539.patch",
"merged_at": "2022-12-05T05:14:06"
} | Signed-off-by: David Mosallanezhad <[email protected]>
# What does this PR do ?
- adds set_start_method for t5, gpt prompt learning and t5 fine-tuning.
- includes a bugfix for function parameters
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [x] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5539/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5539/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5538 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5538/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5538/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5538/events | https://github.com/NVIDIA/NeMo/issues/5538 | 1,472,002,677 | I_kwDOC_bI7s5XvPp1 | 5,538 | ASR_with_SpeakerDiarization Notebook Doesn't Work | {
"login": "superMDguy",
"id": 19691288,
"node_id": "MDQ6VXNlcjE5NjkxMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/19691288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/superMDguy",
"html_url": "https://github.com/superMDguy",
"followers_url": "https://api.github.com/users/superMDguy/followers",
"following_url": "https://api.github.com/users/superMDguy/following{/other_user}",
"gists_url": "https://api.github.com/users/superMDguy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/superMDguy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/superMDguy/subscriptions",
"organizations_url": "https://api.github.com/users/superMDguy/orgs",
"repos_url": "https://api.github.com/users/superMDguy/repos",
"events_url": "https://api.github.com/users/superMDguy/events{/privacy}",
"received_events_url": "https://api.github.com/users/superMDguy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
}
] | closed | false | null | [] | null | null | 2022-12-01T21:51:48 | 2022-12-02T15:43:10 | 2022-12-02T15:43:10 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
Error:
```
[NeMo E 2022-12-01 21:47:24 common:503] Model instantiation failed!
Target class: nemo.collections.asr.models.label_models.EncDecSpeakerLabelModel
Error(s): __new__() missing 1 required positional argument: 'task'
Traceback (most recent call last):
File "/usr/local/lib/python3.8/dist-packages/nemo/core/classes/common.py", line 482, in from_config_dict
instance = imported_cls(cfg=config, trainer=trainer)
File "/usr/local/lib/python3.8/dist-packages/nemo/collections/asr/models/label_models.py", line 168, in __init__
# self._macro_accuracy = Accuracy(num_classes=num_classes, average='macro')
TypeError: __new__() missing 1 required positional argument: 'task'
```
Looks like there's something wrong with how the Accuracy class is instantiated
**Steps/Code to reproduce bug**
1. Load [notebook](https://github.com/NVIDIA/NeMo/blob/stable/tutorials/speaker_tasks/ASR_with_SpeakerDiarization.ipynb#L766) in collab
2. Run all cells
3. Fails on the `Run diarization with the extracted word timestamps` step
**Expected behavior**
The notebook works
**Environment overview (please complete the following information)**
- Environment location: Collab w/GPU | {
"login": "superMDguy",
"id": 19691288,
"node_id": "MDQ6VXNlcjE5NjkxMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/19691288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/superMDguy",
"html_url": "https://github.com/superMDguy",
"followers_url": "https://api.github.com/users/superMDguy/followers",
"following_url": "https://api.github.com/users/superMDguy/following{/other_user}",
"gists_url": "https://api.github.com/users/superMDguy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/superMDguy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/superMDguy/subscriptions",
"organizations_url": "https://api.github.com/users/superMDguy/orgs",
"repos_url": "https://api.github.com/users/superMDguy/repos",
"events_url": "https://api.github.com/users/superMDguy/events{/privacy}",
"received_events_url": "https://api.github.com/users/superMDguy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5538/reactions",
"total_count": 2,
"+1": 2,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5538/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5537 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5537/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5537/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5537/events | https://github.com/NVIDIA/NeMo/pull/5537 | 1,471,703,144 | PR_kwDOC_bI7s5EFJ2j | 5,537 | Upperbound the torchmetrics version | {
"login": "SeanNaren",
"id": 6707363,
"node_id": "MDQ6VXNlcjY3MDczNjM=",
"avatar_url": "https://avatars.githubusercontent.com/u/6707363?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SeanNaren",
"html_url": "https://github.com/SeanNaren",
"followers_url": "https://api.github.com/users/SeanNaren/followers",
"following_url": "https://api.github.com/users/SeanNaren/following{/other_user}",
"gists_url": "https://api.github.com/users/SeanNaren/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SeanNaren/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SeanNaren/subscriptions",
"organizations_url": "https://api.github.com/users/SeanNaren/orgs",
"repos_url": "https://api.github.com/users/SeanNaren/repos",
"events_url": "https://api.github.com/users/SeanNaren/events{/privacy}",
"received_events_url": "https://api.github.com/users/SeanNaren/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-12-01T17:42:15 | 2022-12-01T21:12:47 | 2022-12-01T21:12:44 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5537",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5537",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5537.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5537.patch",
"merged_at": "2022-12-01T21:12:44"
} | # What does this PR do ?
Fixes current failures around torchmetrics new release.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Upper bound torchmetrics version.
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [x] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "SeanNaren",
"id": 6707363,
"node_id": "MDQ6VXNlcjY3MDczNjM=",
"avatar_url": "https://avatars.githubusercontent.com/u/6707363?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SeanNaren",
"html_url": "https://github.com/SeanNaren",
"followers_url": "https://api.github.com/users/SeanNaren/followers",
"following_url": "https://api.github.com/users/SeanNaren/following{/other_user}",
"gists_url": "https://api.github.com/users/SeanNaren/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SeanNaren/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SeanNaren/subscriptions",
"organizations_url": "https://api.github.com/users/SeanNaren/orgs",
"repos_url": "https://api.github.com/users/SeanNaren/repos",
"events_url": "https://api.github.com/users/SeanNaren/events{/privacy}",
"received_events_url": "https://api.github.com/users/SeanNaren/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5537/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5537/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5536 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5536/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5536/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5536/events | https://github.com/NVIDIA/NeMo/issues/5536 | 1,471,002,858 | I_kwDOC_bI7s5Xrbjq | 5,536 | Training stuck for a long time | {
"login": "Johnson-NLP",
"id": 65533749,
"node_id": "MDQ6VXNlcjY1NTMzNzQ5",
"avatar_url": "https://avatars.githubusercontent.com/u/65533749?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Johnson-NLP",
"html_url": "https://github.com/Johnson-NLP",
"followers_url": "https://api.github.com/users/Johnson-NLP/followers",
"following_url": "https://api.github.com/users/Johnson-NLP/following{/other_user}",
"gists_url": "https://api.github.com/users/Johnson-NLP/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Johnson-NLP/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Johnson-NLP/subscriptions",
"organizations_url": "https://api.github.com/users/Johnson-NLP/orgs",
"repos_url": "https://api.github.com/users/Johnson-NLP/repos",
"events_url": "https://api.github.com/users/Johnson-NLP/events{/privacy}",
"received_events_url": "https://api.github.com/users/Johnson-NLP/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-12-01T09:49:28 | 2024-07-12T05:58:18 | 2022-12-02T03:13:00 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | I try to train a Squeezeformer ASR model with speech_to_text_ctc_bpe.py. But the train log stucks for quite a long time after showing the "Total estimated model params size (MB)" ,without giving any error or training information. My GPUs keep 100% occupation as well. How to deal with this problem?
| {
"login": "Johnson-NLP",
"id": 65533749,
"node_id": "MDQ6VXNlcjY1NTMzNzQ5",
"avatar_url": "https://avatars.githubusercontent.com/u/65533749?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Johnson-NLP",
"html_url": "https://github.com/Johnson-NLP",
"followers_url": "https://api.github.com/users/Johnson-NLP/followers",
"following_url": "https://api.github.com/users/Johnson-NLP/following{/other_user}",
"gists_url": "https://api.github.com/users/Johnson-NLP/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Johnson-NLP/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Johnson-NLP/subscriptions",
"organizations_url": "https://api.github.com/users/Johnson-NLP/orgs",
"repos_url": "https://api.github.com/users/Johnson-NLP/repos",
"events_url": "https://api.github.com/users/Johnson-NLP/events{/privacy}",
"received_events_url": "https://api.github.com/users/Johnson-NLP/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5536/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5536/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5534 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5534/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5534/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5534/events | https://github.com/NVIDIA/NeMo/pull/5534 | 1,470,595,556 | PR_kwDOC_bI7s5EBYwZ | 5,534 | Add StreamingFeatureBufferer class for real-life streaming decoding | {
"login": "tango4j",
"id": 13211483,
"node_id": "MDQ6VXNlcjEzMjExNDgz",
"avatar_url": "https://avatars.githubusercontent.com/u/13211483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tango4j",
"html_url": "https://github.com/tango4j",
"followers_url": "https://api.github.com/users/tango4j/followers",
"following_url": "https://api.github.com/users/tango4j/following{/other_user}",
"gists_url": "https://api.github.com/users/tango4j/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tango4j/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tango4j/subscriptions",
"organizations_url": "https://api.github.com/users/tango4j/orgs",
"repos_url": "https://api.github.com/users/tango4j/repos",
"events_url": "https://api.github.com/users/tango4j/events{/privacy}",
"received_events_url": "https://api.github.com/users/tango4j/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-01T03:09:18 | 2022-12-02T22:16:33 | 2022-12-02T03:02:34 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5534",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5534",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5534.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5534.patch",
"merged_at": "2022-12-02T03:02:34"
} | Signed-off-by: Taejin Park <[email protected]>
# What does this PR do ?
Add StreamingFeatureBufferer class for real-life streaming decoding
**Collection**: [Note which collection this PR will affect]
ASR
# Changelog
Added StreamingFeatureBufferer() in streaming_utils.py.
# Usage
Will add usage when the example code is ready and pushed.
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in ASR
# Additional Information
* Related to # (issue)
| {
"login": "tango4j",
"id": 13211483,
"node_id": "MDQ6VXNlcjEzMjExNDgz",
"avatar_url": "https://avatars.githubusercontent.com/u/13211483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tango4j",
"html_url": "https://github.com/tango4j",
"followers_url": "https://api.github.com/users/tango4j/followers",
"following_url": "https://api.github.com/users/tango4j/following{/other_user}",
"gists_url": "https://api.github.com/users/tango4j/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tango4j/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tango4j/subscriptions",
"organizations_url": "https://api.github.com/users/tango4j/orgs",
"repos_url": "https://api.github.com/users/tango4j/repos",
"events_url": "https://api.github.com/users/tango4j/events{/privacy}",
"received_events_url": "https://api.github.com/users/tango4j/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5534/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5534/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5533 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5533/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5533/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5533/events | https://github.com/NVIDIA/NeMo/pull/5533 | 1,470,501,389 | PR_kwDOC_bI7s5EBEU0 | 5,533 | Draft: latent model support | {
"login": "ntadimeti",
"id": 56451898,
"node_id": "MDQ6VXNlcjU2NDUxODk4",
"avatar_url": "https://avatars.githubusercontent.com/u/56451898?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ntadimeti",
"html_url": "https://github.com/ntadimeti",
"followers_url": "https://api.github.com/users/ntadimeti/followers",
"following_url": "https://api.github.com/users/ntadimeti/following{/other_user}",
"gists_url": "https://api.github.com/users/ntadimeti/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ntadimeti/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ntadimeti/subscriptions",
"organizations_url": "https://api.github.com/users/ntadimeti/orgs",
"repos_url": "https://api.github.com/users/ntadimeti/repos",
"events_url": "https://api.github.com/users/ntadimeti/events{/privacy}",
"received_events_url": "https://api.github.com/users/ntadimeti/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | null | [] | null | null | 2022-12-01T01:23:19 | 2022-12-29T01:57:19 | 2022-12-29T01:57:19 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5533",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5533",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5533.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5533.patch",
"merged_at": null
} | # What does this PR do ?
Adds support to token_level_encoder_decoder module to accept Hidden and Loss transformations and apply them.
# Usage
* You can add the following to your config :
hidden_transforms: [MegatronGaussianHiddenTransform()] and it will apply a gaussian transform to the hiddens. | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5533/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5533/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5532 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5532/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5532/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5532/events | https://github.com/NVIDIA/NeMo/pull/5532 | 1,470,394,165 | PR_kwDOC_bI7s5EAr-5 | 5,532 | TTS tutorial update: use speaker 9017 instead of 6097 | {
"login": "redoctopus",
"id": 4268876,
"node_id": "MDQ6VXNlcjQyNjg4NzY=",
"avatar_url": "https://avatars.githubusercontent.com/u/4268876?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/redoctopus",
"html_url": "https://github.com/redoctopus",
"followers_url": "https://api.github.com/users/redoctopus/followers",
"following_url": "https://api.github.com/users/redoctopus/following{/other_user}",
"gists_url": "https://api.github.com/users/redoctopus/gists{/gist_id}",
"starred_url": "https://api.github.com/users/redoctopus/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/redoctopus/subscriptions",
"organizations_url": "https://api.github.com/users/redoctopus/orgs",
"repos_url": "https://api.github.com/users/redoctopus/repos",
"events_url": "https://api.github.com/users/redoctopus/events{/privacy}",
"received_events_url": "https://api.github.com/users/redoctopus/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4418253970,
"node_id": "LA_kwDOC_bI7s8AAAABB1k0kg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/TTS",
"name": "TTS",
"color": "49AF28",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-12-01T00:06:55 | 2022-12-01T23:02:34 | 2022-12-01T23:02:31 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5532",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5532",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5532.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5532.patch",
"merged_at": "2022-12-01T23:02:31"
} | Signed-off-by: Jocelyn Huang <[email protected]>
# What does this PR do ?
Replaces speaker 6097's data with speaker 9017's since the latter has given explicit permission.
**Collection**: TTS
# Changelog
- Updated link to data and recalculated pitch stats for the new speaker.
# Usage
* Tutorial should run as usual.
| {
"login": "redoctopus",
"id": 4268876,
"node_id": "MDQ6VXNlcjQyNjg4NzY=",
"avatar_url": "https://avatars.githubusercontent.com/u/4268876?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/redoctopus",
"html_url": "https://github.com/redoctopus",
"followers_url": "https://api.github.com/users/redoctopus/followers",
"following_url": "https://api.github.com/users/redoctopus/following{/other_user}",
"gists_url": "https://api.github.com/users/redoctopus/gists{/gist_id}",
"starred_url": "https://api.github.com/users/redoctopus/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/redoctopus/subscriptions",
"organizations_url": "https://api.github.com/users/redoctopus/orgs",
"repos_url": "https://api.github.com/users/redoctopus/repos",
"events_url": "https://api.github.com/users/redoctopus/events{/privacy}",
"received_events_url": "https://api.github.com/users/redoctopus/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5532/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5532/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5530 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5530/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5530/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5530/events | https://github.com/NVIDIA/NeMo/pull/5530 | 1,470,079,363 | PR_kwDOC_bI7s5D_ndd | 5,530 | Exporting Megatron GPT | {
"login": "Davood-M",
"id": 3964285,
"node_id": "MDQ6VXNlcjM5NjQyODU=",
"avatar_url": "https://avatars.githubusercontent.com/u/3964285?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Davood-M",
"html_url": "https://github.com/Davood-M",
"followers_url": "https://api.github.com/users/Davood-M/followers",
"following_url": "https://api.github.com/users/Davood-M/following{/other_user}",
"gists_url": "https://api.github.com/users/Davood-M/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Davood-M/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Davood-M/subscriptions",
"organizations_url": "https://api.github.com/users/Davood-M/orgs",
"repos_url": "https://api.github.com/users/Davood-M/repos",
"events_url": "https://api.github.com/users/Davood-M/events{/privacy}",
"received_events_url": "https://api.github.com/users/Davood-M/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
},
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-30T18:50:21 | 2023-03-31T16:20:21 | 2023-01-04T01:58:07 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5530",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5530",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5530.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5530.patch",
"merged_at": null
} | Signed-off-by: David Mosallanezhad <[email protected]>
# What does this PR do ?
Adds export functionality for Megatron GPT
**Collection**: [Note which collection this PR will affect]
# Changelog
- export related functions for Megatron GPT model
# Usage
* exportable using export.py with --autocast option
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5530/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5530/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5529 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5529/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5529/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5529/events | https://github.com/NVIDIA/NeMo/issues/5529 | 1,469,202,846 | I_kwDOC_bI7s5XkkGe | 5,529 | Expression detection in Speaker Diarization | {
"login": "viralrupapara36",
"id": 104073221,
"node_id": "U_kgDOBjQIBQ",
"avatar_url": "https://avatars.githubusercontent.com/u/104073221?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/viralrupapara36",
"html_url": "https://github.com/viralrupapara36",
"followers_url": "https://api.github.com/users/viralrupapara36/followers",
"following_url": "https://api.github.com/users/viralrupapara36/following{/other_user}",
"gists_url": "https://api.github.com/users/viralrupapara36/gists{/gist_id}",
"starred_url": "https://api.github.com/users/viralrupapara36/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/viralrupapara36/subscriptions",
"organizations_url": "https://api.github.com/users/viralrupapara36/orgs",
"repos_url": "https://api.github.com/users/viralrupapara36/repos",
"events_url": "https://api.github.com/users/viralrupapara36/events{/privacy}",
"received_events_url": "https://api.github.com/users/viralrupapara36/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1719393562,
"node_id": "MDU6TGFiZWwxNzE5MzkzNTYy",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/feature",
"name": "feature",
"color": "68F3F0",
"default": false,
"description": "request/PR for a new feature"
},
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | {
"login": "nithinraok",
"id": 19668129,
"node_id": "MDQ6VXNlcjE5NjY4MTI5",
"avatar_url": "https://avatars.githubusercontent.com/u/19668129?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/nithinraok",
"html_url": "https://github.com/nithinraok",
"followers_url": "https://api.github.com/users/nithinraok/followers",
"following_url": "https://api.github.com/users/nithinraok/following{/other_user}",
"gists_url": "https://api.github.com/users/nithinraok/gists{/gist_id}",
"starred_url": "https://api.github.com/users/nithinraok/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/nithinraok/subscriptions",
"organizations_url": "https://api.github.com/users/nithinraok/orgs",
"repos_url": "https://api.github.com/users/nithinraok/repos",
"events_url": "https://api.github.com/users/nithinraok/events{/privacy}",
"received_events_url": "https://api.github.com/users/nithinraok/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "nithinraok",
"id": 19668129,
"node_id": "MDQ6VXNlcjE5NjY4MTI5",
"avatar_url": "https://avatars.githubusercontent.com/u/19668129?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/nithinraok",
"html_url": "https://github.com/nithinraok",
"followers_url": "https://api.github.com/users/nithinraok/followers",
"following_url": "https://api.github.com/users/nithinraok/following{/other_user}",
"gists_url": "https://api.github.com/users/nithinraok/gists{/gist_id}",
"starred_url": "https://api.github.com/users/nithinraok/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/nithinraok/subscriptions",
"organizations_url": "https://api.github.com/users/nithinraok/orgs",
"repos_url": "https://api.github.com/users/nithinraok/repos",
"events_url": "https://api.github.com/users/nithinraok/events{/privacy}",
"received_events_url": "https://api.github.com/users/nithinraok/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
},
{
"login": "okuchaiev",
"id": 22035961,
"node_id": "MDQ6VXNlcjIyMDM1OTYx",
"avatar_url": "https://avatars.githubusercontent.com/u/22035961?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/okuchaiev",
"html_url": "https://github.com/okuchaiev",
"followers_url": "https://api.github.com/users/okuchaiev/followers",
"following_url": "https://api.github.com/users/okuchaiev/following{/other_user}",
"gists_url": "https://api.github.com/users/okuchaiev/gists{/gist_id}",
"starred_url": "https://api.github.com/users/okuchaiev/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/okuchaiev/subscriptions",
"organizations_url": "https://api.github.com/users/okuchaiev/orgs",
"repos_url": "https://api.github.com/users/okuchaiev/repos",
"events_url": "https://api.github.com/users/okuchaiev/events{/privacy}",
"received_events_url": "https://api.github.com/users/okuchaiev/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-11-30T07:57:11 | 2023-02-11T01:57:34 | 2023-02-11T01:57:34 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | While working with speaker diarization task I found that if input audio have expressions (like anger, shouting, whispering, exciting etc.) then model is recognise as another speaker.
So, if these expressions also recognise than it will be good. | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5529/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5529/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5528 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5528/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5528/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5528/events | https://github.com/NVIDIA/NeMo/pull/5528 | 1,469,172,483 | PR_kwDOC_bI7s5D8j_z | 5,528 | NLP docs fixes | {
"login": "vsl9",
"id": 4344862,
"node_id": "MDQ6VXNlcjQzNDQ4NjI=",
"avatar_url": "https://avatars.githubusercontent.com/u/4344862?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/vsl9",
"html_url": "https://github.com/vsl9",
"followers_url": "https://api.github.com/users/vsl9/followers",
"following_url": "https://api.github.com/users/vsl9/following{/other_user}",
"gists_url": "https://api.github.com/users/vsl9/gists{/gist_id}",
"starred_url": "https://api.github.com/users/vsl9/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/vsl9/subscriptions",
"organizations_url": "https://api.github.com/users/vsl9/orgs",
"repos_url": "https://api.github.com/users/vsl9/repos",
"events_url": "https://api.github.com/users/vsl9/events{/privacy}",
"received_events_url": "https://api.github.com/users/vsl9/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-30T07:27:36 | 2022-12-06T18:47:14 | 2022-12-05T22:14:13 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5528",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5528",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5528.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5528.patch",
"merged_at": "2022-12-05T22:14:12"
} | # What does this PR do ?
- Fixes invalid image URL in `Entity Linking` page
- Fixes title format in `Question Answering` page (to resolve issues in NLP Tasks TOC page)
**Collection**: NLP docs
# Changelog
- Resolves issues in NLP docs
# Before your PR is "Ready for review"
**Pre checks**:
- [X] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [X] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [X] Bugfix
- [X] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
| {
"login": "Jorjeous",
"id": 37293288,
"node_id": "MDQ6VXNlcjM3MjkzMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/37293288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Jorjeous",
"html_url": "https://github.com/Jorjeous",
"followers_url": "https://api.github.com/users/Jorjeous/followers",
"following_url": "https://api.github.com/users/Jorjeous/following{/other_user}",
"gists_url": "https://api.github.com/users/Jorjeous/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Jorjeous/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Jorjeous/subscriptions",
"organizations_url": "https://api.github.com/users/Jorjeous/orgs",
"repos_url": "https://api.github.com/users/Jorjeous/repos",
"events_url": "https://api.github.com/users/Jorjeous/events{/privacy}",
"received_events_url": "https://api.github.com/users/Jorjeous/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5528/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5528/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5527 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5527/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5527/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5527/events | https://github.com/NVIDIA/NeMo/pull/5527 | 1,469,008,597 | PR_kwDOC_bI7s5D8A9n | 5,527 | Multiblank Transducer | {
"login": "hainan-xv",
"id": 5440014,
"node_id": "MDQ6VXNlcjU0NDAwMTQ=",
"avatar_url": "https://avatars.githubusercontent.com/u/5440014?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/hainan-xv",
"html_url": "https://github.com/hainan-xv",
"followers_url": "https://api.github.com/users/hainan-xv/followers",
"following_url": "https://api.github.com/users/hainan-xv/following{/other_user}",
"gists_url": "https://api.github.com/users/hainan-xv/gists{/gist_id}",
"starred_url": "https://api.github.com/users/hainan-xv/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/hainan-xv/subscriptions",
"organizations_url": "https://api.github.com/users/hainan-xv/orgs",
"repos_url": "https://api.github.com/users/hainan-xv/repos",
"events_url": "https://api.github.com/users/hainan-xv/events{/privacy}",
"received_events_url": "https://api.github.com/users/hainan-xv/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-30T03:54:47 | 2022-12-21T18:43:33 | 2022-12-21T02:41:37 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5527",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5527",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5527.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5527.patch",
"merged_at": "2022-12-21T02:41:37"
} | # What does this PR do ?
Add support for multi-blank transducers described in https://arxiv.org/pdf/2211.03541
**Collection**:
ASR
# Changelog
- implementation of multi-blank both with numba kernel code as well as pytorch.
- add relevant test code.
- add example .yaml file for multi-blank ASR model configuration.
# Usage
Invoke transducer model training with examples/asr/conf/conformer/conformer_multiblank_transducer_bpe.yaml file. This will train a multi-blank model with big blank durations [2,4,8], and sigma=0.05.
To run multi-blank with custom configurations, e.g. using [2,3,5,8], change the options by,
```
model.loss.multiblank_rnnt_kwargs.big_blank_durations=[2,3,5,8] # used for loss computation in training.
model.loss.multiblank_rnnt_kwargs.sigma=0.05 # 0.05 is the recommended value.
model.decoding.big_blank_durations=[2,3,5,8] # used for decoding. The length of the list should be consistent with above.
model.joint.num_extra_outputs=4 # this needs to be consistent with the duration list you set
```
You could set model.decoding.big_blank_durations to have values smaller than what's used in training (model.loss.multiblank_rnnt_kwargs.big_blank_duration), e.g. [1,1,1,1]. This makes decoding slower but usually gives better accuracy.
To run inference, add option rnnt_decoding.big_blank_durations when calling NeMo/examples/asr/transcribe_speech.py, to specify the big blank durations, e.g.
```
python NeMo/examples/asr/transcribe_speech.py \
model_path=your_big_blank_model.nemo \
dataset_manifest=your_data.json \
++rnnt_decoding.big_blank_durations=[2,3,5,8] \
output_filename=your_output_file
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [x] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "hainan-xv",
"id": 5440014,
"node_id": "MDQ6VXNlcjU0NDAwMTQ=",
"avatar_url": "https://avatars.githubusercontent.com/u/5440014?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/hainan-xv",
"html_url": "https://github.com/hainan-xv",
"followers_url": "https://api.github.com/users/hainan-xv/followers",
"following_url": "https://api.github.com/users/hainan-xv/following{/other_user}",
"gists_url": "https://api.github.com/users/hainan-xv/gists{/gist_id}",
"starred_url": "https://api.github.com/users/hainan-xv/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/hainan-xv/subscriptions",
"organizations_url": "https://api.github.com/users/hainan-xv/orgs",
"repos_url": "https://api.github.com/users/hainan-xv/repos",
"events_url": "https://api.github.com/users/hainan-xv/events{/privacy}",
"received_events_url": "https://api.github.com/users/hainan-xv/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5527/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5527/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5526 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5526/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5526/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5526/events | https://github.com/NVIDIA/NeMo/pull/5526 | 1,468,828,281 | PR_kwDOC_bI7s5D7bCG | 5,526 | Add core classes and functions for online clustering diarizer part 1 | {
"login": "tango4j",
"id": 13211483,
"node_id": "MDQ6VXNlcjEzMjExNDgz",
"avatar_url": "https://avatars.githubusercontent.com/u/13211483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tango4j",
"html_url": "https://github.com/tango4j",
"followers_url": "https://api.github.com/users/tango4j/followers",
"following_url": "https://api.github.com/users/tango4j/following{/other_user}",
"gists_url": "https://api.github.com/users/tango4j/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tango4j/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tango4j/subscriptions",
"organizations_url": "https://api.github.com/users/tango4j/orgs",
"repos_url": "https://api.github.com/users/tango4j/repos",
"events_url": "https://api.github.com/users/tango4j/events{/privacy}",
"received_events_url": "https://api.github.com/users/tango4j/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-29T23:19:29 | 2023-02-22T23:48:57 | 2022-12-15T17:46:23 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5526",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5526",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5526.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5526.patch",
"merged_at": "2022-12-15T17:46:23"
} | Signed-off-by: Taejin Park <[email protected]>
[[NeMo Online Speaker Diarization VIDEO DEMO LINK](
https://drive.google.com/file/d/15ro8HKmzevXh_wx9oTtfRUw9wNtj6LuJ/view?usp=share_link)]
# What does this PR do ?
Add core classes and functions for online clustering diarizer.
Dividing the huge PR into two:
part 1: online clustering (nmesc_clustering.py) + unit tests
part 2: Segmentation + online diarizer class + unit tests
**Collection**:
ASR
# Changelog
#### `OnlineSpeakerClustering` in nmesc_clustering.py (included in this PR)
online version of `SpeakerClustering` class (online speaker counting and permutation matching)
#### `test_diar_utils.py` in NeMo/tests/collections/asr/ (included in this PR)
Tests all clustering and online diarization related functions
-------------
the following will be included in the later PR
#### `OnlineClusteringDiarizer` in online_diarizer.py (inherits ClusteringDiarizer) (will be included in part2)
online version of `ClusteringDiarizer` class in clustering_diarizer.py
#### `OnlineSegmentor` class in speaker_utils.py: (will be included in part2)
online version of speech segmentations
<img width="966" alt="image" src="https://user-images.githubusercontent.com/13211483/204661137-78d0272b-1594-4509-a486-a8720e835bc4.png">
Checkout [Help Slides Link](https://docs.google.com/presentation/d/1LDUTIGRNVuLJsy81V3GaHZlFDUpxZ9rux-fNPfgcLmw/edit?usp=sharing) for better understanding on why online diarizer needs
(1) linear sum assignment (Hungarian alg)
(2) Interval merging
(3) merge_vector function
and data flow of class and functions.
If you are curious about why online diarization needs FIFO-queue, take a look at the most recent paper on online neural diarization
[Online Neural Diarization of Unlimited Numbers of Speakers](https://arxiv.org/pdf/2206.02432.pdf)
# Usage
This is core classes and functions update.
Example script will be pushed in the second round of PR.
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [x] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
[Online Neural Diarization of Unlimited Numbers of Speakers](https://arxiv.org/pdf/2206.02432.pdf)
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "tango4j",
"id": 13211483,
"node_id": "MDQ6VXNlcjEzMjExNDgz",
"avatar_url": "https://avatars.githubusercontent.com/u/13211483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tango4j",
"html_url": "https://github.com/tango4j",
"followers_url": "https://api.github.com/users/tango4j/followers",
"following_url": "https://api.github.com/users/tango4j/following{/other_user}",
"gists_url": "https://api.github.com/users/tango4j/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tango4j/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tango4j/subscriptions",
"organizations_url": "https://api.github.com/users/tango4j/orgs",
"repos_url": "https://api.github.com/users/tango4j/repos",
"events_url": "https://api.github.com/users/tango4j/events{/privacy}",
"received_events_url": "https://api.github.com/users/tango4j/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5526/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5526/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5525 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5525/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5525/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5525/events | https://github.com/NVIDIA/NeMo/pull/5525 | 1,468,803,650 | PR_kwDOC_bI7s5D7Vpt | 5,525 | Conformer local attention | {
"login": "sam1373",
"id": 24465107,
"node_id": "MDQ6VXNlcjI0NDY1MTA3",
"avatar_url": "https://avatars.githubusercontent.com/u/24465107?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sam1373",
"html_url": "https://github.com/sam1373",
"followers_url": "https://api.github.com/users/sam1373/followers",
"following_url": "https://api.github.com/users/sam1373/following{/other_user}",
"gists_url": "https://api.github.com/users/sam1373/gists{/gist_id}",
"starred_url": "https://api.github.com/users/sam1373/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/sam1373/subscriptions",
"organizations_url": "https://api.github.com/users/sam1373/orgs",
"repos_url": "https://api.github.com/users/sam1373/repos",
"events_url": "https://api.github.com/users/sam1373/events{/privacy}",
"received_events_url": "https://api.github.com/users/sam1373/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-29T22:52:43 | 2023-03-03T14:28:26 | 2022-12-15T11:16:04 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5525",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5525",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5525.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5525.patch",
"merged_at": "2022-12-15T11:16:04"
} | Signed-off-by: Samuel Kriman <[email protected]>
# What does this PR do ?
This PR introduces the option of using local attention in Conformer with relative positional encoding, following the approach from Longformer (https://arxiv.org/abs/2004.05150), based on Ankur Kumar's branch https://github.com/iankur/NeMo/tree/longformer_rel_pos. This is useful for running on long audio files, as it can be done faster than with buffered inference. It may also improve performance on audio with long segments of noise that may confuse models with full attention.
This method can be used with checkpoints trained with regular attention. For example, Conformer-Large checkpoints (such as https://catalog.ngc.nvidia.com/orgs/nvidia/teams/nemo/models/stt_en_conformer_ctc_large and https://catalog.ngc.nvidia.com/orgs/nvidia/teams/nemo/models/stt_en_conformer_transducer_large) can be run on ~30 minute audio files on a single 32GB GPU after switching to local attention. The duration can be further increased to ~2 hours by switching to stacking subsampling (this won't work without additional training). Please note that the WER may increase slightly on our main evaluation sets when using small context local attention with models trained with full context (see table below for some WER(%) comparisons). This effect can be mitigated by either increasing the context size or fine-tuning the model using local attention.
| Model | LS test-other | LS test-clean | People Speech | SLR |
|----------|----------|----------|----------|----------|
| Conformer-Large-CTC | 4.53 | 2.19 | 23.04 | 5.94 |
| Conformer-Large-CTC with local attention (64, 64) | 5.39 | 2.57 | 23.89 | 6.2 |
| Conformer-Large-CTC with local attention (192, 192) | 4.53 | 2.19 | 23.2 | 5.96 |
| Conformer-Large-RNNT | 3.73 | 1.8 | 21.69 | 5.4 |
| Conformer-Large-RNNT with local attention (64, 64) | 4.6 | 2.14 | 21.85 | 5.67 |
| Conformer-Large-RNNT with local attention (192, 192) | 3.75 | 1.83 | 21.64 | 5.41 |
**Collection**: ASR
# Changelog
- adds option to use local attention for Conformer
- adds option for overriding attention parameters when running transcribe or eval scripts, or after loading model
# Usage
With an eval script can be used like this to switch to local attention with any Conformer checkpoint
```
python speech_to_text_eval.py \
(...other params...) \
++model_change.conformer.self_attention_model="rel_pos_local_attn" \
++model_change.conformer.att_context_size=[64, 64]
```
Or after loading model
```python
asr_model = ASRModel.from_pretrained('stt_en_conformer_ctc_large')
asr_model.change_attention_model(
self_attention_model="rel_pos_local_attn",
att_context_size=[64, 64]
)
```
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [x] Documentation
| {
"login": "sam1373",
"id": 24465107,
"node_id": "MDQ6VXNlcjI0NDY1MTA3",
"avatar_url": "https://avatars.githubusercontent.com/u/24465107?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sam1373",
"html_url": "https://github.com/sam1373",
"followers_url": "https://api.github.com/users/sam1373/followers",
"following_url": "https://api.github.com/users/sam1373/following{/other_user}",
"gists_url": "https://api.github.com/users/sam1373/gists{/gist_id}",
"starred_url": "https://api.github.com/users/sam1373/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/sam1373/subscriptions",
"organizations_url": "https://api.github.com/users/sam1373/orgs",
"repos_url": "https://api.github.com/users/sam1373/repos",
"events_url": "https://api.github.com/users/sam1373/events{/privacy}",
"received_events_url": "https://api.github.com/users/sam1373/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5525/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5525/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5524 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5524/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5524/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5524/events | https://github.com/NVIDIA/NeMo/pull/5524 | 1,468,765,129 | PR_kwDOC_bI7s5D7NbA | 5,524 | Conformer local attention | {
"login": "sam1373",
"id": 24465107,
"node_id": "MDQ6VXNlcjI0NDY1MTA3",
"avatar_url": "https://avatars.githubusercontent.com/u/24465107?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sam1373",
"html_url": "https://github.com/sam1373",
"followers_url": "https://api.github.com/users/sam1373/followers",
"following_url": "https://api.github.com/users/sam1373/following{/other_user}",
"gists_url": "https://api.github.com/users/sam1373/gists{/gist_id}",
"starred_url": "https://api.github.com/users/sam1373/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/sam1373/subscriptions",
"organizations_url": "https://api.github.com/users/sam1373/orgs",
"repos_url": "https://api.github.com/users/sam1373/repos",
"events_url": "https://api.github.com/users/sam1373/events{/privacy}",
"received_events_url": "https://api.github.com/users/sam1373/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1923495043,
"node_id": "MDU6TGFiZWwxOTIzNDk1MDQz",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/core",
"name": "core",
"color": "f25e3a",
"default": false,
"description": "Changes to NeMo Core"
},
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-29T22:07:02 | 2022-11-29T22:12:57 | 2022-11-29T22:12:57 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | true | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5524",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5524",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5524.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5524.patch",
"merged_at": null
} | # What does this PR do ?
Sliding window local attention for Conformer.
**Collection**: ASR
# Changelog
- adds option to use local attention for Conformer
- adds option for overriding/merging parameters into model config when using transcribe/eval scripts or loading model
# Usage
With an eval script can be used like this to switch to local attention with any Conformer checkpoint
```
python speech_to_text_eval.py \
(...other params...) \
++merge_into_model_config.encoder.att_context_size=[64,64] \
++merge_into_model_config.encoder.self_attention_model="rel_pos_local_attn"
```
Or when loading model
```python
merge_into_model_config = {
'encoder': dict({
'self_attention_model': "rel_pos_local_attn",
'att_context_size': [64, 64]
}),
}
merge_into_model_config = DictConfig(merge_into_model_config)
asr_model = ASRModel.from_pretrained('stt_en_conformer_ctc_large', map_location='cuda',
merge_into_model_config=merge_into_model_config)
```
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
| {
"login": "sam1373",
"id": 24465107,
"node_id": "MDQ6VXNlcjI0NDY1MTA3",
"avatar_url": "https://avatars.githubusercontent.com/u/24465107?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sam1373",
"html_url": "https://github.com/sam1373",
"followers_url": "https://api.github.com/users/sam1373/followers",
"following_url": "https://api.github.com/users/sam1373/following{/other_user}",
"gists_url": "https://api.github.com/users/sam1373/gists{/gist_id}",
"starred_url": "https://api.github.com/users/sam1373/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/sam1373/subscriptions",
"organizations_url": "https://api.github.com/users/sam1373/orgs",
"repos_url": "https://api.github.com/users/sam1373/repos",
"events_url": "https://api.github.com/users/sam1373/events{/privacy}",
"received_events_url": "https://api.github.com/users/sam1373/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5524/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5524/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5523 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5523/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5523/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5523/events | https://github.com/NVIDIA/NeMo/pull/5523 | 1,468,583,250 | PR_kwDOC_bI7s5D6l4A | 5,523 | prompt tuning fix for unscale grad errors | {
"login": "arendu",
"id": 108822655,
"node_id": "U_kgDOBnyAfw",
"avatar_url": "https://avatars.githubusercontent.com/u/108822655?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/arendu",
"html_url": "https://github.com/arendu",
"followers_url": "https://api.github.com/users/arendu/followers",
"following_url": "https://api.github.com/users/arendu/following{/other_user}",
"gists_url": "https://api.github.com/users/arendu/gists{/gist_id}",
"starred_url": "https://api.github.com/users/arendu/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/arendu/subscriptions",
"organizations_url": "https://api.github.com/users/arendu/orgs",
"repos_url": "https://api.github.com/users/arendu/repos",
"events_url": "https://api.github.com/users/arendu/events{/privacy}",
"received_events_url": "https://api.github.com/users/arendu/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
},
{
"id": 4840216855,
"node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI",
"name": "CI",
"color": "1D76DB",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-29T19:22:18 | 2023-02-21T08:34:48 | 2022-11-29T21:10:27 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5523",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5523",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5523.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5523.patch",
"merged_at": "2022-11-29T21:10:26"
} | # What does this PR do ?
Fixes the unscale gradient error.
**Collection**: [NLP]
# Changelog
- Adds CI test for simple prompt tuning
- Initializes the prompt embeddings with full prec
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [x] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "arendu",
"id": 108822655,
"node_id": "U_kgDOBnyAfw",
"avatar_url": "https://avatars.githubusercontent.com/u/108822655?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/arendu",
"html_url": "https://github.com/arendu",
"followers_url": "https://api.github.com/users/arendu/followers",
"following_url": "https://api.github.com/users/arendu/following{/other_user}",
"gists_url": "https://api.github.com/users/arendu/gists{/gist_id}",
"starred_url": "https://api.github.com/users/arendu/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/arendu/subscriptions",
"organizations_url": "https://api.github.com/users/arendu/orgs",
"repos_url": "https://api.github.com/users/arendu/repos",
"events_url": "https://api.github.com/users/arendu/events{/privacy}",
"received_events_url": "https://api.github.com/users/arendu/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5523/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5523/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5522 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5522/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5522/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5522/events | https://github.com/NVIDIA/NeMo/pull/5522 | 1,468,564,688 | PR_kwDOC_bI7s5D6h25 | 5,522 | T5 Eval bugfix | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4556669155,
"node_id": "LA_kwDOC_bI7s8AAAABD5lA4w",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/cherry-pick",
"name": "cherry-pick",
"color": "ededed",
"default": false,
"description": null
},
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-29T19:06:08 | 2022-12-05T05:14:43 | 2022-12-05T05:14:37 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5522",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5522",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5522.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5522.patch",
"merged_at": "2022-12-05T05:14:37"
} | # What does this PR do ?
minor bugfix
**Collection**: [Note which collection this PR will affect]
# Changelog
- Added function arg
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [X] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [X] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5522/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5522/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5521 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5521/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5521/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5521/events | https://github.com/NVIDIA/NeMo/pull/5521 | 1,468,445,967 | PR_kwDOC_bI7s5D6IVa | 5,521 | T5 Eval bugfix | {
"login": "Davood-M",
"id": 3964285,
"node_id": "MDQ6VXNlcjM5NjQyODU=",
"avatar_url": "https://avatars.githubusercontent.com/u/3964285?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Davood-M",
"html_url": "https://github.com/Davood-M",
"followers_url": "https://api.github.com/users/Davood-M/followers",
"following_url": "https://api.github.com/users/Davood-M/following{/other_user}",
"gists_url": "https://api.github.com/users/Davood-M/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Davood-M/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Davood-M/subscriptions",
"organizations_url": "https://api.github.com/users/Davood-M/orgs",
"repos_url": "https://api.github.com/users/Davood-M/repos",
"events_url": "https://api.github.com/users/Davood-M/events{/privacy}",
"received_events_url": "https://api.github.com/users/Davood-M/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-11-29T17:29:14 | 2022-11-29T19:05:47 | 2022-11-29T19:05:37 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5521",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5521",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5521.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5521.patch",
"merged_at": "2022-11-29T19:05:37"
} | # What does this PR do ?
minor bugfix
**Collection**: [Note which collection this PR will affect]
# Changelog
- Added function arg
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [X] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [X] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "okuchaiev",
"id": 22035961,
"node_id": "MDQ6VXNlcjIyMDM1OTYx",
"avatar_url": "https://avatars.githubusercontent.com/u/22035961?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/okuchaiev",
"html_url": "https://github.com/okuchaiev",
"followers_url": "https://api.github.com/users/okuchaiev/followers",
"following_url": "https://api.github.com/users/okuchaiev/following{/other_user}",
"gists_url": "https://api.github.com/users/okuchaiev/gists{/gist_id}",
"starred_url": "https://api.github.com/users/okuchaiev/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/okuchaiev/subscriptions",
"organizations_url": "https://api.github.com/users/okuchaiev/orgs",
"repos_url": "https://api.github.com/users/okuchaiev/repos",
"events_url": "https://api.github.com/users/okuchaiev/events{/privacy}",
"received_events_url": "https://api.github.com/users/okuchaiev/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5521/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5521/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5520 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5520/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5520/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5520/events | https://github.com/NVIDIA/NeMo/issues/5520 | 1,468,102,385 | I_kwDOC_bI7s5XgXbx | 5,520 | Way to use `prefetch_factor` in ASR training | {
"login": "menon92",
"id": 11752205,
"node_id": "MDQ6VXNlcjExNzUyMjA1",
"avatar_url": "https://avatars.githubusercontent.com/u/11752205?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/menon92",
"html_url": "https://github.com/menon92",
"followers_url": "https://api.github.com/users/menon92/followers",
"following_url": "https://api.github.com/users/menon92/following{/other_user}",
"gists_url": "https://api.github.com/users/menon92/gists{/gist_id}",
"starred_url": "https://api.github.com/users/menon92/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/menon92/subscriptions",
"organizations_url": "https://api.github.com/users/menon92/orgs",
"repos_url": "https://api.github.com/users/menon92/repos",
"events_url": "https://api.github.com/users/menon92/events{/privacy}",
"received_events_url": "https://api.github.com/users/menon92/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1719393562,
"node_id": "MDU6TGFiZWwxNzE5MzkzNTYy",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/feature",
"name": "feature",
"color": "68F3F0",
"default": false,
"description": "request/PR for a new feature"
},
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-11-29T13:52:14 | 2023-09-04T01:44:36 | 2023-09-04T01:44:35 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Is your feature request related to a problem? Please describe.**
The PyTorch [dataloader](https://pytorch.org/docs/stable/data.html) supports pre-fetching the batch of data so that GPU doesn't wait for data processing. But I could not find any options to set the `prefetch_factor` parameter in the example config.YAML file
**Describe the solution you'd like**
Like others dataloader parameter, we can also set this prefetch_factor param in the config YAML file and use this during training
A clear and concise description of what you want to happen.
Thanks
| {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5520/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5520/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5519 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5519/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5519/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5519/events | https://github.com/NVIDIA/NeMo/issues/5519 | 1,467,508,989 | I_kwDOC_bI7s5XeGj9 | 5,519 | ssl conformer learning rate & warmup | {
"login": "alexandercesarr",
"id": 105767588,
"node_id": "U_kgDOBk3ipA",
"avatar_url": "https://avatars.githubusercontent.com/u/105767588?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/alexandercesarr",
"html_url": "https://github.com/alexandercesarr",
"followers_url": "https://api.github.com/users/alexandercesarr/followers",
"following_url": "https://api.github.com/users/alexandercesarr/following{/other_user}",
"gists_url": "https://api.github.com/users/alexandercesarr/gists{/gist_id}",
"starred_url": "https://api.github.com/users/alexandercesarr/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/alexandercesarr/subscriptions",
"organizations_url": "https://api.github.com/users/alexandercesarr/orgs",
"repos_url": "https://api.github.com/users/alexandercesarr/repos",
"events_url": "https://api.github.com/users/alexandercesarr/events{/privacy}",
"received_events_url": "https://api.github.com/users/alexandercesarr/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
},
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | {
"login": "sam1373",
"id": 24465107,
"node_id": "MDQ6VXNlcjI0NDY1MTA3",
"avatar_url": "https://avatars.githubusercontent.com/u/24465107?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sam1373",
"html_url": "https://github.com/sam1373",
"followers_url": "https://api.github.com/users/sam1373/followers",
"following_url": "https://api.github.com/users/sam1373/following{/other_user}",
"gists_url": "https://api.github.com/users/sam1373/gists{/gist_id}",
"starred_url": "https://api.github.com/users/sam1373/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/sam1373/subscriptions",
"organizations_url": "https://api.github.com/users/sam1373/orgs",
"repos_url": "https://api.github.com/users/sam1373/repos",
"events_url": "https://api.github.com/users/sam1373/events{/privacy}",
"received_events_url": "https://api.github.com/users/sam1373/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "sam1373",
"id": 24465107,
"node_id": "MDQ6VXNlcjI0NDY1MTA3",
"avatar_url": "https://avatars.githubusercontent.com/u/24465107?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sam1373",
"html_url": "https://github.com/sam1373",
"followers_url": "https://api.github.com/users/sam1373/followers",
"following_url": "https://api.github.com/users/sam1373/following{/other_user}",
"gists_url": "https://api.github.com/users/sam1373/gists{/gist_id}",
"starred_url": "https://api.github.com/users/sam1373/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/sam1373/subscriptions",
"organizations_url": "https://api.github.com/users/sam1373/orgs",
"repos_url": "https://api.github.com/users/sam1373/repos",
"events_url": "https://api.github.com/users/sam1373/events{/privacy}",
"received_events_url": "https://api.github.com/users/sam1373/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-11-29T06:08:22 | 2023-02-05T02:02:03 | 2023-02-05T02:02:03 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | Hi, I've trained a **SSL conformer** model. Now I wanna **fine-tune** it with labeled data.
I encountered two problems:
1. I want the learning rate decrease linearly.
I don't want warmup. So I set it 0.
But in this situation, it first increase and then decrease.
2. the second problem is I want to control the learning rate decrease speed. For this one I changed the weight_decay and also min_lr but it decreased very fast. How can I control it?
I want it be constant. | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5519/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5519/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5517 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5517/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5517/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5517/events | https://github.com/NVIDIA/NeMo/issues/5517 | 1,467,394,978 | I_kwDOC_bI7s5Xdqui | 5,517 | Convert megatron lm ckpt to nemo | {
"login": "bew-pbwt",
"id": 78535791,
"node_id": "MDQ6VXNlcjc4NTM1Nzkx",
"avatar_url": "https://avatars.githubusercontent.com/u/78535791?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/bew-pbwt",
"html_url": "https://github.com/bew-pbwt",
"followers_url": "https://api.github.com/users/bew-pbwt/followers",
"following_url": "https://api.github.com/users/bew-pbwt/following{/other_user}",
"gists_url": "https://api.github.com/users/bew-pbwt/gists{/gist_id}",
"starred_url": "https://api.github.com/users/bew-pbwt/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/bew-pbwt/subscriptions",
"organizations_url": "https://api.github.com/users/bew-pbwt/orgs",
"repos_url": "https://api.github.com/users/bew-pbwt/repos",
"events_url": "https://api.github.com/users/bew-pbwt/events{/privacy}",
"received_events_url": "https://api.github.com/users/bew-pbwt/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
},
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | {
"login": "yidong72",
"id": 43824965,
"node_id": "MDQ6VXNlcjQzODI0OTY1",
"avatar_url": "https://avatars.githubusercontent.com/u/43824965?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/yidong72",
"html_url": "https://github.com/yidong72",
"followers_url": "https://api.github.com/users/yidong72/followers",
"following_url": "https://api.github.com/users/yidong72/following{/other_user}",
"gists_url": "https://api.github.com/users/yidong72/gists{/gist_id}",
"starred_url": "https://api.github.com/users/yidong72/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/yidong72/subscriptions",
"organizations_url": "https://api.github.com/users/yidong72/orgs",
"repos_url": "https://api.github.com/users/yidong72/repos",
"events_url": "https://api.github.com/users/yidong72/events{/privacy}",
"received_events_url": "https://api.github.com/users/yidong72/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "yidong72",
"id": 43824965,
"node_id": "MDQ6VXNlcjQzODI0OTY1",
"avatar_url": "https://avatars.githubusercontent.com/u/43824965?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/yidong72",
"html_url": "https://github.com/yidong72",
"followers_url": "https://api.github.com/users/yidong72/followers",
"following_url": "https://api.github.com/users/yidong72/following{/other_user}",
"gists_url": "https://api.github.com/users/yidong72/gists{/gist_id}",
"starred_url": "https://api.github.com/users/yidong72/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/yidong72/subscriptions",
"organizations_url": "https://api.github.com/users/yidong72/orgs",
"repos_url": "https://api.github.com/users/yidong72/repos",
"events_url": "https://api.github.com/users/yidong72/events{/privacy}",
"received_events_url": "https://api.github.com/users/yidong72/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-11-29T03:34:04 | 2023-01-19T02:00:28 | 2023-01-19T02:00:28 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
I have tried converting megatron lm ckpt to nemo (pytorch .pt to .nemo file ) with model_optim_rng.pt from nvidia/megatron_bert_345m by using [https://github.com/NVIDIA/NeMo/tree/main/examples/nlp/language_modeling)/megatron_ckpt_to_nemo.py](url) script
> wget --content-disposition https://api.ngc.nvidia.com/v2/models/nvidia/megatron_bert_345m/versions/v0.1_uncased/zip -O megatron_bert_345m_v0.1_uncased.zip
> unzip megatron_bert_345m_v0.1_uncased.zip
> python -m torch.distributed.launch --nproc_per_node=1 megatron_lm_ckpt_to_nemo.py \
> --checkpoint_folder release/mp_rank_00/ \
> --checkpoint_name model_optim_rng.pt \
> --nemo_file_path all-thai-lm.nemo \
> --model_type bert \
> --tensor_model_parallel_size 2 \
> --pipeline_model_parallel_size 2
seem like torch has some issue with GPU
```
[NeMo W 2022-11-29 02:35:59 optimizers:77] Could not import distributed_fused_adam optimizer from Apex
OMP: Info #276: omp_set_nested routine deprecated, please use omp_set_max_active_levels instead.
[NeMo W 2022-11-29 02:36:11 experimental:27] Module <class 'nemo.collections.nlp.data.language_modeling.megatron.megatron_batch_samplers.MegatronPretrainingRandomBatchSampler'> is experimental, not ready for production and is not fully supported. Use at your own risk.
[NeMo W 2022-11-29 02:36:11 experimental:27] Module <class 'nemo.collections.nlp.models.text_normalization_as_tagging.thutmose_tagger.ThutmoseTaggerModel'> is experimental, not ready for production and is not fully supported. Use at your own risk.
[NeMo I 2022-11-29 02:36:11 distributed:31] Initializing torch.distributed with local_rank: 0, rank: 0, world_size: 1
Traceback (most recent call last):
File "megatron_lm_ckpt_to_nemo.py", line 476, in <module>
assert world_size % args.tensor_model_parallel_size == 0
AssertionError
/opt/conda/lib/python3.8/site-packages/torch/distributed/launch.py:178: FutureWarning: The module torch.distributed.launch is deprecated
and will be removed in future. Use torchrun.
Note that --use_env is set by default in torchrun.
If your script expects `--local_rank` argument to be set, please
change it to read from `os.environ['LOCAL_RANK']` instead. See
https://pytorch.org/docs/stable/distributed.html#launch-utility for
further instructions
warnings.warn(
ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank: 0 (pid: 9735) of binary: /opt/conda/bin/python
Traceback (most recent call last):
File "/opt/conda/lib/python3.8/runpy.py", line 194, in _run_module_as_main
return _run_code(code, main_globals, None,
File "/opt/conda/lib/python3.8/runpy.py", line 87, in _run_code
exec(code, run_globals)
File "/opt/conda/lib/python3.8/site-packages/torch/distributed/launch.py", line 193, in <module>
main()
File "/opt/conda/lib/python3.8/site-packages/torch/distributed/launch.py", line 189, in main
launch(args)
File "/opt/conda/lib/python3.8/site-packages/torch/distributed/launch.py", line 174, in launch
run(args)
File "/opt/conda/lib/python3.8/site-packages/torch/distributed/run.py", line 752, in run
elastic_launch(
File "/opt/conda/lib/python3.8/site-packages/torch/distributed/launcher/api.py", line 132, in __call__
return launch_agent(self._config, self._entrypoint, list(args))
File "/opt/conda/lib/python3.8/site-packages/torch/distributed/launcher/api.py", line 246, in launch_agent
raise ChildFailedError(
torch.distributed.elastic.multiprocessing.errors.ChildFailedError:
============================================================
megatron_lm_ckpt_to_nemo.py FAILED
------------------------------------------------------------
Failures:
<NO_OTHER_FAILURES>
------------------------------------------------------------
Root Cause (first observed failure):
[0]:
time : 2022-11-29_02:36:16
host : e792987d9454
rank : 0 (local_rank: 0)
exitcode : 1 (pid: 9735)
error_file: <N/A>
traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html
============================================================
```
inside docker can detect all GPUs (nvidia-smi)
**Expected behavior**
Should get nemo LM model
**Environment overview (please complete the following information)**
- Environment location: nvcr.io/nvidia/nemo:22.08 docker
- Method of NeMo install: inside image already had nemo
> sudo docker pull nvcr.io/nvidia/nemo:22.08 && sudo nvidia-docker run -it -v --shm-size=16g -p 8888:8888 -p 6006:6006 --ulimit memlock=-1 --ulimit stack=67108864 nvcr.io/nvidia/nemo:22.08
**Additional context**
GPU: Tesla V100
server: nvidia DGX1 | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5517/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5517/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5514 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5514/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5514/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5514/events | https://github.com/NVIDIA/NeMo/pull/5514 | 1,466,811,146 | PR_kwDOC_bI7s5D0rvi | 5,514 | Multi-blank Transducers | {
"login": "hainan-xv",
"id": 5440014,
"node_id": "MDQ6VXNlcjU0NDAwMTQ=",
"avatar_url": "https://avatars.githubusercontent.com/u/5440014?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/hainan-xv",
"html_url": "https://github.com/hainan-xv",
"followers_url": "https://api.github.com/users/hainan-xv/followers",
"following_url": "https://api.github.com/users/hainan-xv/following{/other_user}",
"gists_url": "https://api.github.com/users/hainan-xv/gists{/gist_id}",
"starred_url": "https://api.github.com/users/hainan-xv/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/hainan-xv/subscriptions",
"organizations_url": "https://api.github.com/users/hainan-xv/orgs",
"repos_url": "https://api.github.com/users/hainan-xv/repos",
"events_url": "https://api.github.com/users/hainan-xv/events{/privacy}",
"received_events_url": "https://api.github.com/users/hainan-xv/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-28T17:30:08 | 2022-11-30T04:01:42 | 2022-11-30T04:01:41 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | true | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5514",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5514",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5514.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5514.patch",
"merged_at": null
} | # What does this PR do ?
Add support for multi-blank transducers described in https://arxiv.org/pdf/2211.03541
**Collection**: [Note which collection this PR will affect]
ASR
# Changelog
- implementation of multi-blank both with numba kernel code as well as pytorch.
- add example .yaml file for multi-blank ASR model configuration.
# Usage
Multi-blank model training could be invoked by using the .yaml file included in this PR and call.
```
python NeMo/examples/asr/asr_transducer/speech_to_text_rnnt_bpe.py \\
[your options here]
```
with your options for data/hyperparameters/etc. This trains a multi-blank model with big-blank durations [2, 4, 8].
To make changes in the big blank durations, change the options by,
```
model.loss.multiblank_rnnt_kwargs.big_blank_durations=[2,3,5,8] # or any other set of durations you want
model.decoding.big_blank_durations=[2,3,5,8] # used for decoding
model.joint.num_extra_output=4 # this needs to be consistent with the duration list you set
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "hainan-xv",
"id": 5440014,
"node_id": "MDQ6VXNlcjU0NDAwMTQ=",
"avatar_url": "https://avatars.githubusercontent.com/u/5440014?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/hainan-xv",
"html_url": "https://github.com/hainan-xv",
"followers_url": "https://api.github.com/users/hainan-xv/followers",
"following_url": "https://api.github.com/users/hainan-xv/following{/other_user}",
"gists_url": "https://api.github.com/users/hainan-xv/gists{/gist_id}",
"starred_url": "https://api.github.com/users/hainan-xv/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/hainan-xv/subscriptions",
"organizations_url": "https://api.github.com/users/hainan-xv/orgs",
"repos_url": "https://api.github.com/users/hainan-xv/repos",
"events_url": "https://api.github.com/users/hainan-xv/events{/privacy}",
"received_events_url": "https://api.github.com/users/hainan-xv/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5514/reactions",
"total_count": 1,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 1,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5514/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5513 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5513/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5513/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5513/events | https://github.com/NVIDIA/NeMo/issues/5513 | 1,466,212,227 | I_kwDOC_bI7s5XZJ-D | 5,513 | Conformer_Transducer's encoder exported to ONNX fails to instantiate an `onnxruntime.InferenceSession` | {
"login": "NiziL",
"id": 5002613,
"node_id": "MDQ6VXNlcjUwMDI2MTM=",
"avatar_url": "https://avatars.githubusercontent.com/u/5002613?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/NiziL",
"html_url": "https://github.com/NiziL",
"followers_url": "https://api.github.com/users/NiziL/followers",
"following_url": "https://api.github.com/users/NiziL/following{/other_user}",
"gists_url": "https://api.github.com/users/NiziL/gists{/gist_id}",
"starred_url": "https://api.github.com/users/NiziL/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/NiziL/subscriptions",
"organizations_url": "https://api.github.com/users/NiziL/orgs",
"repos_url": "https://api.github.com/users/NiziL/repos",
"events_url": "https://api.github.com/users/NiziL/events{/privacy}",
"received_events_url": "https://api.github.com/users/NiziL/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
}
] | closed | false | null | [] | null | null | 2022-11-28T10:57:10 | 2022-11-29T11:39:19 | 2022-11-29T10:48:23 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
Creating an onnx `InferenceSession` using the exported encoder from a conformer_transducer model (using the default parameters of `export`) leads to an `ONNXRuntimeError`.
```
[ONNXRuntimeError] : 1 : FAIL : Load model from /tmp/tmp_ssi0eq1/encoder-test.onnx failed:Type Error: Type parameter (Tind) of Optype (Slice) bound to different types (tensor(int32) and tensor(int64) in node (/pos_enc/Slice).
```
**Steps/Code to reproduce bug**
Here's a minimal example to reproduce the bug.
```python
#! /usr/bin/env python3
# -*- coding: utf-8 -*-
import tempfile
import onnxruntime as ort
from nemo.collections.asr.models import EncDecRNNTBPEModel
def main():
model = EncDecRNNTBPEModel.from_pretrained(
"stt_fr_conformer_transducer_large"
)
with tempfile.TemporaryDirectory() as temp_dir:
model.export(temp_dir + "/test.onnx")
try:
ort.InferenceSession(temp_dir + "/encoder-test.onnx")
print("encoder has been sucessfully initialized")
except Exception as e:
print("Error on encoder")
print(e)
try:
ort.InferenceSession(temp_dir + "/decoder_joint-test.onnx")
print("decoder_joint has been sucessfully initialized")
except Exception as e:
print("Error on decoder_joint")
print(e)
if __name__ == "__main__":
main()
```
**Expected behavior**
Well... The `InferenceSession` should be correctly initialized from the exported onnx file, shouldn't it ?
**Environment overview and details**
- Environment location:
- Method of NeMo install:
```
> conda create -n NeMo python==3.9
> pip install Cython
> pip install "nemo_toolkit[all]"
```
- OS version
```
> uname -srvpio
Linux 6.0.9-arch1-1 #1 SMP PREEMPT_DYNAMIC Wed, 16 Nov 2022 17:01:17 +0000 unknown unknown GNU/Linux
```
- PyTorch version
```
> python
>>> import torch
>>> torch.__version__
'1.13.0'
```
- Python version
```
> python --version
Python 3.9.0
```
| {
"login": "NiziL",
"id": 5002613,
"node_id": "MDQ6VXNlcjUwMDI2MTM=",
"avatar_url": "https://avatars.githubusercontent.com/u/5002613?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/NiziL",
"html_url": "https://github.com/NiziL",
"followers_url": "https://api.github.com/users/NiziL/followers",
"following_url": "https://api.github.com/users/NiziL/following{/other_user}",
"gists_url": "https://api.github.com/users/NiziL/gists{/gist_id}",
"starred_url": "https://api.github.com/users/NiziL/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/NiziL/subscriptions",
"organizations_url": "https://api.github.com/users/NiziL/orgs",
"repos_url": "https://api.github.com/users/NiziL/repos",
"events_url": "https://api.github.com/users/NiziL/events{/privacy}",
"received_events_url": "https://api.github.com/users/NiziL/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5513/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5513/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5512 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5512/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5512/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5512/events | https://github.com/NVIDIA/NeMo/pull/5512 | 1,465,548,343 | PR_kwDOC_bI7s5DwdWl | 5,512 | add exportable mel spec | {
"login": "1-800-BAD-CODE",
"id": 50530592,
"node_id": "MDQ6VXNlcjUwNTMwNTky",
"avatar_url": "https://avatars.githubusercontent.com/u/50530592?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/1-800-BAD-CODE",
"html_url": "https://github.com/1-800-BAD-CODE",
"followers_url": "https://api.github.com/users/1-800-BAD-CODE/followers",
"following_url": "https://api.github.com/users/1-800-BAD-CODE/following{/other_user}",
"gists_url": "https://api.github.com/users/1-800-BAD-CODE/gists{/gist_id}",
"starred_url": "https://api.github.com/users/1-800-BAD-CODE/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/1-800-BAD-CODE/subscriptions",
"organizations_url": "https://api.github.com/users/1-800-BAD-CODE/orgs",
"repos_url": "https://api.github.com/users/1-800-BAD-CODE/repos",
"events_url": "https://api.github.com/users/1-800-BAD-CODE/events{/privacy}",
"received_events_url": "https://api.github.com/users/1-800-BAD-CODE/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-27T19:15:16 | 2022-11-28T19:25:16 | 2022-11-28T19:25:16 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5512",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5512",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5512.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5512.patch",
"merged_at": "2022-11-28T19:25:16"
} | Signed-off-by: shane carroll <[email protected]>
# What does this PR do ?
`AudioToMelSpectrogramPreprocessor` accepts a bool argument `use_torchaudio` which, if True, switches the `featurizer` to a `torchaudio`-based extractor which produces the same features but with an exportable graph.
New preprocessor mimics the old implementation sufficiently well to swap out the preprocessor of pre-trained models and export them.
Preprocessor can be exported to JIT; ONNX is blocked by https://github.com/pytorch/pytorch/issues/81075
**Collection**: ASR
# Changelog
Add an option to `AudioToMelSpectrogramPreprocessor` which can alter `featurizer`.
Add a class `FilterbankFeaturesTA` which is analogous to `FilterbankFeatures`
# Usage
The following script will
1. Load a pre-trained English Conformer
2. Create a copy of the model's preprocessor, but with a `torchaudio` backend
3. Compare new and old preprocessor outputs
4. Export the preprocessor to JIT
5. Compare JIT and PyTorch outputs
6. Swap the pre-trained model's preprocessor to the new one, check WER matches old
```python
from pathlib import Path
import torch
import hydra
from pytorch_lightning import seed_everything
from omegaconf import open_dict
from nemo.utils import logging
from nemo.utils.nemo_logging import Logger
from nemo.collections.asr.metrics.wer import word_error_rate
from nemo.collections.asr.models import ASRModel
from nemo.collections.asr.modules.audio_preprocessing import AudioToMelSpectrogramPreprocessor
# Optionally use a seed
# seed_everything(42)
logging.set_verbosity(Logger.CRITICAL)
# Get a pre-trained ASR model to compare preprocessors. Any model with a mel spec extractor should work.
m: ASRModel = ASRModel.from_pretrained("stt_en_conformer_ctc_small", map_location=torch.device("cpu"))
m.eval()
old_preprocessor = m.preprocessor
# Extract preprocessor config and set the flag to use the torchaudio-based extractor; keep all other arguments the same
new_config = m.cfg.preprocessor
with open_dict(new_config):
new_config.use_torchaudio = True
# Instantiate an instance that uses torchaudio on the backend
new_preprocessor: AudioToMelSpectrogramPreprocessor = hydra.utils.instantiate(config=new_config)
new_preprocessor.eval()
print(f"New preprocessor featurizer type: {type(new_preprocessor.featurizer)}")
# Export the torchaudio preprocessor and load it back in as a `ScriptModule`.
new_preprocessor.export("tmp.pt")
jit_preprocessor = torch.jit.load("tmp.pt")
# Generate random input
batch_size = 4
max_length = 16000
signals = torch.randn(size=[batch_size, max_length])
lengths = torch.randint(low=200, high=max_length, size=[batch_size])
lengths[0] = max_length
# Extract features with all preprocessors
old_feats, old_feat_lens = old_preprocessor(input_signal=signals, length=lengths)
new_feats, new_feat_lens = new_preprocessor(input_signal=signals, length=lengths)
jit_feats, jit_feat_lens = jit_preprocessor(input_signal=signals, length=lengths)
# Make sure new output matches old output
# Need to relax the tolerance from defaults. We will check WER also, as an alternative verification of correctness.
rel_tolerance = 1e-2
abs_tolerance = 1e-4
torch.testing.assert_allclose(actual=new_feats, expected=old_feats, atol=abs_tolerance, rtol=rel_tolerance)
# Zero tolerance for integer lengths.
torch.testing.assert_allclose(actual=new_feat_lens, expected=old_feat_lens, atol=0, rtol=0)
print(f"Output comparison passed with relative tolerance {rel_tolerance} and absolute tolerance {abs_tolerance}.")
# Make sure JIT output matches PyTorch output
# Keep tolerance at defaults for JIT comparison
torch.testing.assert_allclose(actual=jit_feats, expected=new_feats)
torch.testing.assert_allclose(actual=jit_feat_lens, expected=new_feat_lens, atol=0, rtol=0)
print(f"Jit comparison passed with default tolerance.")
# To run a WER check you'll need to comment out some assumptions in the CTC transcribe method, as addressed in https://github.com/NVIDIA/NeMo/pull/2762
# print("Testing WER with old/new preprocessor with some LibriSpeech data")
# We only need audio files; we're comparing model outputs to each other, not references
# dev_other_dir = "/path/to/LibriSpeech/dev-other"
# num_files_to_use = 100
# audio_files = [str(x) for x in Path(dev_other_dir).rglob("*.flac")]
# audio_files = audio_files[:num_files_to_use]
# print("Transcribing with the baseline model")
# old_output = m.transcribe(audio_files)
# m.preprocessor = new_preprocessor
# print("Transcribing after switching the preprocessor")
# new_output = m.transcribe(audio_files)
# wer = word_error_rate(hypotheses=new_output, references=old_output)
# print(f"WER with {len(audio_files)} audio files using old vs. new preprocessor is {wer}")
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [x] Did you write any new necessary tests?
- [x] Did you add or update any necessary documentation?
- [x] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to: this issue comes up once in a while and is generally brushed aside.
| {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5512/reactions",
"total_count": 2,
"+1": 1,
"-1": 0,
"laugh": 0,
"hooray": 1,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5512/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5510 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5510/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5510/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5510/events | https://github.com/NVIDIA/NeMo/issues/5510 | 1,465,349,382 | I_kwDOC_bI7s5XV3UG | 5,510 | Benchmarks on BPE vs Char based tokens | {
"login": "mehadi92",
"id": 104126711,
"node_id": "U_kgDOBjTY9w",
"avatar_url": "https://avatars.githubusercontent.com/u/104126711?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/mehadi92",
"html_url": "https://github.com/mehadi92",
"followers_url": "https://api.github.com/users/mehadi92/followers",
"following_url": "https://api.github.com/users/mehadi92/following{/other_user}",
"gists_url": "https://api.github.com/users/mehadi92/gists{/gist_id}",
"starred_url": "https://api.github.com/users/mehadi92/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/mehadi92/subscriptions",
"organizations_url": "https://api.github.com/users/mehadi92/orgs",
"repos_url": "https://api.github.com/users/mehadi92/repos",
"events_url": "https://api.github.com/users/mehadi92/events{/privacy}",
"received_events_url": "https://api.github.com/users/mehadi92/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | {
"login": "okuchaiev",
"id": 22035961,
"node_id": "MDQ6VXNlcjIyMDM1OTYx",
"avatar_url": "https://avatars.githubusercontent.com/u/22035961?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/okuchaiev",
"html_url": "https://github.com/okuchaiev",
"followers_url": "https://api.github.com/users/okuchaiev/followers",
"following_url": "https://api.github.com/users/okuchaiev/following{/other_user}",
"gists_url": "https://api.github.com/users/okuchaiev/gists{/gist_id}",
"starred_url": "https://api.github.com/users/okuchaiev/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/okuchaiev/subscriptions",
"organizations_url": "https://api.github.com/users/okuchaiev/orgs",
"repos_url": "https://api.github.com/users/okuchaiev/repos",
"events_url": "https://api.github.com/users/okuchaiev/events{/privacy}",
"received_events_url": "https://api.github.com/users/okuchaiev/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "okuchaiev",
"id": 22035961,
"node_id": "MDQ6VXNlcjIyMDM1OTYx",
"avatar_url": "https://avatars.githubusercontent.com/u/22035961?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/okuchaiev",
"html_url": "https://github.com/okuchaiev",
"followers_url": "https://api.github.com/users/okuchaiev/followers",
"following_url": "https://api.github.com/users/okuchaiev/following{/other_user}",
"gists_url": "https://api.github.com/users/okuchaiev/gists{/gist_id}",
"starred_url": "https://api.github.com/users/okuchaiev/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/okuchaiev/subscriptions",
"organizations_url": "https://api.github.com/users/okuchaiev/orgs",
"repos_url": "https://api.github.com/users/okuchaiev/repos",
"events_url": "https://api.github.com/users/okuchaiev/events{/privacy}",
"received_events_url": "https://api.github.com/users/okuchaiev/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-11-27T06:44:36 | 2022-11-27T10:48:44 | 2022-11-27T06:45:34 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | Hi, Is there any comparison of using PBE vs Character base tokens? How much WER changes when using BPE instead of character - Thanks
| {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5510/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5510/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5509 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5509/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5509/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5509/events | https://github.com/NVIDIA/NeMo/issues/5509 | 1,465,316,675 | I_kwDOC_bI7s5XVvVD | 5,509 | Cannot finetune FastPitch on a custom dataset | {
"login": "godspirit00",
"id": 16176151,
"node_id": "MDQ6VXNlcjE2MTc2MTUx",
"avatar_url": "https://avatars.githubusercontent.com/u/16176151?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/godspirit00",
"html_url": "https://github.com/godspirit00",
"followers_url": "https://api.github.com/users/godspirit00/followers",
"following_url": "https://api.github.com/users/godspirit00/following{/other_user}",
"gists_url": "https://api.github.com/users/godspirit00/gists{/gist_id}",
"starred_url": "https://api.github.com/users/godspirit00/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/godspirit00/subscriptions",
"organizations_url": "https://api.github.com/users/godspirit00/orgs",
"repos_url": "https://api.github.com/users/godspirit00/repos",
"events_url": "https://api.github.com/users/godspirit00/events{/privacy}",
"received_events_url": "https://api.github.com/users/godspirit00/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-11-27T03:12:24 | 2022-11-27T03:17:39 | 2022-11-27T03:17:39 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | I followed the finetuning tutorial and tried to finetune the pretrained LJSpeech model with the command
`python examples/tts/fastpitch_finetune.py --config-name=fastpitch_align_v1.05.yaml train_dataset=/root/autodl-tmp/ryanspeech.train.json validation_datasets=/root/autodl-tmp/ryanspeech.val.json sup_data_path=/root/autodl-tmp/nemo-training/ryan_22k/sup_data exp_manager.exp_dir=/root/autodl-tmp/nemo-training/ryan_22k +init_from_nemo_model=/root/autodl-tmp/tts_en_fastpitch_align.nemo model.train_ds.dataloader_params.batch_size=24 pitch_mean=150.32728576660156 pitch_std=42.45840835571289 pitch_fmin=65.4063949584961 pitch_fmax=1760.0 model.optim.lr=2e-4 ~model.optim.sched`
But I got the error:
```
Traceback (most recent call last):
File "examples/tts/fastpitch_finetune.py", line 33, in main
model.maybe_init_from_pretrained_checkpoint(cfg=cfg)
File "/root/miniconda3/lib/python3.8/site-packages/pytorch_lightning/utilities/rank_zero.py", line 32, in wrapped_fn
return fn(*args, **kwargs)
File "/root/NeMo/nemo/core/classes/modelPT.py", line 1066, in maybe_init_from_pretrained_checkpoint
self.load_state_dict(restored_model.state_dict(), strict=False)
File "/root/miniconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1482, in load_state_dict
raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
RuntimeError: Error(s) in loading state_dict for FastPitchModel:
size mismatch for fastpitch.encoder.word_emb.weight: copying a param with shape torch.Size([115, 384]) from checkpoint, the shape in current model is torch.Size([114, 384]).
```
-------------------
EDIT:
seems like it has something to do with the `+model.text_tokenizer.add_blank_at=true` in the tutorial.
Problem solved. | {
"login": "godspirit00",
"id": 16176151,
"node_id": "MDQ6VXNlcjE2MTc2MTUx",
"avatar_url": "https://avatars.githubusercontent.com/u/16176151?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/godspirit00",
"html_url": "https://github.com/godspirit00",
"followers_url": "https://api.github.com/users/godspirit00/followers",
"following_url": "https://api.github.com/users/godspirit00/following{/other_user}",
"gists_url": "https://api.github.com/users/godspirit00/gists{/gist_id}",
"starred_url": "https://api.github.com/users/godspirit00/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/godspirit00/subscriptions",
"organizations_url": "https://api.github.com/users/godspirit00/orgs",
"repos_url": "https://api.github.com/users/godspirit00/repos",
"events_url": "https://api.github.com/users/godspirit00/events{/privacy}",
"received_events_url": "https://api.github.com/users/godspirit00/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5509/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5509/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5508 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5508/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5508/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5508/events | https://github.com/NVIDIA/NeMo/pull/5508 | 1,465,270,856 | PR_kwDOC_bI7s5DvoCM | 5,508 | Exportable mel spectrogram preprocessor | {
"login": "1-800-BAD-CODE",
"id": 50530592,
"node_id": "MDQ6VXNlcjUwNTMwNTky",
"avatar_url": "https://avatars.githubusercontent.com/u/50530592?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/1-800-BAD-CODE",
"html_url": "https://github.com/1-800-BAD-CODE",
"followers_url": "https://api.github.com/users/1-800-BAD-CODE/followers",
"following_url": "https://api.github.com/users/1-800-BAD-CODE/following{/other_user}",
"gists_url": "https://api.github.com/users/1-800-BAD-CODE/gists{/gist_id}",
"starred_url": "https://api.github.com/users/1-800-BAD-CODE/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/1-800-BAD-CODE/subscriptions",
"organizations_url": "https://api.github.com/users/1-800-BAD-CODE/orgs",
"repos_url": "https://api.github.com/users/1-800-BAD-CODE/repos",
"events_url": "https://api.github.com/users/1-800-BAD-CODE/events{/privacy}",
"received_events_url": "https://api.github.com/users/1-800-BAD-CODE/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
},
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-26T22:33:25 | 2022-11-27T19:01:53 | 2022-11-26T22:59:13 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5508",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5508",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5508.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5508.patch",
"merged_at": null
} | # What does this PR do ?
`AudioToMelSpectrogramPreprocessor` accepts a bool argument `use_torchaudio` which, if True, switches the `featurizer` to a `torchaudio`-based extractor which produces the same features but with an exportable graph.
New preprocessor mimics the old implementation sufficiently well to swap out the preprocessor of pre-trained models and export them.
Preprocessor can be exported to JIT; ONNX is blocked by https://github.com/pytorch/pytorch/issues/81075
**Collection**: ASR
# Changelog
Add an option to `AudioToMelSpectrogramPreprocessor` which can alter `featurizer`.
Add a class `FilterbankFeaturesTA` which is analogous to `FilterbankFeatures`
# Usage
The following script will
1. Load a pre-trained English Conformer
2. Create a copy of the model's preprocessor, but with a `torchaudio` backend
3. Compare new and old preprocessor outputs
4. Export the preprocessor to JIT
5. Compare JIT and PyTorch outputs
6. Swap the pre-trained model's preprocessor to the new one, check WER matches old
```python
from pathlib import Path
import torch
import hydra
from pytorch_lightning import seed_everything
from omegaconf import open_dict
from nemo.utils import logging
from nemo.utils.nemo_logging import Logger
from nemo.collections.asr.metrics.wer import word_error_rate
from nemo.collections.asr.models import ASRModel
from nemo.collections.asr.modules.audio_preprocessing import AudioToMelSpectrogramPreprocessor
# Optionally use a seed
# seed_everything(42)
logging.set_verbosity(Logger.CRITICAL)
# Get a pre-trained ASR model to compare preprocessors. Any model with a mel spec extractor should work.
m: ASRModel = ASRModel.from_pretrained("stt_en_conformer_ctc_small", map_location=torch.device("cpu"))
m.eval()
old_preprocessor = m.preprocessor
# Extract preprocessor config and set the flag to use the torchaudio-based extractor; keep all other arguments the same
new_config = m.cfg.preprocessor
with open_dict(new_config):
new_config.use_torchaudio = True
# Instantiate an instance that uses torchaudio on the backend
new_preprocessor: AudioToMelSpectrogramPreprocessor = hydra.utils.instantiate(config=new_config)
new_preprocessor.eval()
print(f"New preprocessor featurizer type: {type(new_preprocessor.featurizer)}")
# Export the torchaudio preprocessor and load it back in as a `ScriptModule`.
new_preprocessor.export("tmp.pt")
jit_preprocessor = torch.jit.load("tmp.pt")
# Generate random input
batch_size = 4
max_length = 16000
signals = torch.randn(size=[batch_size, max_length])
lengths = torch.randint(low=200, high=max_length, size=[batch_size])
lengths[0] = max_length
# Extract features with all preprocessors
old_feats, old_feat_lens = old_preprocessor(input_signal=signals, length=lengths)
new_feats, new_feat_lens = new_preprocessor(input_signal=signals, length=lengths)
jit_feats, jit_feat_lens = jit_preprocessor(input_signal=signals, length=lengths)
# Make sure new output matches old output
# Need to relax the tolerance from defaults. We will check WER also, as an alternative verification of correctness.
rel_tolerance = 1e-2
abs_tolerance = 1e-4
torch.testing.assert_allclose(actual=new_feats, expected=old_feats, atol=abs_tolerance, rtol=rel_tolerance)
# Zero tolerance for integer lengths.
torch.testing.assert_allclose(actual=new_feat_lens, expected=old_feat_lens, atol=0, rtol=0)
print(f"Output comparison passed with relative tolerance {rel_tolerance} and absolute tolerance {abs_tolerance}.")
# Make sure JIT output matches PyTorch output
# Keep tolerance at defaults for JIT comparison
torch.testing.assert_allclose(actual=jit_feats, expected=new_feats)
torch.testing.assert_allclose(actual=jit_feat_lens, expected=new_feat_lens, atol=0, rtol=0)
print(f"Jit comparison passed with default tolerance.")
# To run a WER check you'll need to comment out some assumptions in the CTC transcribe method, as addressed in https://github.com/NVIDIA/NeMo/pull/2762
# print("Testing WER with old/new preprocessor with some LibriSpeech data")
# We only need audio files; we're comparing model outputs to each other, not references
# dev_other_dir = "/path/to/LibriSpeech/dev-other"
# num_files_to_use = 100
# audio_files = [str(x) for x in Path(dev_other_dir).rglob("*.flac")]
# audio_files = audio_files[:num_files_to_use]
# print("Transcribing with the baseline model")
# old_output = m.transcribe(audio_files)
# m.preprocessor = new_preprocessor
# print("Transcribing after switching the preprocessor")
# new_output = m.transcribe(audio_files)
# wer = word_error_rate(hypotheses=new_output, references=old_output)
# print(f"WER with {len(audio_files)} audio files using old vs. new preprocessor is {wer}")
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [x] Did you write any new necessary tests?
- [x] Did you add or update any necessary documentation?
- [x] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to: this issue comes up once in a while and is generally brushed aside.
| {
"login": "1-800-BAD-CODE",
"id": 50530592,
"node_id": "MDQ6VXNlcjUwNTMwNTky",
"avatar_url": "https://avatars.githubusercontent.com/u/50530592?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/1-800-BAD-CODE",
"html_url": "https://github.com/1-800-BAD-CODE",
"followers_url": "https://api.github.com/users/1-800-BAD-CODE/followers",
"following_url": "https://api.github.com/users/1-800-BAD-CODE/following{/other_user}",
"gists_url": "https://api.github.com/users/1-800-BAD-CODE/gists{/gist_id}",
"starred_url": "https://api.github.com/users/1-800-BAD-CODE/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/1-800-BAD-CODE/subscriptions",
"organizations_url": "https://api.github.com/users/1-800-BAD-CODE/orgs",
"repos_url": "https://api.github.com/users/1-800-BAD-CODE/repos",
"events_url": "https://api.github.com/users/1-800-BAD-CODE/events{/privacy}",
"received_events_url": "https://api.github.com/users/1-800-BAD-CODE/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5508/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5508/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5505 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5505/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5505/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5505/events | https://github.com/NVIDIA/NeMo/issues/5505 | 1,464,845,232 | I_kwDOC_bI7s5XT8Ow | 5,505 | TTS Speech marks | {
"login": "lumpidu",
"id": 734966,
"node_id": "MDQ6VXNlcjczNDk2Ng==",
"avatar_url": "https://avatars.githubusercontent.com/u/734966?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/lumpidu",
"html_url": "https://github.com/lumpidu",
"followers_url": "https://api.github.com/users/lumpidu/followers",
"following_url": "https://api.github.com/users/lumpidu/following{/other_user}",
"gists_url": "https://api.github.com/users/lumpidu/gists{/gist_id}",
"starred_url": "https://api.github.com/users/lumpidu/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/lumpidu/subscriptions",
"organizations_url": "https://api.github.com/users/lumpidu/orgs",
"repos_url": "https://api.github.com/users/lumpidu/repos",
"events_url": "https://api.github.com/users/lumpidu/events{/privacy}",
"received_events_url": "https://api.github.com/users/lumpidu/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4418253970,
"node_id": "LA_kwDOC_bI7s8AAAABB1k0kg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/TTS",
"name": "TTS",
"color": "49AF28",
"default": false,
"description": ""
}
] | closed | false | {
"login": "redoctopus",
"id": 4268876,
"node_id": "MDQ6VXNlcjQyNjg4NzY=",
"avatar_url": "https://avatars.githubusercontent.com/u/4268876?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/redoctopus",
"html_url": "https://github.com/redoctopus",
"followers_url": "https://api.github.com/users/redoctopus/followers",
"following_url": "https://api.github.com/users/redoctopus/following{/other_user}",
"gists_url": "https://api.github.com/users/redoctopus/gists{/gist_id}",
"starred_url": "https://api.github.com/users/redoctopus/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/redoctopus/subscriptions",
"organizations_url": "https://api.github.com/users/redoctopus/orgs",
"repos_url": "https://api.github.com/users/redoctopus/repos",
"events_url": "https://api.github.com/users/redoctopus/events{/privacy}",
"received_events_url": "https://api.github.com/users/redoctopus/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "redoctopus",
"id": 4268876,
"node_id": "MDQ6VXNlcjQyNjg4NzY=",
"avatar_url": "https://avatars.githubusercontent.com/u/4268876?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/redoctopus",
"html_url": "https://github.com/redoctopus",
"followers_url": "https://api.github.com/users/redoctopus/followers",
"following_url": "https://api.github.com/users/redoctopus/following{/other_user}",
"gists_url": "https://api.github.com/users/redoctopus/gists{/gist_id}",
"starred_url": "https://api.github.com/users/redoctopus/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/redoctopus/subscriptions",
"organizations_url": "https://api.github.com/users/redoctopus/orgs",
"repos_url": "https://api.github.com/users/redoctopus/repos",
"events_url": "https://api.github.com/users/redoctopus/events{/privacy}",
"received_events_url": "https://api.github.com/users/redoctopus/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-11-25T18:49:44 | 2023-01-31T19:24:59 | 2023-01-31T19:24:59 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Is your feature request related to a problem? Please describe.**
In TTS, if one wants to provide timing hints of the currently spoken word to the user, one needs to map the timing information (beginning/end time offset in generated pcm) to a word or characters or phonemes. This is necessary/applicable for e.g. screen readers or language learning apps.
**Describe the solution you'd like**
The acoustic model should be able to output timing information of each spoken word / phoneme related to the input phonemes/words. Ideally, the information could be provided together with the output MEL spectrograms as an extra tensor, so that inferencing doesn't need to be done twice. Alternatively, providing a function, that takes the generated MEL spectrograms and maps these to the aforementioned timing information would also be a very good solution and could possibly be implemented generically for all TTS models that output MEL spectrograms.
**Describe alternatives you've considered**
Alternatively, of course without any lip-synchronization accuracy, one could just devide the overall generated audio time of the synthesized PCM by the number of input characters/phonemes, naively assuming every character/phoneme needs an equal amount of time to speak. Although easily to implement, timing results are however off most of the time.
**Additional context**
Add any other context or screenshots about the feature request here.
| {
"login": "redoctopus",
"id": 4268876,
"node_id": "MDQ6VXNlcjQyNjg4NzY=",
"avatar_url": "https://avatars.githubusercontent.com/u/4268876?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/redoctopus",
"html_url": "https://github.com/redoctopus",
"followers_url": "https://api.github.com/users/redoctopus/followers",
"following_url": "https://api.github.com/users/redoctopus/following{/other_user}",
"gists_url": "https://api.github.com/users/redoctopus/gists{/gist_id}",
"starred_url": "https://api.github.com/users/redoctopus/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/redoctopus/subscriptions",
"organizations_url": "https://api.github.com/users/redoctopus/orgs",
"repos_url": "https://api.github.com/users/redoctopus/repos",
"events_url": "https://api.github.com/users/redoctopus/events{/privacy}",
"received_events_url": "https://api.github.com/users/redoctopus/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5505/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5505/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5504 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5504/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5504/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5504/events | https://github.com/NVIDIA/NeMo/pull/5504 | 1,464,496,001 | PR_kwDOC_bI7s5DtM45 | 5,504 | Add more glob patterns for labeler | {
"login": "SeanNaren",
"id": 6707363,
"node_id": "MDQ6VXNlcjY3MDczNjM=",
"avatar_url": "https://avatars.githubusercontent.com/u/6707363?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SeanNaren",
"html_url": "https://github.com/SeanNaren",
"followers_url": "https://api.github.com/users/SeanNaren/followers",
"following_url": "https://api.github.com/users/SeanNaren/following{/other_user}",
"gists_url": "https://api.github.com/users/SeanNaren/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SeanNaren/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SeanNaren/subscriptions",
"organizations_url": "https://api.github.com/users/SeanNaren/orgs",
"repos_url": "https://api.github.com/users/SeanNaren/repos",
"events_url": "https://api.github.com/users/SeanNaren/events{/privacy}",
"received_events_url": "https://api.github.com/users/SeanNaren/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-11-25T12:41:16 | 2022-11-27T22:14:01 | 2022-11-27T22:13:58 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5504",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5504",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5504.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5504.patch",
"merged_at": "2022-11-27T22:13:58"
} | # What does this PR do ?
Larger coverage for the auto-labeler!
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5504/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5504/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5503 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5503/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5503/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5503/events | https://github.com/NVIDIA/NeMo/pull/5503 | 1,464,487,240 | PR_kwDOC_bI7s5DtLAO | 5,503 | See if auto labeler labels | {
"login": "SeanNaren",
"id": 6707363,
"node_id": "MDQ6VXNlcjY3MDczNjM=",
"avatar_url": "https://avatars.githubusercontent.com/u/6707363?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SeanNaren",
"html_url": "https://github.com/SeanNaren",
"followers_url": "https://api.github.com/users/SeanNaren/followers",
"following_url": "https://api.github.com/users/SeanNaren/following{/other_user}",
"gists_url": "https://api.github.com/users/SeanNaren/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SeanNaren/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SeanNaren/subscriptions",
"organizations_url": "https://api.github.com/users/SeanNaren/orgs",
"repos_url": "https://api.github.com/users/SeanNaren/repos",
"events_url": "https://api.github.com/users/SeanNaren/events{/privacy}",
"received_events_url": "https://api.github.com/users/SeanNaren/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
},
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-25T12:33:25 | 2022-11-25T16:14:50 | 2022-11-25T16:14:47 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | true | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5503",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5503",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5503.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5503.patch",
"merged_at": null
} | Signed-off-by: SeanNaren <[email protected]>
# What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "SeanNaren",
"id": 6707363,
"node_id": "MDQ6VXNlcjY3MDczNjM=",
"avatar_url": "https://avatars.githubusercontent.com/u/6707363?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SeanNaren",
"html_url": "https://github.com/SeanNaren",
"followers_url": "https://api.github.com/users/SeanNaren/followers",
"following_url": "https://api.github.com/users/SeanNaren/following{/other_user}",
"gists_url": "https://api.github.com/users/SeanNaren/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SeanNaren/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SeanNaren/subscriptions",
"organizations_url": "https://api.github.com/users/SeanNaren/orgs",
"repos_url": "https://api.github.com/users/SeanNaren/repos",
"events_url": "https://api.github.com/users/SeanNaren/events{/privacy}",
"received_events_url": "https://api.github.com/users/SeanNaren/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5503/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5503/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5502 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5502/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5502/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5502/events | https://github.com/NVIDIA/NeMo/issues/5502 | 1,464,044,396 | I_kwDOC_bI7s5XQ4ts | 5,502 | [ASR] predictions during training by Conformer-CTC seems not improve. | {
"login": "triumph9989",
"id": 48488138,
"node_id": "MDQ6VXNlcjQ4NDg4MTM4",
"avatar_url": "https://avatars.githubusercontent.com/u/48488138?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/triumph9989",
"html_url": "https://github.com/triumph9989",
"followers_url": "https://api.github.com/users/triumph9989/followers",
"following_url": "https://api.github.com/users/triumph9989/following{/other_user}",
"gists_url": "https://api.github.com/users/triumph9989/gists{/gist_id}",
"starred_url": "https://api.github.com/users/triumph9989/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/triumph9989/subscriptions",
"organizations_url": "https://api.github.com/users/triumph9989/orgs",
"repos_url": "https://api.github.com/users/triumph9989/repos",
"events_url": "https://api.github.com/users/triumph9989/events{/privacy}",
"received_events_url": "https://api.github.com/users/triumph9989/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
}
] | closed | false | null | [] | null | null | 2022-11-25T05:11:07 | 2022-12-01T02:36:17 | 2022-12-01T02:36:16 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
Hi,
I would like to use Conformer-CTC to train a Mandarin ASR model but even reach the 29th epoch, the predictions seem not to improve.
Meanwhile, another one uses QuartzNet15x5 looks good.
Is there something I have not noticed, please?
**Parts of prediction results**
```
Epoch 29: 100%|████████████████▉| 4198/4202 [21:52<00:01, 3.20it/s, loss=99.1, v_num=7-38][NeMo I 2022-11-25 03:17:08 wer:984]
dation DataLoader 0: 99%|██████████████████████████▊| 444/448 [00:55<00:00, 8.04it/s]
[NeMo I 2022-11-25 03:17:08 wer:985] reference:由於 身體 並 沒 有 處 於 最 佳 狀態
[NeMo I 2022-11-25 03:17:08 wer:986] predicted:但 道
Epoch 29: 100%|████████████████▉| 4199/4202 [21:52<00:00, 3.20it/s, loss=99.1, v_num=7-38][NeMo I 2022-11-25 03:17:08 wer:984]
dation DataLoader 0: 99%|██████████████████████████▊| 445/448 [00:55<00:00, 8.04it/s]
[NeMo I 2022-11-25 03:17:08 wer:985] reference:但 近日 身體 檢查 終於 還是 敵 不過 病魔 折磨
[NeMo I 2022-11-25 03:17:08 wer:986] predicted:中 元
Epoch 29: 100%|████████████████▉| 4200/4202 [21:53<00:00, 3.20it/s, loss=99.1, v_num=7-38][NeMo I 2022-11-25 03:17:09 wer:984]
dation DataLoader 0: 100%|██████████████████████████▉| 446/448 [00:55<00:00, 8.05it/s]
[NeMo I 2022-11-25 03:17:09 wer:985] reference:遠赴 廣西 連續 搗毀 兩 個 電信 詐騙 犯罪 團夥
[NeMo I 2022-11-25 03:17:09 wer:986] predicted:中 元
Epoch 29: 100%|████████████████▉| 4201/4202 [21:53<00:00, 3.20it/s, loss=99.1, v_num=7-38][NeMo I 2022-11-25 03:17:09 wer:984]
dation DataLoader 0: 100%|██████████████████████████▉| 447/448 [00:55<00:00, 8.04it/s]
[NeMo I 2022-11-25 03:17:09 wer:985] reference:自己 不 能 總 在 孩子 面前 打 遊戲 刷 微 博 微信
[NeMo I 2022-11-25 03:17:09 wer:986] predicted:中 元
```
**Steps/Code to reproduce bug**
`CUDA_VISIBLE_DEVICES=0,1 python speech_to_text_ctc.py`
**conformer_ctc_char.yaml**
```
# It contains the default values for training a Conformer-CTC ASR model, large size (~120M) with CTC loss and char-based vocabulary.
# Char-based encoding may give lower accuracy than sub-word encoding (conformer_ctc_bpe.yaml).
# You may find more detail on Conformer-CTC at `examples/asr/conf/conformer/conformer_ctc_bpe.yaml`
name: "Conformer-CTC-Char-Aishell-30ep-lr2.0"
model:
sample_rate: 16000
labels: [" ", "a", "b", "c", "d", "e", "f", "g", "h", "i", "j", "k", "l", "m",
"n", "o", "p", "q", "r", "s", "t", "u", "v", "w", "x", "y", "z", "'"]
log_prediction: true # enables logging sample predictions in the output during training
ctc_reduction: 'mean_batch'
skip_nan_grad: false
train_ds:
manifest_filepath: ???
labels: ${model.labels}
sample_rate: ${model.sample_rate}
batch_size: 16 # you may increase batch_size if your memory allows
shuffle: true
num_workers: 8
pin_memory: true
trim_silence: false
max_duration: 16.7 # it is set for LibriSpeech, you may need to update it for your dataset
min_duration: 0.1
# tarred datasets
is_tarred: false
tarred_audio_filepaths: null
shuffle_n: 2048
# bucketing params
bucketing_strategy: "synced_randomized"
bucketing_batch_size: null
validation_ds:
manifest_filepath: ???
labels: ${model.labels}
sample_rate: ${model.sample_rate}
batch_size: 16 # you may increase batch_size if your memory allows
shuffle: false
num_workers: 8
pin_memory: true
test_ds:
manifest_filepath: null
labels: ${model.labels}
sample_rate: ${model.sample_rate}
batch_size: 16 # you may increase batch_size if your memory allows
shuffle: false
num_workers: 8
pin_memory: true
preprocessor:
_target_: nemo.collections.asr.modules.AudioToMelSpectrogramPreprocessor
sample_rate: ${model.sample_rate}
normalize: "per_feature"
window_size: 0.025
window_stride: 0.01
window: "hann"
features: 80
n_fft: 512
log: true
frame_splicing: 1
dither: 0.00001
pad_to: 0
pad_value: 0.0
spec_augment:
_target_: nemo.collections.asr.modules.SpectrogramAugmentation
freq_masks: 2 # set to zero to disable it
# you may use lower time_masks for smaller models to have a faster convergence
time_masks: 10 # set to zero to disable it
freq_width: 27
time_width: 0.05
encoder:
_target_: nemo.collections.asr.modules.ConformerEncoder
feat_in: ${model.preprocessor.features}
feat_out: -1 # you may set it if you need different output size other than the default d_model
n_layers: 16
d_model: 256
# Sub-sampling params
subsampling: striding # vggnet, striding, stacking or stacking_norm, dw_striding
subsampling_factor: 4 # must be power of 2 for striding and vggnet
subsampling_conv_channels: -1 # set to -1 to make it equal to the d_model
causal_downsampling: false
# Feed forward module's params
ff_expansion_factor: 4
# Multi-headed Attention Module's params
self_attention_model: rel_pos # rel_pos or abs_pos
n_heads: 8 # may need to be lower for smaller d_models
# [left, right] specifies the number of steps to be seen from left and right of each step in self-attention
att_context_size: [-1, -1] # -1 means unlimited context
att_context_style: regular # regular or chunked_limited
xscaling: true # scales up the input embeddings by sqrt(d_model)
untie_biases: true # unties the biases of the TransformerXL layers
pos_emb_max_len: 5000
# Convolution module's params
conv_kernel_size: 31
conv_norm_type: 'batch_norm' # batch_norm or layer_norm or groupnormN (N specifies the number of groups)
# conv_context_size can be"causal" or a list of two integers while conv_context_size[0]+conv_context_size[1]+1==conv_kernel_size
# null means [(kernel_size-1)//2, (kernel_size-1)//2], and 'causal' means [(kernel_size-1), 0]
conv_context_size: null
### regularization
dropout: 0.1 # The dropout used in most of the Conformer Modules
dropout_pre_encoder: 0.1 # The dropout used before the encoder
dropout_emb: 0.0 # The dropout used for embeddings
dropout_att: 0.1 # The dropout for multi-headed attention modules
decoder:
_target_: nemo.collections.asr.modules.ConvASRDecoder
feat_in: null
num_classes: -1
vocabulary: ${model.labels}
optim:
name: adamw
lr: 3 #0.01 很慢
# optimizer arguments
betas: [0.9, 0.98]
# less necessity for weight_decay as we already have large augmentations with SpecAug
# you may need weight_decay for large models, stable AMP training, small datasets, or when lower augmentations are used
# weight decay of 0.0 with lr of 2.0 also works fine
weight_decay: 1e-3
# scheduler setup
sched:
name: NoamAnnealing
d_model: ${model.encoder.d_model}
# scheduler config override
warmup_steps: 10000
warmup_ratio: null
min_lr: 1e-6
trainer:
devices: 2 # number of GPUs, -1 would use all available GPUs
num_nodes: 1
max_epochs: 30
max_steps: -1 # computed at runtime if not set
val_check_interval: 1.0 # Set to 0.25 to check 4 times per epoch, or an int for number of iterations
accelerator: gpu
strategy: ddp
accumulate_grad_batches: 1
gradient_clip_val: 0.0
precision: 32 # Should be set to 16 for O1 and O2 to enable the AMP.
log_every_n_steps: 10 # Interval of logging.
enable_progress_bar: True
resume_from_checkpoint: null # The path to a checkpoint file to continue the training, restores the whole state including the epoch, step, LR schedulers, apex, etc.
num_sanity_val_steps: 0 # number of steps to perform validation steps for sanity check the validation process before starting the training, setting to 0 disables it
check_val_every_n_epoch: 1 # number of evaluations on validation every n epochs
sync_batchnorm: true
enable_checkpointing: False # Provided by exp_manager
logger: false # Provided by exp_manager
benchmark: false # needs to be false for models with variable-length speech input as it slows down training
exp_manager:
exp_dir: ../exp
name: ${name}
create_tensorboard_logger: true
create_checkpoint_callback: true
checkpoint_callback_params:
# in case of multiple validation sets, first one is used
monitor: "val_wer"
mode: "min"
save_top_k: 5
always_save_nemo: True # saves the checkpoints as nemo files instead of PTL checkpoints
# you need to set these two to True to continue the training
resume_if_exists: false
resume_ignore_no_checkpoint: false
# You may use this section to create a W&B logger
create_wandb_logger: false
wandb_logger_kwargs:
name: null
project: null
```
**speech_to_text_ctc.py**
```
from curses import use_default_colors
import pytorch_lightning as pl
from omegaconf import OmegaConf
from nemo.collections.asr.models import EncDecCTCModel
from nemo.core.config import hydra_runner
from nemo.utils import logging
from nemo.utils.exp_manager import exp_manager
use_cer = True
log_prediction = True
manifest_path = "/home/face/NeMo/examples/asr/"
train_manifest = f"{manifest_path}train_manifest.json"
dev_manifest = f"{manifest_path}dev_manifest.json"
test_manifest = f"{manifest_path}test_manifest.json"
charset = []
#將 vocab.txt 讀入 list 中，使其能寫入 config 的 labels
with open(f"{manifest_path}Aishell_vocab/train_dev.txt ", "r") as vocab_read:
for char in vocab_read.read():
charset.append(char) #EX: ['余', '耍', '汞', '圳']
# print(charset)
# print(len(charset))
@hydra_runner(config_path="../conf/conformer", config_name="conformer_ctc_char")
def main(cfg):
cfg.model.labels = charset
cfg.model.train_ds.manifest_filepath = train_manifest
cfg.model.validation_ds.manifest_filepath = dev_manifest
cfg.model.test_ds.manifest_filepath = test_manifest
logging.info(f'Hydra config: {OmegaConf.to_yaml(cfg)}')
trainer = pl.Trainer(**cfg.trainer)
exp_manager(trainer, cfg.get("exp_manager", None))
asr_model = EncDecCTCModel(cfg=cfg.model, trainer=trainer)
asr_model._wer.use_cer = use_cer
asr_model._wer.log_prediction = log_prediction
# Initialize the weights of the model from another model, if provided via config
asr_model.maybe_init_from_pretrained_checkpoint(cfg)
trainer.fit(asr_model)
if hasattr(cfg.model, 'test_ds') and cfg.model.test_ds.manifest_filepath is not None:
if asr_model.prepare_test(trainer):
trainer.test(asr_model)
if __name__ == '__main__':
main() # noqa pylint: disable=no-value-for-parameter
```
**Environment overview (please complete the following information)**
- Environment location: Bare-metal
- Method of NeMo install: From source
**Environment details**
-nemo-tookit
1.12.0
- OS version
Linux version 4.15.0-142-generic (buildd@lgw01-amd64-039) (gcc version 5.4.0 20160609 (Ubuntu 5.4.0-6ubuntu1~16.04.12)) #146~16.04.1-Ubuntu SMP Tue Apr 13 09:27:15 UTC 2021
- PyTorch version
1.13.0
- Python version
3.8
**Additional context**
Add any other context about the problem here.
GPU: NVIDIA GeForce 2080TI
| {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5502/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5502/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5500 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5500/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5500/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5500/events | https://github.com/NVIDIA/NeMo/issues/5500 | 1,462,693,484 | I_kwDOC_bI7s5XLu5s | 5,500 | help | {
"login": "Pydataman",
"id": 17594431,
"node_id": "MDQ6VXNlcjE3NTk0NDMx",
"avatar_url": "https://avatars.githubusercontent.com/u/17594431?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Pydataman",
"html_url": "https://github.com/Pydataman",
"followers_url": "https://api.github.com/users/Pydataman/followers",
"following_url": "https://api.github.com/users/Pydataman/following{/other_user}",
"gists_url": "https://api.github.com/users/Pydataman/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Pydataman/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Pydataman/subscriptions",
"organizations_url": "https://api.github.com/users/Pydataman/orgs",
"repos_url": "https://api.github.com/users/Pydataman/repos",
"events_url": "https://api.github.com/users/Pydataman/events{/privacy}",
"received_events_url": "https://api.github.com/users/Pydataman/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-11-24T03:18:11 | 2023-02-16T04:06:45 | 2023-02-16T04:06:45 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | hello，I reproduced this paper （ADAPTER-BASED EXTENSION OF MULTI-SPEAKER TEXT-TO-SPEECH MODEL FOR NEW SPEAKERS），but the result is not as good as the paper， very poor。I only finetune adapter，please ask me where is the paper code？ | {
"login": "ericharper",
"id": 11999610,
"node_id": "MDQ6VXNlcjExOTk5NjEw",
"avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ericharper",
"html_url": "https://github.com/ericharper",
"followers_url": "https://api.github.com/users/ericharper/followers",
"following_url": "https://api.github.com/users/ericharper/following{/other_user}",
"gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ericharper/subscriptions",
"organizations_url": "https://api.github.com/users/ericharper/orgs",
"repos_url": "https://api.github.com/users/ericharper/repos",
"events_url": "https://api.github.com/users/ericharper/events{/privacy}",
"received_events_url": "https://api.github.com/users/ericharper/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5500/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5500/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5499 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5499/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5499/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5499/events | https://github.com/NVIDIA/NeMo/pull/5499 | 1,462,612,662 | PR_kwDOC_bI7s5Dm6k3 | 5,499 | [TTS] Add adapter for new speaker adaptation | {
"login": "subhankar-ghosh",
"id": 23703547,
"node_id": "MDQ6VXNlcjIzNzAzNTQ3",
"avatar_url": "https://avatars.githubusercontent.com/u/23703547?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/subhankar-ghosh",
"html_url": "https://github.com/subhankar-ghosh",
"followers_url": "https://api.github.com/users/subhankar-ghosh/followers",
"following_url": "https://api.github.com/users/subhankar-ghosh/following{/other_user}",
"gists_url": "https://api.github.com/users/subhankar-ghosh/gists{/gist_id}",
"starred_url": "https://api.github.com/users/subhankar-ghosh/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/subhankar-ghosh/subscriptions",
"organizations_url": "https://api.github.com/users/subhankar-ghosh/orgs",
"repos_url": "https://api.github.com/users/subhankar-ghosh/repos",
"events_url": "https://api.github.com/users/subhankar-ghosh/events{/privacy}",
"received_events_url": "https://api.github.com/users/subhankar-ghosh/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4418253970,
"node_id": "LA_kwDOC_bI7s8AAAABB1k0kg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/TTS",
"name": "TTS",
"color": "49AF28",
"default": false,
"description": ""
},
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | null | [] | null | null | 2022-11-24T01:02:28 | 2023-01-11T01:59:06 | 2023-01-11T01:59:06 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5499",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5499",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5499.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5499.patch",
"merged_at": null
} | # What does this PR do ?
Adds adapter support for new speaker adaptation. The following are the contributions of this PR:
1) Adds adapter support to TTS (FastPitch) pipeline for new speaker adaptation.
2) Adds Global Style Token based speaker representation to multi-speaker TTS pipeline.
3) Adds Speaker Verification based speaker representation to multi-speaker TTS pipeline.
**Collection**: [tts]
# Changelog
- `examples/tts/conf/fastpitch_speaker_adaptation.yaml` - A new config file that is used for adapter training or using GST speaker representation.
- `examples/tts/fastpitch_finetune_adapters.py` - Script to Finetune adapters for new speaker adaptation. This script edits the config to add adapters, freeze the pretrained model and trains just the adapters.
- `nemo/collections/tts/parts/mixins/fastpitch_adapter_mixins.py` - This is a new file that contains all adapter related methods for FastPitch model. `parts/mixins` follows ASR folder structure.
- `nemo/collections/tts/modules/aligner.py, fastpitch.py, transformer.py` - the changes in these files are corresponding to inserting adapters to each of these modules.
- `nemo/collections/tts/modules/submodules.py` - GST based speaker representation module's implementation is in this file.
- `nemo/collections/tts/torch/tts_data_types.py, data.py` - GST related changes made to dataloader. GST needs a reference audio for a particular speaker. So a new datatype `GSTRefAudio` is introduced, this datatype refers to the reference audio.
- `tutorials/tts/FastPitch_Speaker_Adaptation.ipynb` - Tutorial explaining step by step how to use adapters to adapt a new speaker in FastPitch.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5499/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5499/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5498 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5498/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5498/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5498/events | https://github.com/NVIDIA/NeMo/pull/5498 | 1,462,555,554 | PR_kwDOC_bI7s5DmuSI | 5,498 | Add auto-labeler | {
"login": "SeanNaren",
"id": 6707363,
"node_id": "MDQ6VXNlcjY3MDczNjM=",
"avatar_url": "https://avatars.githubusercontent.com/u/6707363?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SeanNaren",
"html_url": "https://github.com/SeanNaren",
"followers_url": "https://api.github.com/users/SeanNaren/followers",
"following_url": "https://api.github.com/users/SeanNaren/following{/other_user}",
"gists_url": "https://api.github.com/users/SeanNaren/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SeanNaren/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SeanNaren/subscriptions",
"organizations_url": "https://api.github.com/users/SeanNaren/orgs",
"repos_url": "https://api.github.com/users/SeanNaren/repos",
"events_url": "https://api.github.com/users/SeanNaren/events{/privacy}",
"received_events_url": "https://api.github.com/users/SeanNaren/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-11-23T23:31:21 | 2022-11-25T04:43:42 | 2022-11-25T04:43:38 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5498",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5498",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5498.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5498.patch",
"merged_at": "2022-11-25T04:43:38"
} | # What does this PR do ?
Test to see if adding this action will allow us to auto-label based on the files changed.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5498/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5498/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5497 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5497/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5497/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5497/events | https://github.com/NVIDIA/NeMo/pull/5497 | 1,462,547,189 | PR_kwDOC_bI7s5DmsfQ | 5,497 | Export fixes for Riva | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4556669155,
"node_id": "LA_kwDOC_bI7s8AAAABD5lA4w",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/cherry-pick",
"name": "cherry-pick",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | null | [] | null | null | 2022-11-23T23:20:33 | 2023-02-21T08:28:59 | 2022-11-24T01:27:34 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5497",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5497",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5497.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5497.patch",
"merged_at": "2022-11-24T01:27:34"
} | Signed-off-by: Boris Fomitchev <[email protected]>
# What does this PR do ?
Fixes to make Riva networks export. Most of those are actually already in main. | {
"login": "borisfom",
"id": 14189615,
"node_id": "MDQ6VXNlcjE0MTg5NjE1",
"avatar_url": "https://avatars.githubusercontent.com/u/14189615?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/borisfom",
"html_url": "https://github.com/borisfom",
"followers_url": "https://api.github.com/users/borisfom/followers",
"following_url": "https://api.github.com/users/borisfom/following{/other_user}",
"gists_url": "https://api.github.com/users/borisfom/gists{/gist_id}",
"starred_url": "https://api.github.com/users/borisfom/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/borisfom/subscriptions",
"organizations_url": "https://api.github.com/users/borisfom/orgs",
"repos_url": "https://api.github.com/users/borisfom/repos",
"events_url": "https://api.github.com/users/borisfom/events{/privacy}",
"received_events_url": "https://api.github.com/users/borisfom/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5497/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5497/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5496 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5496/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5496/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5496/events | https://github.com/NVIDIA/NeMo/pull/5496 | 1,462,466,378 | PR_kwDOC_bI7s5DmaRj | 5,496 | Export fixes for Riva | {
"login": "borisfom",
"id": 14189615,
"node_id": "MDQ6VXNlcjE0MTg5NjE1",
"avatar_url": "https://avatars.githubusercontent.com/u/14189615?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/borisfom",
"html_url": "https://github.com/borisfom",
"followers_url": "https://api.github.com/users/borisfom/followers",
"following_url": "https://api.github.com/users/borisfom/following{/other_user}",
"gists_url": "https://api.github.com/users/borisfom/gists{/gist_id}",
"starred_url": "https://api.github.com/users/borisfom/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/borisfom/subscriptions",
"organizations_url": "https://api.github.com/users/borisfom/orgs",
"repos_url": "https://api.github.com/users/borisfom/repos",
"events_url": "https://api.github.com/users/borisfom/events{/privacy}",
"received_events_url": "https://api.github.com/users/borisfom/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-11-23T22:06:35 | 2022-11-23T23:19:58 | 2022-11-23T23:19:58 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5496",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5496",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5496.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5496.patch",
"merged_at": "2022-11-23T23:19:58"
} | Signed-off-by: Boris Fomitchev <[email protected]>
# What does this PR do ?
Fixes to make Riva networks export. Most of those are actually already in main. | {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5496/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5496/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5495 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5495/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5495/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5495/events | https://github.com/NVIDIA/NeMo/pull/5495 | 1,462,412,904 | PR_kwDOC_bI7s5DmOhe | 5,495 | Draft: Add activations_checkpointing_granularity to Bert models. | {
"login": "markelsanz14",
"id": 10426497,
"node_id": "MDQ6VXNlcjEwNDI2NDk3",
"avatar_url": "https://avatars.githubusercontent.com/u/10426497?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/markelsanz14",
"html_url": "https://github.com/markelsanz14",
"followers_url": "https://api.github.com/users/markelsanz14/followers",
"following_url": "https://api.github.com/users/markelsanz14/following{/other_user}",
"gists_url": "https://api.github.com/users/markelsanz14/gists{/gist_id}",
"starred_url": "https://api.github.com/users/markelsanz14/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/markelsanz14/subscriptions",
"organizations_url": "https://api.github.com/users/markelsanz14/orgs",
"repos_url": "https://api.github.com/users/markelsanz14/repos",
"events_url": "https://api.github.com/users/markelsanz14/events{/privacy}",
"received_events_url": "https://api.github.com/users/markelsanz14/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-11-23T21:12:38 | 2022-11-28T20:17:46 | 2022-11-23T22:28:02 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5495",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5495",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5495.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5495.patch",
"merged_at": null
} | # What does this PR do ?
Adds activations_checkpoint_granularity to BERT models, which is now required to run any sort of activations checkpointing.
**Collection**: NLP
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "markelsanz14",
"id": 10426497,
"node_id": "MDQ6VXNlcjEwNDI2NDk3",
"avatar_url": "https://avatars.githubusercontent.com/u/10426497?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/markelsanz14",
"html_url": "https://github.com/markelsanz14",
"followers_url": "https://api.github.com/users/markelsanz14/followers",
"following_url": "https://api.github.com/users/markelsanz14/following{/other_user}",
"gists_url": "https://api.github.com/users/markelsanz14/gists{/gist_id}",
"starred_url": "https://api.github.com/users/markelsanz14/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/markelsanz14/subscriptions",
"organizations_url": "https://api.github.com/users/markelsanz14/orgs",
"repos_url": "https://api.github.com/users/markelsanz14/repos",
"events_url": "https://api.github.com/users/markelsanz14/events{/privacy}",
"received_events_url": "https://api.github.com/users/markelsanz14/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5495/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5495/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5494 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5494/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5494/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5494/events | https://github.com/NVIDIA/NeMo/pull/5494 | 1,462,354,652 | PR_kwDOC_bI7s5DmBrp | 5,494 | Bert sequence parallel support | {
"login": "shanmugamr1992",
"id": 111910568,
"node_id": "U_kgDOBqueqA",
"avatar_url": "https://avatars.githubusercontent.com/u/111910568?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/shanmugamr1992",
"html_url": "https://github.com/shanmugamr1992",
"followers_url": "https://api.github.com/users/shanmugamr1992/followers",
"following_url": "https://api.github.com/users/shanmugamr1992/following{/other_user}",
"gists_url": "https://api.github.com/users/shanmugamr1992/gists{/gist_id}",
"starred_url": "https://api.github.com/users/shanmugamr1992/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/shanmugamr1992/subscriptions",
"organizations_url": "https://api.github.com/users/shanmugamr1992/orgs",
"repos_url": "https://api.github.com/users/shanmugamr1992/repos",
"events_url": "https://api.github.com/users/shanmugamr1992/events{/privacy}",
"received_events_url": "https://api.github.com/users/shanmugamr1992/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
},
{
"id": 4840216855,
"node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI",
"name": "CI",
"color": "1D76DB",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-23T20:18:20 | 2022-12-05T19:15:43 | 2022-12-05T19:13:17 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5494",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5494",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5494.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5494.patch",
"merged_at": "2022-12-05T19:13:17"
} | # What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "shanmugamr1992",
"id": 111910568,
"node_id": "U_kgDOBqueqA",
"avatar_url": "https://avatars.githubusercontent.com/u/111910568?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/shanmugamr1992",
"html_url": "https://github.com/shanmugamr1992",
"followers_url": "https://api.github.com/users/shanmugamr1992/followers",
"following_url": "https://api.github.com/users/shanmugamr1992/following{/other_user}",
"gists_url": "https://api.github.com/users/shanmugamr1992/gists{/gist_id}",
"starred_url": "https://api.github.com/users/shanmugamr1992/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/shanmugamr1992/subscriptions",
"organizations_url": "https://api.github.com/users/shanmugamr1992/orgs",
"repos_url": "https://api.github.com/users/shanmugamr1992/repos",
"events_url": "https://api.github.com/users/shanmugamr1992/events{/privacy}",
"received_events_url": "https://api.github.com/users/shanmugamr1992/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5494/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5494/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5493 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5493/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5493/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5493/events | https://github.com/NVIDIA/NeMo/pull/5493 | 1,462,323,873 | PR_kwDOC_bI7s5Dl62Y | 5,493 | Bug fix/gpt | {
"login": "shanmugamr1992",
"id": 111910568,
"node_id": "U_kgDOBqueqA",
"avatar_url": "https://avatars.githubusercontent.com/u/111910568?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/shanmugamr1992",
"html_url": "https://github.com/shanmugamr1992",
"followers_url": "https://api.github.com/users/shanmugamr1992/followers",
"following_url": "https://api.github.com/users/shanmugamr1992/following{/other_user}",
"gists_url": "https://api.github.com/users/shanmugamr1992/gists{/gist_id}",
"starred_url": "https://api.github.com/users/shanmugamr1992/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/shanmugamr1992/subscriptions",
"organizations_url": "https://api.github.com/users/shanmugamr1992/orgs",
"repos_url": "https://api.github.com/users/shanmugamr1992/repos",
"events_url": "https://api.github.com/users/shanmugamr1992/events{/privacy}",
"received_events_url": "https://api.github.com/users/shanmugamr1992/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-11-23T19:53:26 | 2022-11-23T21:49:37 | 2022-11-23T21:49:23 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5493",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5493",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5493.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5493.patch",
"merged_at": "2022-11-23T21:49:23"
} | # What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "shanmugamr1992",
"id": 111910568,
"node_id": "U_kgDOBqueqA",
"avatar_url": "https://avatars.githubusercontent.com/u/111910568?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/shanmugamr1992",
"html_url": "https://github.com/shanmugamr1992",
"followers_url": "https://api.github.com/users/shanmugamr1992/followers",
"following_url": "https://api.github.com/users/shanmugamr1992/following{/other_user}",
"gists_url": "https://api.github.com/users/shanmugamr1992/gists{/gist_id}",
"starred_url": "https://api.github.com/users/shanmugamr1992/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/shanmugamr1992/subscriptions",
"organizations_url": "https://api.github.com/users/shanmugamr1992/orgs",
"repos_url": "https://api.github.com/users/shanmugamr1992/repos",
"events_url": "https://api.github.com/users/shanmugamr1992/events{/privacy}",
"received_events_url": "https://api.github.com/users/shanmugamr1992/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5493/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5493/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5490 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5490/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5490/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5490/events | https://github.com/NVIDIA/NeMo/pull/5490 | 1,461,808,005 | PR_kwDOC_bI7s5DkKT1 | 5,490 | Suggest users to add GitHub labels to PRs | {
"login": "SeanNaren",
"id": 6707363,
"node_id": "MDQ6VXNlcjY3MDczNjM=",
"avatar_url": "https://avatars.githubusercontent.com/u/6707363?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SeanNaren",
"html_url": "https://github.com/SeanNaren",
"followers_url": "https://api.github.com/users/SeanNaren/followers",
"following_url": "https://api.github.com/users/SeanNaren/following{/other_user}",
"gists_url": "https://api.github.com/users/SeanNaren/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SeanNaren/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SeanNaren/subscriptions",
"organizations_url": "https://api.github.com/users/SeanNaren/orgs",
"repos_url": "https://api.github.com/users/SeanNaren/repos",
"events_url": "https://api.github.com/users/SeanNaren/events{/privacy}",
"received_events_url": "https://api.github.com/users/SeanNaren/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | {
"login": "SeanNaren",
"id": 6707363,
"node_id": "MDQ6VXNlcjY3MDczNjM=",
"avatar_url": "https://avatars.githubusercontent.com/u/6707363?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SeanNaren",
"html_url": "https://github.com/SeanNaren",
"followers_url": "https://api.github.com/users/SeanNaren/followers",
"following_url": "https://api.github.com/users/SeanNaren/following{/other_user}",
"gists_url": "https://api.github.com/users/SeanNaren/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SeanNaren/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SeanNaren/subscriptions",
"organizations_url": "https://api.github.com/users/SeanNaren/orgs",
"repos_url": "https://api.github.com/users/SeanNaren/repos",
"events_url": "https://api.github.com/users/SeanNaren/events{/privacy}",
"received_events_url": "https://api.github.com/users/SeanNaren/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "SeanNaren",
"id": 6707363,
"node_id": "MDQ6VXNlcjY3MDczNjM=",
"avatar_url": "https://avatars.githubusercontent.com/u/6707363?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SeanNaren",
"html_url": "https://github.com/SeanNaren",
"followers_url": "https://api.github.com/users/SeanNaren/followers",
"following_url": "https://api.github.com/users/SeanNaren/following{/other_user}",
"gists_url": "https://api.github.com/users/SeanNaren/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SeanNaren/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SeanNaren/subscriptions",
"organizations_url": "https://api.github.com/users/SeanNaren/orgs",
"repos_url": "https://api.github.com/users/SeanNaren/repos",
"events_url": "https://api.github.com/users/SeanNaren/events{/privacy}",
"received_events_url": "https://api.github.com/users/SeanNaren/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-11-23T13:55:19 | 2022-11-23T23:24:27 | 2022-11-23T23:24:23 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5490",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5490",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5490.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5490.patch",
"merged_at": null
} | # What does this PR do ?
Suggests users to add labels to their PR. This helps classify for maintainers.
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "SeanNaren",
"id": 6707363,
"node_id": "MDQ6VXNlcjY3MDczNjM=",
"avatar_url": "https://avatars.githubusercontent.com/u/6707363?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SeanNaren",
"html_url": "https://github.com/SeanNaren",
"followers_url": "https://api.github.com/users/SeanNaren/followers",
"following_url": "https://api.github.com/users/SeanNaren/following{/other_user}",
"gists_url": "https://api.github.com/users/SeanNaren/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SeanNaren/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SeanNaren/subscriptions",
"organizations_url": "https://api.github.com/users/SeanNaren/orgs",
"repos_url": "https://api.github.com/users/SeanNaren/repos",
"events_url": "https://api.github.com/users/SeanNaren/events{/privacy}",
"received_events_url": "https://api.github.com/users/SeanNaren/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5490/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5490/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5489 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5489/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5489/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5489/events | https://github.com/NVIDIA/NeMo/pull/5489 | 1,461,762,166 | PR_kwDOC_bI7s5DkAVz | 5,489 | fixed missing fields | {
"login": "Jorjeous",
"id": 37293288,
"node_id": "MDQ6VXNlcjM3MjkzMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/37293288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Jorjeous",
"html_url": "https://github.com/Jorjeous",
"followers_url": "https://api.github.com/users/Jorjeous/followers",
"following_url": "https://api.github.com/users/Jorjeous/following{/other_user}",
"gists_url": "https://api.github.com/users/Jorjeous/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Jorjeous/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Jorjeous/subscriptions",
"organizations_url": "https://api.github.com/users/Jorjeous/orgs",
"repos_url": "https://api.github.com/users/Jorjeous/repos",
"events_url": "https://api.github.com/users/Jorjeous/events{/privacy}",
"received_events_url": "https://api.github.com/users/Jorjeous/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
}
] | closed | false | {
"login": "Jorjeous",
"id": 37293288,
"node_id": "MDQ6VXNlcjM3MjkzMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/37293288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Jorjeous",
"html_url": "https://github.com/Jorjeous",
"followers_url": "https://api.github.com/users/Jorjeous/followers",
"following_url": "https://api.github.com/users/Jorjeous/following{/other_user}",
"gists_url": "https://api.github.com/users/Jorjeous/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Jorjeous/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Jorjeous/subscriptions",
"organizations_url": "https://api.github.com/users/Jorjeous/orgs",
"repos_url": "https://api.github.com/users/Jorjeous/repos",
"events_url": "https://api.github.com/users/Jorjeous/events{/privacy}",
"received_events_url": "https://api.github.com/users/Jorjeous/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "Jorjeous",
"id": 37293288,
"node_id": "MDQ6VXNlcjM3MjkzMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/37293288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Jorjeous",
"html_url": "https://github.com/Jorjeous",
"followers_url": "https://api.github.com/users/Jorjeous/followers",
"following_url": "https://api.github.com/users/Jorjeous/following{/other_user}",
"gists_url": "https://api.github.com/users/Jorjeous/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Jorjeous/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Jorjeous/subscriptions",
"organizations_url": "https://api.github.com/users/Jorjeous/orgs",
"repos_url": "https://api.github.com/users/Jorjeous/repos",
"events_url": "https://api.github.com/users/Jorjeous/events{/privacy}",
"received_events_url": "https://api.github.com/users/Jorjeous/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-11-23T13:25:02 | 2022-12-08T13:18:54 | 2022-12-08T13:18:54 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5489",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5489",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5489.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5489.patch",
"merged_at": null
} | # What does this PR do ?
bugfix
Add a one line overview of what this PR aims to accomplish.
**Collection**: [Note which collection this PR will affect]
ASR
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [+] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "Jorjeous",
"id": 37293288,
"node_id": "MDQ6VXNlcjM3MjkzMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/37293288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Jorjeous",
"html_url": "https://github.com/Jorjeous",
"followers_url": "https://api.github.com/users/Jorjeous/followers",
"following_url": "https://api.github.com/users/Jorjeous/following{/other_user}",
"gists_url": "https://api.github.com/users/Jorjeous/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Jorjeous/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Jorjeous/subscriptions",
"organizations_url": "https://api.github.com/users/Jorjeous/orgs",
"repos_url": "https://api.github.com/users/Jorjeous/repos",
"events_url": "https://api.github.com/users/Jorjeous/events{/privacy}",
"received_events_url": "https://api.github.com/users/Jorjeous/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5489/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5489/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5488 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5488/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5488/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5488/events | https://github.com/NVIDIA/NeMo/pull/5488 | 1,461,747,387 | PR_kwDOC_bI7s5Dj9Fn | 5,488 | Sde transcribe 22_11 | {
"login": "Jorjeous",
"id": 37293288,
"node_id": "MDQ6VXNlcjM3MjkzMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/37293288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Jorjeous",
"html_url": "https://github.com/Jorjeous",
"followers_url": "https://api.github.com/users/Jorjeous/followers",
"following_url": "https://api.github.com/users/Jorjeous/following{/other_user}",
"gists_url": "https://api.github.com/users/Jorjeous/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Jorjeous/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Jorjeous/subscriptions",
"organizations_url": "https://api.github.com/users/Jorjeous/orgs",
"repos_url": "https://api.github.com/users/Jorjeous/repos",
"events_url": "https://api.github.com/users/Jorjeous/events{/privacy}",
"received_events_url": "https://api.github.com/users/Jorjeous/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | {
"login": "Jorjeous",
"id": 37293288,
"node_id": "MDQ6VXNlcjM3MjkzMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/37293288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Jorjeous",
"html_url": "https://github.com/Jorjeous",
"followers_url": "https://api.github.com/users/Jorjeous/followers",
"following_url": "https://api.github.com/users/Jorjeous/following{/other_user}",
"gists_url": "https://api.github.com/users/Jorjeous/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Jorjeous/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Jorjeous/subscriptions",
"organizations_url": "https://api.github.com/users/Jorjeous/orgs",
"repos_url": "https://api.github.com/users/Jorjeous/repos",
"events_url": "https://api.github.com/users/Jorjeous/events{/privacy}",
"received_events_url": "https://api.github.com/users/Jorjeous/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "Jorjeous",
"id": 37293288,
"node_id": "MDQ6VXNlcjM3MjkzMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/37293288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Jorjeous",
"html_url": "https://github.com/Jorjeous",
"followers_url": "https://api.github.com/users/Jorjeous/followers",
"following_url": "https://api.github.com/users/Jorjeous/following{/other_user}",
"gists_url": "https://api.github.com/users/Jorjeous/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Jorjeous/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Jorjeous/subscriptions",
"organizations_url": "https://api.github.com/users/Jorjeous/orgs",
"repos_url": "https://api.github.com/users/Jorjeous/repos",
"events_url": "https://api.github.com/users/Jorjeous/events{/privacy}",
"received_events_url": "https://api.github.com/users/Jorjeous/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-11-23T13:16:26 | 2022-11-23T13:17:06 | 2022-11-23T13:17:06 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5488",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5488",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5488.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5488.patch",
"merged_at": null
} | # What does this PR do ?
bugfix
Add a one line overview of what this PR aims to accomplish.
**Collection**: [Note which collection this PR will affect]
ASR
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [+] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "Jorjeous",
"id": 37293288,
"node_id": "MDQ6VXNlcjM3MjkzMjg4",
"avatar_url": "https://avatars.githubusercontent.com/u/37293288?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Jorjeous",
"html_url": "https://github.com/Jorjeous",
"followers_url": "https://api.github.com/users/Jorjeous/followers",
"following_url": "https://api.github.com/users/Jorjeous/following{/other_user}",
"gists_url": "https://api.github.com/users/Jorjeous/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Jorjeous/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Jorjeous/subscriptions",
"organizations_url": "https://api.github.com/users/Jorjeous/orgs",
"repos_url": "https://api.github.com/users/Jorjeous/repos",
"events_url": "https://api.github.com/users/Jorjeous/events{/privacy}",
"received_events_url": "https://api.github.com/users/Jorjeous/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5488/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5488/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5487 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5487/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5487/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5487/events | https://github.com/NVIDIA/NeMo/pull/5487 | 1,461,437,067 | PR_kwDOC_bI7s5Di4t3 | 5,487 | Add missing import in transcribe_utils.py | {
"login": "jonghwanhyeon",
"id": 1588996,
"node_id": "MDQ6VXNlcjE1ODg5OTY=",
"avatar_url": "https://avatars.githubusercontent.com/u/1588996?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jonghwanhyeon",
"html_url": "https://github.com/jonghwanhyeon",
"followers_url": "https://api.github.com/users/jonghwanhyeon/followers",
"following_url": "https://api.github.com/users/jonghwanhyeon/following{/other_user}",
"gists_url": "https://api.github.com/users/jonghwanhyeon/gists{/gist_id}",
"starred_url": "https://api.github.com/users/jonghwanhyeon/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/jonghwanhyeon/subscriptions",
"organizations_url": "https://api.github.com/users/jonghwanhyeon/orgs",
"repos_url": "https://api.github.com/users/jonghwanhyeon/repos",
"events_url": "https://api.github.com/users/jonghwanhyeon/events{/privacy}",
"received_events_url": "https://api.github.com/users/jonghwanhyeon/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-11-23T10:06:32 | 2022-11-23T10:22:41 | 2022-11-23T10:22:41 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5487",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5487",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5487.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5487.patch",
"merged_at": "2022-11-23T10:22:41"
} | # What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
**Collection**: ASR
# Changelog
- Add missing import `model_utils` in `nemo/collections/asr/parts/utils/transcribe_utils.py`
- When I use `examples/asr/transcribe_speech.py`, the following error has been occurred:
```
Traceback (most recent call last):
File "/workspace/NeMo/examples/asr/transcribe_speech.py", line 146, in main
asr_model, model_name = setup_model(cfg, map_location)
File "/workspace/NeMo/nemo/collections/asr/parts/utils/transcribe_utils.py", line 184, in setup_model
imported_class = model_utils.import_class_by_path(classpath) # type: ASRModel
NameError: name 'model_utils' is not defined
```
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [X] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5487/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5487/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5486 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5486/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5486/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5486/events | https://github.com/NVIDIA/NeMo/pull/5486 | 1,461,272,816 | PR_kwDOC_bI7s5DiUOX | 5,486 | Add core classes and functions for online clustering diarizer | {
"login": "tango4j",
"id": 13211483,
"node_id": "MDQ6VXNlcjEzMjExNDgz",
"avatar_url": "https://avatars.githubusercontent.com/u/13211483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tango4j",
"html_url": "https://github.com/tango4j",
"followers_url": "https://api.github.com/users/tango4j/followers",
"following_url": "https://api.github.com/users/tango4j/following{/other_user}",
"gists_url": "https://api.github.com/users/tango4j/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tango4j/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tango4j/subscriptions",
"organizations_url": "https://api.github.com/users/tango4j/orgs",
"repos_url": "https://api.github.com/users/tango4j/repos",
"events_url": "https://api.github.com/users/tango4j/events{/privacy}",
"received_events_url": "https://api.github.com/users/tango4j/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1719393562,
"node_id": "MDU6TGFiZWwxNzE5MzkzNTYy",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/feature",
"name": "feature",
"color": "68F3F0",
"default": false,
"description": "request/PR for a new feature"
},
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-23T08:30:34 | 2022-11-30T22:38:51 | 2022-11-30T22:38:51 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | true | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5486",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5486",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5486.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5486.patch",
"merged_at": null
} | Signed-off-by: Taejin Park <[email protected]>
# What does this PR do ?
Add core classes and functions for online clustering diarizer
**Collection**: [Note which collection this PR will affect]
# Changelog
#### `OnlineSegmentor` class in speaker_utils.py:
online version of speech segmentations
#### `OnlineSpeakerClustering` in nmesc_clustering.py
online version of `SpeakerClustering` class (online speaker counting and permutation matching)
#### `OnlineClusteringDiarizer` in online_diarizer.py (inherits ClusteringDiarizer)
online version of `ClusteringDiarizer` class in clustering_diarizer.py
#### `test_diar_utils.py` in NeMo/tests/collections/asr/
Tests all clustering and online diarization related functions
<img width="966" alt="image" src="https://user-images.githubusercontent.com/13211483/204661137-78d0272b-1594-4509-a486-a8720e835bc4.png">
Checkout this slide for better understanding on why online diarizer needs
(1) linear sum assignment (Hungarian alg)
(2) Interval merging
(3) merge_vector function
[
Slides link
](https://docs.google.com/presentation/d/1LDUTIGRNVuLJsy81V3GaHZlFDUpxZ9rux-fNPfgcLmw/edit?usp=sharing)
If you are curious about why online diarization needs FIFO-queue, take a look at the most recent paper on online neural diarization
[Online Neural Diarization of Unlimited Numbers of Speakers](https://arxiv.org/pdf/2206.02432.pdf)
# Usage
This is core classes and functions update.
Example script will be pushed in the second round of PR.
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [x] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
[Online Neural Diarization of Unlimited Numbers of Speakers](https://arxiv.org/pdf/2206.02432.pdf)
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "tango4j",
"id": 13211483,
"node_id": "MDQ6VXNlcjEzMjExNDgz",
"avatar_url": "https://avatars.githubusercontent.com/u/13211483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tango4j",
"html_url": "https://github.com/tango4j",
"followers_url": "https://api.github.com/users/tango4j/followers",
"following_url": "https://api.github.com/users/tango4j/following{/other_user}",
"gists_url": "https://api.github.com/users/tango4j/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tango4j/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tango4j/subscriptions",
"organizations_url": "https://api.github.com/users/tango4j/orgs",
"repos_url": "https://api.github.com/users/tango4j/repos",
"events_url": "https://api.github.com/users/tango4j/events{/privacy}",
"received_events_url": "https://api.github.com/users/tango4j/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5486/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5486/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5485 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5485/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5485/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5485/events | https://github.com/NVIDIA/NeMo/issues/5485 | 1,461,079,618 | I_kwDOC_bI7s5XFk5C | 5,485 | RuntimeError: Invalid scalar type | {
"login": "triumph9989",
"id": 48488138,
"node_id": "MDQ6VXNlcjQ4NDg4MTM4",
"avatar_url": "https://avatars.githubusercontent.com/u/48488138?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/triumph9989",
"html_url": "https://github.com/triumph9989",
"followers_url": "https://api.github.com/users/triumph9989/followers",
"following_url": "https://api.github.com/users/triumph9989/following{/other_user}",
"gists_url": "https://api.github.com/users/triumph9989/gists{/gist_id}",
"starred_url": "https://api.github.com/users/triumph9989/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/triumph9989/subscriptions",
"organizations_url": "https://api.github.com/users/triumph9989/orgs",
"repos_url": "https://api.github.com/users/triumph9989/repos",
"events_url": "https://api.github.com/users/triumph9989/events{/privacy}",
"received_events_url": "https://api.github.com/users/triumph9989/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
},
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-23T05:58:28 | 2022-11-25T16:12:13 | 2022-11-25T16:12:13 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
Hi,
I tried to train an ASR model, then the following error was appearance.
It seems to be some error about distribution.
```
Initializing distributed: GLOBAL_RANK: 0, MEMBER: 1/1
Added key: store_based_barrier_key:1 to store for rank: 0
Rank 0: Completed store-based barrier for key:store_based_barrier_key:1 with 1 nodes.
----------------------------------------------------------------------------------------------------
distributed_backend=gloo
All distributed processes registered. Starting with 1 processes
----------------------------------------------------------------------------------------------------
Error executing job with overrides: []
Traceback (most recent call last):
File "speech_to_text_ctc.py", line 114, in main
trainer.fit(asr_model)
File "/home/face/.conda/envs/nemo/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py", line 696, in fit
self._call_and_handle_interrupt(
File "/home/face/.conda/envs/nemo/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py", line 648, in _call_and_handle_interrupt
return self.strategy.launcher.launch(trainer_fn, *args, trainer=self, **kwargs)
File "/home/face/.conda/envs/nemo/lib/python3.8/site-packages/pytorch_lightning/strategies/launchers/subprocess_script.py", line 93, in launch
return function(*args, **kwargs)
File "/home/face/.conda/envs/nemo/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py", line 735, in _fit_impl
results = self._run(model, ckpt_path=self.ckpt_path)
File "/home/face/.conda/envs/nemo/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py", line 1147, in _run
self.strategy.setup(self)
File "/home/face/.conda/envs/nemo/lib/python3.8/site-packages/pytorch_lightning/strategies/ddp.py", line 181, in setup
self.configure_ddp()
File "/home/face/.conda/envs/nemo/lib/python3.8/site-packages/pytorch_lightning/strategies/ddp.py", line 302, in configure_ddp
self.model = self._setup_model(LightningDistributedModule(self.model))
File "/home/face/.conda/envs/nemo/lib/python3.8/site-packages/pytorch_lightning/strategies/ddp.py", line 197, in _setup_model
return DistributedDataParallel(module=model, device_ids=device_ids, **self._ddp_kwargs)
File "/home/face/.conda/envs/nemo/lib/python3.8/site-packages/torch/nn/parallel/distributed.py", line 657, in __init__
_sync_module_states(
File "/home/face/.conda/envs/nemo/lib/python3.8/site-packages/torch/distributed/utils.py", line 136, in _sync_module_states
_sync_params_and_buffers(
File "/home/face/.conda/envs/nemo/lib/python3.8/site-packages/torch/distributed/utils.py", line 154, in _sync_params_and_buffers
dist._broadcast_coalesced(
RuntimeError: Invalid scalar type
```
**Steps/Code to reproduce bug**
`CUDA_VISIBLE_DEVICES=0,1python speech_to_text_ctc.py`
**conformer_ctc_char.yaml**
```
# It contains the default values for training a Conformer-CTC ASR model, large size (~120M) with CTC loss and char-based vocabulary.
# Char-based encoding may give lower accuracy than sub-word encoding (conformer_ctc_bpe.yaml).
# You may find more detail on Conformer-CTC at `examples/asr/conf/conformer/conformer_ctc_bpe.yaml`
name: "Conformer-CTC-Char-Aishell-10ep-nun-classes-4466"
model:
sample_rate: 16000
labels: [" ", "a", "b", "c", "d", "e", "f", "g", "h", "i", "j", "k", "l", "m",
"n", "o", "p", "q", "r", "s", "t", "u", "v", "w", "x", "y", "z", "'"]
log_prediction: true # enables logging sample predictions in the output during training
ctc_reduction: 'mean_batch'
skip_nan_grad: false
train_ds:
manifest_filepath: ???
labels: ${model.labels}
sample_rate: ${model.sample_rate}
batch_size: 16 # you may increase batch_size if your memory allows
shuffle: true
num_workers: 8
pin_memory: true
trim_silence: false
max_duration: 16.7 # it is set for LibriSpeech, you may need to update it for your dataset
min_duration: 0.1
# tarred datasets
is_tarred: false
tarred_audio_filepaths: null
shuffle_n: 2048
# bucketing params
bucketing_strategy: "synced_randomized"
bucketing_batch_size: null
validation_ds:
manifest_filepath: ???
labels: ${model.labels}
sample_rate: ${model.sample_rate}
batch_size: 16 # you may increase batch_size if your memory allows
shuffle: false
num_workers: 8
pin_memory: true
test_ds:
manifest_filepath: null
labels: ${model.labels}
sample_rate: ${model.sample_rate}
batch_size: 16 # you may increase batch_size if your memory allows
shuffle: false
num_workers: 8
pin_memory: true
preprocessor:
_target_: nemo.collections.asr.modules.AudioToMelSpectrogramPreprocessor
sample_rate: ${model.sample_rate}
normalize: "per_feature"
window_size: 0.025
window_stride: 0.01
window: "hann"
features: 80
n_fft: 512
log: true
frame_splicing: 1
dither: 0.00001
pad_to: 0
pad_value: 0.0
spec_augment:
_target_: nemo.collections.asr.modules.SpectrogramAugmentation
freq_masks: 2 # set to zero to disable it
# you may use lower time_masks for smaller models to have a faster convergence
time_masks: 10 # set to zero to disable it
freq_width: 27
time_width: 0.05
encoder:
_target_: nemo.collections.asr.modules.ConformerEncoder
feat_in: ${model.preprocessor.features}
feat_out: -1 # you may set it if you need different output size other than the default d_model
n_layers: 16
d_model: 256
# Sub-sampling params
subsampling: striding # vggnet, striding, stacking or stacking_norm, dw_striding
subsampling_factor: 4 # must be power of 2 for striding and vggnet
subsampling_conv_channels: -1 # set to -1 to make it equal to the d_model
causal_downsampling: True
# Feed forward module's params
ff_expansion_factor: 4
# Multi-headed Attention Module's params
self_attention_model: rel_pos # rel_pos or abs_pos
n_heads: 8 # may need to be lower for smaller d_models
# [left, right] specifies the number of steps to be seen from left and right of each step in self-attention
att_context_size: [-1, -1] # -1 means unlimited context
att_context_style: regular # regular or chunked_limited
xscaling: true # scales up the input embeddings by sqrt(d_model)
untie_biases: true # unties the biases of the TransformerXL layers
pos_emb_max_len: 5000
# Convolution module's params
conv_kernel_size: 31
conv_norm_type: 'batch_norm' # batch_norm or layer_norm or groupnormN (N specifies the number of groups)
# conv_context_size can be"causal" or a list of two integers while conv_context_size[0]+conv_context_size[1]+1==conv_kernel_size
# null means [(kernel_size-1)//2, (kernel_size-1)//2], and 'causal' means [(kernel_size-1), 0]
conv_context_size: null
### regularization
dropout: 0.1 # The dropout used in most of the Conformer Modules
dropout_emb: 0.0 # The dropout used for embeddings
dropout_att: 0.1 # The dropout for multi-headed attention modules
decoder:
_target_: nemo.collections.asr.modules.ConvASRDecoder
feat_in: null
num_classes: 4465
vocabulary: ${model.labels}
optim:
name: adamw
lr: 1.5 #2.0
# optimizer arguments
betas: [0.9, 0.98]
# less necessity for weight_decay as we already have large augmentations with SpecAug
# you may need weight_decay for large models, stable AMP training, small datasets, or when lower augmentations are used
# weight decay of 0.0 with lr of 2.0 also works fine
weight_decay: 1e-3
# scheduler setup
sched:
name: NoamAnnealing
d_model: ${model.encoder.d_model}
# scheduler config override
warmup_steps: 10000
warmup_ratio: null
min_lr: 1e-6
trainer:
devices: 1 # number of GPUs, -1 would use all available GPUs
num_nodes: 1
max_epochs: 10
max_steps: -1 # computed at runtime if not set
val_check_interval: 1.0 # Set to 0.25 to check 4 times per epoch, or an int for number of iterations
accelerator: auto
strategy: ddp_find_unused_parameters_false
accumulate_grad_batches: 1
gradient_clip_val: 0.0
precision: 32 # Should be set to 16 for O1 and O2 to enable the AMP.
log_every_n_steps: 10 # Interval of logging.
enable_progress_bar: True
resume_from_checkpoint: null # The path to a checkpoint file to continue the training, restores the whole state including the epoch, step, LR schedulers, apex, etc.
num_sanity_val_steps: 0 # number of steps to perform validation steps for sanity check the validation process before starting the training, setting to 0 disables it
check_val_every_n_epoch: 1 # number of evaluations on validation every n epochs
sync_batchnorm: true
enable_checkpointing: False # Provided by exp_manager
logger: false # Provided by exp_manager
benchmark: false # needs to be false for models with variable-length speech input as it slows down training
exp_manager:
exp_dir: ../exp
name: ${name}
create_tensorboard_logger: true
create_checkpoint_callback: true
checkpoint_callback_params:
# in case of multiple validation sets, first one is used
monitor: "val_wer"
mode: "min"
save_top_k: 5
always_save_nemo: True # saves the checkpoints as nemo files instead of PTL checkpoints
# you need to set these two to True to continue the training
resume_if_exists: false
resume_ignore_no_checkpoint: false
# You may use this section to create a W&B logger
create_wandb_logger: false
wandb_logger_kwargs:
name: null
project: null
```
**speech_to_text_ctc.py**
```
from curses import use_default_colors
import pytorch_lightning as pl
from omegaconf import OmegaConf
from nemo.collections.asr.models import EncDecCTCModel
from nemo.core.config import hydra_runner
from nemo.utils import logging
from nemo.utils.exp_manager import exp_manager
use_cer = True
log_prediction = True
manifest_path = "/home/face/NeMo/examples/asr/"
train_manifest = f"{manifest_path}train_manifest.json"
dev_manifest = f"{manifest_path}dev_manifest.json"
test_manifest = f"{manifest_path}test_manifest.json"
charset = []
#將 vocab.txt 讀入 list 中，使其能寫入 config 的 labels
with open(f"{manifest_path}Aishell_vocab/train_dev.txt ", "r") as vocab_read:
for char in vocab_read.read():
charset.append(char) #EX: ['余', '耍', '汞', '圳']
# print(charset)
# print(len(charset))
@hydra_runner(config_path="../conf/conformer", config_name="conformer_ctc_char")
def main(cfg):
cfg.model.labels = charset
cfg.model.train_ds.manifest_filepath = train_manifest
cfg.model.validation_ds.manifest_filepath = dev_manifest
cfg.model.test_ds.manifest_filepath = test_manifest
logging.info(f'Hydra config: {OmegaConf.to_yaml(cfg)}')
trainer = pl.Trainer(**cfg.trainer)
exp_manager(trainer, cfg.get("exp_manager", None))
asr_model = EncDecCTCModel(cfg=cfg.model, trainer=trainer)
asr_model._wer.use_cer = use_cer
asr_model._wer.log_prediction = log_prediction
# Initialize the weights of the model from another model, if provided via config
asr_model.maybe_init_from_pretrained_checkpoint(cfg)
trainer.fit(asr_model)
if hasattr(cfg.model, 'test_ds') and cfg.model.test_ds.manifest_filepath is not None:
if asr_model.prepare_test(trainer):
trainer.test(asr_model)
if __name__ == '__main__':
main() # noqa pylint: disable=no-value-for-parameter
```
**Environment overview (please complete the following information)**
- Environment location: Bare-metal
- Method of NeMo install: From source
**Environment details**
-nemo-tookit
1.12.0
- OS version
Linux version 4.15.0-142-generic (buildd@lgw01-amd64-039) (gcc version 5.4.0 20160609 (Ubuntu 5.4.0-6ubuntu1~16.04.12)) #146~16.04.1-Ubuntu SMP Tue Apr 13 09:27:15 UTC 2021
- PyTorch version
1.13.0
- Python version
3.8
**Additional context**
Add any other context about the problem here.
GPU: NVIDIA GeForce 2080TI
| {
"login": "SeanNaren",
"id": 6707363,
"node_id": "MDQ6VXNlcjY3MDczNjM=",
"avatar_url": "https://avatars.githubusercontent.com/u/6707363?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SeanNaren",
"html_url": "https://github.com/SeanNaren",
"followers_url": "https://api.github.com/users/SeanNaren/followers",
"following_url": "https://api.github.com/users/SeanNaren/following{/other_user}",
"gists_url": "https://api.github.com/users/SeanNaren/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SeanNaren/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SeanNaren/subscriptions",
"organizations_url": "https://api.github.com/users/SeanNaren/orgs",
"repos_url": "https://api.github.com/users/SeanNaren/repos",
"events_url": "https://api.github.com/users/SeanNaren/events{/privacy}",
"received_events_url": "https://api.github.com/users/SeanNaren/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5485/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5485/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5484 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5484/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5484/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5484/events | https://github.com/NVIDIA/NeMo/pull/5484 | 1,461,045,231 | PR_kwDOC_bI7s5DhhDb | 5,484 | Cleanup docs on r1.13.0 | {
"login": "okuchaiev",
"id": 22035961,
"node_id": "MDQ6VXNlcjIyMDM1OTYx",
"avatar_url": "https://avatars.githubusercontent.com/u/22035961?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/okuchaiev",
"html_url": "https://github.com/okuchaiev",
"followers_url": "https://api.github.com/users/okuchaiev/followers",
"following_url": "https://api.github.com/users/okuchaiev/following{/other_user}",
"gists_url": "https://api.github.com/users/okuchaiev/gists{/gist_id}",
"starred_url": "https://api.github.com/users/okuchaiev/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/okuchaiev/subscriptions",
"organizations_url": "https://api.github.com/users/okuchaiev/orgs",
"repos_url": "https://api.github.com/users/okuchaiev/repos",
"events_url": "https://api.github.com/users/okuchaiev/events{/privacy}",
"received_events_url": "https://api.github.com/users/okuchaiev/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | null | [] | null | null | 2022-11-23T05:25:02 | 2023-01-31T01:03:45 | 2023-01-31T01:03:45 | MEMBER | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5484",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5484",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5484.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5484.patch",
"merged_at": null
} | Signed-off-by: Oleksii Kuchaiev <[email protected]>
# What does this PR do ?
Some fixes and minor improvements to README.rst and RTD docs,
# Changelog
- [x] Moves Requirements and Installation into single source of truth (on RTD)
- [x] Cleans up README.rst to make it less verbose and removing redundant and obsolete links
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [x] Documentation
| {
"login": "titu1994",
"id": 3048602,
"node_id": "MDQ6VXNlcjMwNDg2MDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/3048602?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/titu1994",
"html_url": "https://github.com/titu1994",
"followers_url": "https://api.github.com/users/titu1994/followers",
"following_url": "https://api.github.com/users/titu1994/following{/other_user}",
"gists_url": "https://api.github.com/users/titu1994/gists{/gist_id}",
"starred_url": "https://api.github.com/users/titu1994/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/titu1994/subscriptions",
"organizations_url": "https://api.github.com/users/titu1994/orgs",
"repos_url": "https://api.github.com/users/titu1994/repos",
"events_url": "https://api.github.com/users/titu1994/events{/privacy}",
"received_events_url": "https://api.github.com/users/titu1994/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5484/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5484/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5483 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5483/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5483/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5483/events | https://github.com/NVIDIA/NeMo/issues/5483 | 1,460,997,342 | I_kwDOC_bI7s5XFQze | 5,483 | Inference Size Error Speaker_ Diarization | {
"login": "arnavmehta7",
"id": 65492948,
"node_id": "MDQ6VXNlcjY1NDkyOTQ4",
"avatar_url": "https://avatars.githubusercontent.com/u/65492948?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/arnavmehta7",
"html_url": "https://github.com/arnavmehta7",
"followers_url": "https://api.github.com/users/arnavmehta7/followers",
"following_url": "https://api.github.com/users/arnavmehta7/following{/other_user}",
"gists_url": "https://api.github.com/users/arnavmehta7/gists{/gist_id}",
"starred_url": "https://api.github.com/users/arnavmehta7/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/arnavmehta7/subscriptions",
"organizations_url": "https://api.github.com/users/arnavmehta7/orgs",
"repos_url": "https://api.github.com/users/arnavmehta7/repos",
"events_url": "https://api.github.com/users/arnavmehta7/events{/privacy}",
"received_events_url": "https://api.github.com/users/arnavmehta7/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
},
{
"id": 4585431311,
"node_id": "LA_kwDOC_bI7s8AAAABEVAhDw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale",
"name": "stale",
"color": "ededed",
"default": false,
"description": null
},
{
"id": 4811888028,
"node_id": "LA_kwDOC_bI7s8AAAABHs-VnA",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/Speaker%20Tasks",
"name": "Speaker Tasks",
"color": "871AD4",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-23T04:39:00 | 2023-01-09T01:58:14 | 2023-01-09T01:58:14 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | On the speaker_diarization ipynb file. If I switch the link to the audio to a custom audio then I get a shape error on inference. It shows that I have a tensor but they required in some other format.
**Steps/Code to reproduce bug**
[Link of Notebook
](https://github.com/NVIDIA/NeMo/blob/main/tutorials/speaker_tasks/Speaker_Diarization_Inference.ipynb)
On this notebook, just switch the link of audio or upload a audio and replace it with the an4 one.
**Expected behavior**
Diarization perfectly finished.
**Environment overview (please complete the following information)**
Colab.
**Environment details**
Colab.
| {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5483/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5483/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5482 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5482/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5482/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5482/events | https://github.com/NVIDIA/NeMo/pull/5482 | 1,460,820,854 | PR_kwDOC_bI7s5Dgtqn | 5,482 | export_utils bugfix | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4556669155,
"node_id": "LA_kwDOC_bI7s8AAAABD5lA4w",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/cherry-pick",
"name": "cherry-pick",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | null | [] | null | null | 2022-11-23T02:06:12 | 2022-11-23T04:17:10 | 2022-11-23T04:17:05 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5482",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5482",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5482.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5482.patch",
"merged_at": "2022-11-23T04:17:05"
} | # What does this PR do ?
Updates MatchedScaleMaskSoftmax replacement
**Collection**: [Note which collection this PR will affect]
# Changelog
- Removes redundant class in export_utils
# Usage
* N/A
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [X] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [X] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "borisfom",
"id": 14189615,
"node_id": "MDQ6VXNlcjE0MTg5NjE1",
"avatar_url": "https://avatars.githubusercontent.com/u/14189615?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/borisfom",
"html_url": "https://github.com/borisfom",
"followers_url": "https://api.github.com/users/borisfom/followers",
"following_url": "https://api.github.com/users/borisfom/following{/other_user}",
"gists_url": "https://api.github.com/users/borisfom/gists{/gist_id}",
"starred_url": "https://api.github.com/users/borisfom/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/borisfom/subscriptions",
"organizations_url": "https://api.github.com/users/borisfom/orgs",
"repos_url": "https://api.github.com/users/borisfom/repos",
"events_url": "https://api.github.com/users/borisfom/events{/privacy}",
"received_events_url": "https://api.github.com/users/borisfom/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5482/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5482/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5481 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5481/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5481/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5481/events | https://github.com/NVIDIA/NeMo/pull/5481 | 1,460,746,818 | PR_kwDOC_bI7s5DgdmS | 5,481 | [BugFix] Removing <unk> tokens from decoding timestamp | {
"login": "tango4j",
"id": 13211483,
"node_id": "MDQ6VXNlcjEzMjExNDgz",
"avatar_url": "https://avatars.githubusercontent.com/u/13211483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tango4j",
"html_url": "https://github.com/tango4j",
"followers_url": "https://api.github.com/users/tango4j/followers",
"following_url": "https://api.github.com/users/tango4j/following{/other_user}",
"gists_url": "https://api.github.com/users/tango4j/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tango4j/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tango4j/subscriptions",
"organizations_url": "https://api.github.com/users/tango4j/orgs",
"repos_url": "https://api.github.com/users/tango4j/repos",
"events_url": "https://api.github.com/users/tango4j/events{/privacy}",
"received_events_url": "https://api.github.com/users/tango4j/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-11-23T00:18:01 | 2022-11-29T04:24:28 | 2022-11-29T04:24:24 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5481",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5481",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5481.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5481.patch",
"merged_at": "2022-11-29T04:24:23"
} | Signed-off-by: Taejin Park <[email protected]>
# What does this PR do ?
Removing `<unk>` tokens from decoding timestamp.
This solves length mismatching error in audio files contain lots of silence.
**Collection**: [Note which collection this PR will affect]
ASR (speaker tasks)
# Changelog
in decoder_timestamp_utils.py, this PR is making change to exclude
`<unk>` tokens for both word and word-timestamp output.
`<unk>` tokens are outputted as " " (blank space) in text hypothesis output,
and this created mismatch of length since split() command removes all kinds of space
at the start/end
- Also fixed ASR with diar notebook bugs
- Also changed hard-coded variable to a function argument
# Usage
```python
python offline_diar_with_asr_infer.py \
diarizer.manifest_filepath=<path to manifest file> \
diarizer.out_dir='demo_asr_output' \
diarizer.speaker_embeddings.model_path=<pretrained modelname or path to .nemo> \
diarizer.asr.model_path=<pretrained modelname or path to .nemo> \
diarizer.asr.parameters.asr_based_vad=True \
diarizer.speaker_embeddings.parameters.save_embeddings=False
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [x] Did you write any new necessary tests?
- [x] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [x] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in ASR
| {
"login": "nithinraok",
"id": 19668129,
"node_id": "MDQ6VXNlcjE5NjY4MTI5",
"avatar_url": "https://avatars.githubusercontent.com/u/19668129?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/nithinraok",
"html_url": "https://github.com/nithinraok",
"followers_url": "https://api.github.com/users/nithinraok/followers",
"following_url": "https://api.github.com/users/nithinraok/following{/other_user}",
"gists_url": "https://api.github.com/users/nithinraok/gists{/gist_id}",
"starred_url": "https://api.github.com/users/nithinraok/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/nithinraok/subscriptions",
"organizations_url": "https://api.github.com/users/nithinraok/orgs",
"repos_url": "https://api.github.com/users/nithinraok/repos",
"events_url": "https://api.github.com/users/nithinraok/events{/privacy}",
"received_events_url": "https://api.github.com/users/nithinraok/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5481/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5481/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5480 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5480/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5480/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5480/events | https://github.com/NVIDIA/NeMo/pull/5480 | 1,460,675,889 | PR_kwDOC_bI7s5DgNqk | 5,480 | export_utils bugfix | {
"login": "Davood-M",
"id": 3964285,
"node_id": "MDQ6VXNlcjM5NjQyODU=",
"avatar_url": "https://avatars.githubusercontent.com/u/3964285?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Davood-M",
"html_url": "https://github.com/Davood-M",
"followers_url": "https://api.github.com/users/Davood-M/followers",
"following_url": "https://api.github.com/users/Davood-M/following{/other_user}",
"gists_url": "https://api.github.com/users/Davood-M/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Davood-M/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Davood-M/subscriptions",
"organizations_url": "https://api.github.com/users/Davood-M/orgs",
"repos_url": "https://api.github.com/users/Davood-M/repos",
"events_url": "https://api.github.com/users/Davood-M/events{/privacy}",
"received_events_url": "https://api.github.com/users/Davood-M/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-11-22T23:04:26 | 2022-11-23T02:05:51 | 2022-11-23T02:05:41 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5480",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5480",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5480.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5480.patch",
"merged_at": "2022-11-23T02:05:41"
} | # What does this PR do ?
Updates MatchedScaleMaskSoftmax replacement
**Collection**: [Note which collection this PR will affect]
# Changelog
- Removes redundant class in export_utils
# Usage
* N/A
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [X] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [X] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "borisfom",
"id": 14189615,
"node_id": "MDQ6VXNlcjE0MTg5NjE1",
"avatar_url": "https://avatars.githubusercontent.com/u/14189615?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/borisfom",
"html_url": "https://github.com/borisfom",
"followers_url": "https://api.github.com/users/borisfom/followers",
"following_url": "https://api.github.com/users/borisfom/following{/other_user}",
"gists_url": "https://api.github.com/users/borisfom/gists{/gist_id}",
"starred_url": "https://api.github.com/users/borisfom/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/borisfom/subscriptions",
"organizations_url": "https://api.github.com/users/borisfom/orgs",
"repos_url": "https://api.github.com/users/borisfom/repos",
"events_url": "https://api.github.com/users/borisfom/events{/privacy}",
"received_events_url": "https://api.github.com/users/borisfom/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5480/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5480/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5479 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5479/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5479/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5479/events | https://github.com/NVIDIA/NeMo/pull/5479 | 1,460,435,467 | PR_kwDOC_bI7s5DfY_k | 5,479 | [ASR] Transcribe for multi-channel signals | {
"login": "anteju",
"id": 108555623,
"node_id": "U_kgDOBnhtZw",
"avatar_url": "https://avatars.githubusercontent.com/u/108555623?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/anteju",
"html_url": "https://github.com/anteju",
"followers_url": "https://api.github.com/users/anteju/followers",
"following_url": "https://api.github.com/users/anteju/following{/other_user}",
"gists_url": "https://api.github.com/users/anteju/gists{/gist_id}",
"starred_url": "https://api.github.com/users/anteju/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/anteju/subscriptions",
"organizations_url": "https://api.github.com/users/anteju/orgs",
"repos_url": "https://api.github.com/users/anteju/repos",
"events_url": "https://api.github.com/users/anteju/events{/privacy}",
"received_events_url": "https://api.github.com/users/anteju/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1719393562,
"node_id": "MDU6TGFiZWwxNzE5MzkzNTYy",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/feature",
"name": "feature",
"color": "68F3F0",
"default": false,
"description": "request/PR for a new feature"
},
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
}
] | closed | false | {
"login": "anteju",
"id": 108555623,
"node_id": "U_kgDOBnhtZw",
"avatar_url": "https://avatars.githubusercontent.com/u/108555623?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/anteju",
"html_url": "https://github.com/anteju",
"followers_url": "https://api.github.com/users/anteju/followers",
"following_url": "https://api.github.com/users/anteju/following{/other_user}",
"gists_url": "https://api.github.com/users/anteju/gists{/gist_id}",
"starred_url": "https://api.github.com/users/anteju/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/anteju/subscriptions",
"organizations_url": "https://api.github.com/users/anteju/orgs",
"repos_url": "https://api.github.com/users/anteju/repos",
"events_url": "https://api.github.com/users/anteju/events{/privacy}",
"received_events_url": "https://api.github.com/users/anteju/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "anteju",
"id": 108555623,
"node_id": "U_kgDOBnhtZw",
"avatar_url": "https://avatars.githubusercontent.com/u/108555623?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/anteju",
"html_url": "https://github.com/anteju",
"followers_url": "https://api.github.com/users/anteju/followers",
"following_url": "https://api.github.com/users/anteju/following{/other_user}",
"gists_url": "https://api.github.com/users/anteju/gists{/gist_id}",
"starred_url": "https://api.github.com/users/anteju/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/anteju/subscriptions",
"organizations_url": "https://api.github.com/users/anteju/orgs",
"repos_url": "https://api.github.com/users/anteju/repos",
"events_url": "https://api.github.com/users/anteju/events{/privacy}",
"received_events_url": "https://api.github.com/users/anteju/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-11-22T19:24:39 | 2022-11-22T22:13:31 | 2022-11-22T22:13:31 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5479",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5479",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5479.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5479.patch",
"merged_at": "2022-11-22T22:13:31"
} |
# What does this PR do ?
This PR adds the following optional parameters to ASR transcription:
- select a single channel from multi-channel input
- override the default audio key (`"audio_filepath"`) when loading paths from a manifest
**Collection**: ASR
# Changelog
- Added optional `channel_selector` parameter to `TranscriptionConfig`
- Added optional `audio_key` parameter to `TranscriptionConfig`
- Propagated the two parameters to `transcribe` and `transcribe_utils`
# Usage
Transcription config can be updated, for example as
```python
cfg.channel_selector = 0 # Evaluate on the first channel
cfg.audio_key = 'alternative_audio_filepath' # Load audio from an alternative filepath provided in the manifest
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [x] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "anteju",
"id": 108555623,
"node_id": "U_kgDOBnhtZw",
"avatar_url": "https://avatars.githubusercontent.com/u/108555623?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/anteju",
"html_url": "https://github.com/anteju",
"followers_url": "https://api.github.com/users/anteju/followers",
"following_url": "https://api.github.com/users/anteju/following{/other_user}",
"gists_url": "https://api.github.com/users/anteju/gists{/gist_id}",
"starred_url": "https://api.github.com/users/anteju/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/anteju/subscriptions",
"organizations_url": "https://api.github.com/users/anteju/orgs",
"repos_url": "https://api.github.com/users/anteju/repos",
"events_url": "https://api.github.com/users/anteju/events{/privacy}",
"received_events_url": "https://api.github.com/users/anteju/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5479/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5479/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5478 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5478/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5478/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5478/events | https://github.com/NVIDIA/NeMo/pull/5478 | 1,460,362,278 | PR_kwDOC_bI7s5DfIyS | 5,478 | Cherry Pick T5 finetuning changes into 1.13 | {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-11-22T18:28:25 | 2022-11-22T19:52:07 | 2022-11-22T19:51:59 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5478",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5478",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5478.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5478.patch",
"merged_at": "2022-11-22T19:51:59"
} | …tch size refactoring (#5339)
* Initial refactor
Signed-off-by: MaximumEntropy <[email protected]>
* Resolve config before passing to load_from_checkpoint
Signed-off-by: MaximumEntropy <[email protected]>
* Fixes for model parallel and nemo restore
Signed-off-by: MaximumEntropy <[email protected]>
* Fixes for eval
Signed-off-by: MaximumEntropy <[email protected]>
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* Revert config changes
Signed-off-by: MaximumEntropy <[email protected]>
* Refactor
Signed-off-by: MaximumEntropy <[email protected]>
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* Fix typo
Signed-off-by: MaximumEntropy <[email protected]>
* Remove comments
Signed-off-by: MaximumEntropy <[email protected]>
* Minor
Signed-off-by: MaximumEntropy <[email protected]>
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* Fix validation reconfiguration
Signed-off-by: MaximumEntropy <[email protected]>
* Remove old comment
Signed-off-by: MaximumEntropy <[email protected]>
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* Fixes for test_ds
Signed-off-by: MaximumEntropy <[email protected]>
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
Signed-off-by: MaximumEntropy <[email protected]>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
# What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5478/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5478/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5477 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5477/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5477/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5477/events | https://github.com/NVIDIA/NeMo/pull/5477 | 1,460,332,255 | PR_kwDOC_bI7s5DfCHH | 5,477 | Fix a bug with positional vs key-word based argument passing in the transformer layer | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
},
{
"id": 4556669155,
"node_id": "LA_kwDOC_bI7s8AAAABD5lA4w",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/cherry-pick",
"name": "cherry-pick",
"color": "ededed",
"default": false,
"description": null
},
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-22T18:08:37 | 2023-02-21T08:28:49 | 2022-11-22T19:52:22 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5477",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5477",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5477.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5477.patch",
"merged_at": "2022-11-22T19:52:22"
} | Signed-off-by: MaximumEntropy <[email protected]>
# What does this PR do ?
Fixes a bug that affects T5-RPE due to positional vs key-word based argument passing.
**Collection**: NLP
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5477/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5477/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5476 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5476/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5476/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5476/events | https://github.com/NVIDIA/NeMo/pull/5476 | 1,459,161,955 | PR_kwDOC_bI7s5Da_uZ | 5,476 | Add Silence Augmentation | {
"login": "fayejf",
"id": 36722593,
"node_id": "MDQ6VXNlcjM2NzIyNTkz",
"avatar_url": "https://avatars.githubusercontent.com/u/36722593?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/fayejf",
"html_url": "https://github.com/fayejf",
"followers_url": "https://api.github.com/users/fayejf/followers",
"following_url": "https://api.github.com/users/fayejf/following{/other_user}",
"gists_url": "https://api.github.com/users/fayejf/gists{/gist_id}",
"starred_url": "https://api.github.com/users/fayejf/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/fayejf/subscriptions",
"organizations_url": "https://api.github.com/users/fayejf/orgs",
"repos_url": "https://api.github.com/users/fayejf/repos",
"events_url": "https://api.github.com/users/fayejf/events{/privacy}",
"received_events_url": "https://api.github.com/users/fayejf/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1719393562,
"node_id": "MDU6TGFiZWwxNzE5MzkzNTYy",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/feature",
"name": "feature",
"color": "68F3F0",
"default": false,
"description": "request/PR for a new feature"
},
{
"id": 4811884691,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR",
"name": "ASR",
"color": "F3B33E",
"default": false,
"description": ""
},
{
"id": 4811888028,
"node_id": "LA_kwDOC_bI7s8AAAABHs-VnA",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/Speaker%20Tasks",
"name": "Speaker Tasks",
"color": "871AD4",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-22T06:01:02 | 2022-11-23T18:10:08 | 2022-11-23T18:10:03 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5476",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5476",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5476.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5476.patch",
"merged_at": "2022-11-23T18:10:03"
} | Signed-off-by: fayejf <[email protected]>
# What does this PR do ?
Add support for online silence perturbation for data augmentation and robustness test.
**Collection**: ASR
# Changelog
- Add silence at the beginning and/or end of an utterance.
# Usage
* You can potentially add a usage example below
```python
train_ds:
.....
augmentor:
silence:
prob: 0.8
min_start_silence_secs: 0
max_start_silence_secs: 5
min_end_silence_secs: 0
max_end_silence_secs: 5
```
# Before your PR is "Ready for review"
**Pre checks**:
- [x] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [x] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
Will add this to [Online_Noise_Augmentation tutorial](https://github.com/NVIDIA/NeMo/blob/main/tutorials/asr/Online_Noise_Augmentation.ipynb) later if necessary
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [x] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "fayejf",
"id": 36722593,
"node_id": "MDQ6VXNlcjM2NzIyNTkz",
"avatar_url": "https://avatars.githubusercontent.com/u/36722593?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/fayejf",
"html_url": "https://github.com/fayejf",
"followers_url": "https://api.github.com/users/fayejf/followers",
"following_url": "https://api.github.com/users/fayejf/following{/other_user}",
"gists_url": "https://api.github.com/users/fayejf/gists{/gist_id}",
"starred_url": "https://api.github.com/users/fayejf/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/fayejf/subscriptions",
"organizations_url": "https://api.github.com/users/fayejf/orgs",
"repos_url": "https://api.github.com/users/fayejf/repos",
"events_url": "https://api.github.com/users/fayejf/events{/privacy}",
"received_events_url": "https://api.github.com/users/fayejf/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5476/reactions",
"total_count": 1,
"+1": 1,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5476/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5475 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5475/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5475/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5475/events | https://github.com/NVIDIA/NeMo/pull/5475 | 1,458,962,421 | PR_kwDOC_bI7s5DaSTY | 5,475 | Fix a bug with positional vs key-word based argument passing in the transformer layer | {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
},
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-22T03:18:10 | 2022-11-22T18:08:12 | 2022-11-22T18:08:02 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5475",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5475",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5475.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5475.patch",
"merged_at": "2022-11-22T18:08:02"
} | Signed-off-by: MaximumEntropy <[email protected]>
# What does this PR do ?
Fixes a bug that affects T5-RPE due to positional vs key-word based argument passing.
**Collection**: NLP
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5475/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5475/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5474 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5474/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5474/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5474/events | https://github.com/NVIDIA/NeMo/pull/5474 | 1,458,766,555 | PR_kwDOC_bI7s5DZnKp | 5,474 | Bug Fix for bert to run on entire validation dataset | {
"login": "shanmugamr1992",
"id": 111910568,
"node_id": "U_kgDOBqueqA",
"avatar_url": "https://avatars.githubusercontent.com/u/111910568?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/shanmugamr1992",
"html_url": "https://github.com/shanmugamr1992",
"followers_url": "https://api.github.com/users/shanmugamr1992/followers",
"following_url": "https://api.github.com/users/shanmugamr1992/following{/other_user}",
"gists_url": "https://api.github.com/users/shanmugamr1992/gists{/gist_id}",
"starred_url": "https://api.github.com/users/shanmugamr1992/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/shanmugamr1992/subscriptions",
"organizations_url": "https://api.github.com/users/shanmugamr1992/orgs",
"repos_url": "https://api.github.com/users/shanmugamr1992/repos",
"events_url": "https://api.github.com/users/shanmugamr1992/events{/privacy}",
"received_events_url": "https://api.github.com/users/shanmugamr1992/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
},
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-21T23:52:11 | 2022-11-23T19:17:32 | 2022-11-23T19:17:31 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5474",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5474",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5474.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5474.patch",
"merged_at": null
} | Signed-off-by: Shanmugam Ramasamy <[email protected]>
# What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "shanmugamr1992",
"id": 111910568,
"node_id": "U_kgDOBqueqA",
"avatar_url": "https://avatars.githubusercontent.com/u/111910568?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/shanmugamr1992",
"html_url": "https://github.com/shanmugamr1992",
"followers_url": "https://api.github.com/users/shanmugamr1992/followers",
"following_url": "https://api.github.com/users/shanmugamr1992/following{/other_user}",
"gists_url": "https://api.github.com/users/shanmugamr1992/gists{/gist_id}",
"starred_url": "https://api.github.com/users/shanmugamr1992/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/shanmugamr1992/subscriptions",
"organizations_url": "https://api.github.com/users/shanmugamr1992/orgs",
"repos_url": "https://api.github.com/users/shanmugamr1992/repos",
"events_url": "https://api.github.com/users/shanmugamr1992/events{/privacy}",
"received_events_url": "https://api.github.com/users/shanmugamr1992/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5474/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5474/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5473 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5473/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5473/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5473/events | https://github.com/NVIDIA/NeMo/pull/5473 | 1,458,690,185 | PR_kwDOC_bI7s5DZWH1 | 5,473 | Add num layers check for full activation checkpointing | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4556669155,
"node_id": "LA_kwDOC_bI7s8AAAABD5lA4w",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/cherry-pick",
"name": "cherry-pick",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | null | [] | null | null | 2022-11-21T22:37:16 | 2023-02-21T08:28:51 | 2022-11-22T03:16:50 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5473",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5473",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5473.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5473.patch",
"merged_at": "2022-11-22T03:16:50"
} | Signed-off-by: MaximumEntropy <[email protected]>
# What does this PR do ?
Adds a check that performs full activation checkpointing only if activation_checkpoint_num_layers > 0
**Collection**: NLP
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5473/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5473/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5471 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5471/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5471/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5471/events | https://github.com/NVIDIA/NeMo/pull/5471 | 1,458,584,086 | PR_kwDOC_bI7s5DY-cX | 5,471 | Radtts 1.13 plus | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 4418253970,
"node_id": "LA_kwDOC_bI7s8AAAABB1k0kg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/TTS",
"name": "TTS",
"color": "49AF28",
"default": false,
"description": ""
},
{
"id": 4556669155,
"node_id": "LA_kwDOC_bI7s8AAAABD5lA4w",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/cherry-pick",
"name": "cherry-pick",
"color": "ededed",
"default": false,
"description": null
}
] | closed | false | null | [] | null | null | 2022-11-21T21:12:06 | 2022-11-21T22:25:25 | 2022-11-21T22:21:47 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5471",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5471",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5471.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5471.patch",
"merged_at": "2022-11-21T22:21:47"
} | # What does this PR do ?
I finally managed to properly script only necessary part, and added a guard that works around the issue with auto-casted TorchScript not workin in non-Python environment (Triton)
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "borisfom",
"id": 14189615,
"node_id": "MDQ6VXNlcjE0MTg5NjE1",
"avatar_url": "https://avatars.githubusercontent.com/u/14189615?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/borisfom",
"html_url": "https://github.com/borisfom",
"followers_url": "https://api.github.com/users/borisfom/followers",
"following_url": "https://api.github.com/users/borisfom/following{/other_user}",
"gists_url": "https://api.github.com/users/borisfom/gists{/gist_id}",
"starred_url": "https://api.github.com/users/borisfom/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/borisfom/subscriptions",
"organizations_url": "https://api.github.com/users/borisfom/orgs",
"repos_url": "https://api.github.com/users/borisfom/repos",
"events_url": "https://api.github.com/users/borisfom/events{/privacy}",
"received_events_url": "https://api.github.com/users/borisfom/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5471/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5471/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5470 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5470/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5470/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5470/events | https://github.com/NVIDIA/NeMo/pull/5470 | 1,458,560,121 | PR_kwDOC_bI7s5DY5IN | 5,470 | Add num layers check for full activation checkpointing | {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2022-11-21T20:53:46 | 2022-11-21T22:36:35 | 2022-11-21T22:36:32 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5470",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5470",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5470.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5470.patch",
"merged_at": "2022-11-21T22:36:32"
} | Signed-off-by: MaximumEntropy <[email protected]>
# What does this PR do ?
Adds a check that performs full activation checkpointing only if activation_checkpoint_num_layers > 0
**Collection**: NLP
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "MaximumEntropy",
"id": 9114321,
"node_id": "MDQ6VXNlcjkxMTQzMjE=",
"avatar_url": "https://avatars.githubusercontent.com/u/9114321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MaximumEntropy",
"html_url": "https://github.com/MaximumEntropy",
"followers_url": "https://api.github.com/users/MaximumEntropy/followers",
"following_url": "https://api.github.com/users/MaximumEntropy/following{/other_user}",
"gists_url": "https://api.github.com/users/MaximumEntropy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/MaximumEntropy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/MaximumEntropy/subscriptions",
"organizations_url": "https://api.github.com/users/MaximumEntropy/orgs",
"repos_url": "https://api.github.com/users/MaximumEntropy/repos",
"events_url": "https://api.github.com/users/MaximumEntropy/events{/privacy}",
"received_events_url": "https://api.github.com/users/MaximumEntropy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5470/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5470/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/5468 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5468/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5468/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5468/events | https://github.com/NVIDIA/NeMo/issues/5468 | 1,457,922,044 | I_kwDOC_bI7s5W5h_8 | 5,468 | ZeroDivisionError: integer division or modulo by zero || Ambernet fine-tuning | {
"login": "Thayabaran",
"id": 46021275,
"node_id": "MDQ6VXNlcjQ2MDIxMjc1",
"avatar_url": "https://avatars.githubusercontent.com/u/46021275?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Thayabaran",
"html_url": "https://github.com/Thayabaran",
"followers_url": "https://api.github.com/users/Thayabaran/followers",
"following_url": "https://api.github.com/users/Thayabaran/following{/other_user}",
"gists_url": "https://api.github.com/users/Thayabaran/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Thayabaran/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Thayabaran/subscriptions",
"organizations_url": "https://api.github.com/users/Thayabaran/orgs",
"repos_url": "https://api.github.com/users/Thayabaran/repos",
"events_url": "https://api.github.com/users/Thayabaran/events{/privacy}",
"received_events_url": "https://api.github.com/users/Thayabaran/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1485815557,
"node_id": "MDU6TGFiZWwxNDg1ODE1NTU3",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
}
] | closed | false | {
"login": "fayejf",
"id": 36722593,
"node_id": "MDQ6VXNlcjM2NzIyNTkz",
"avatar_url": "https://avatars.githubusercontent.com/u/36722593?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/fayejf",
"html_url": "https://github.com/fayejf",
"followers_url": "https://api.github.com/users/fayejf/followers",
"following_url": "https://api.github.com/users/fayejf/following{/other_user}",
"gists_url": "https://api.github.com/users/fayejf/gists{/gist_id}",
"starred_url": "https://api.github.com/users/fayejf/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/fayejf/subscriptions",
"organizations_url": "https://api.github.com/users/fayejf/orgs",
"repos_url": "https://api.github.com/users/fayejf/repos",
"events_url": "https://api.github.com/users/fayejf/events{/privacy}",
"received_events_url": "https://api.github.com/users/fayejf/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "fayejf",
"id": 36722593,
"node_id": "MDQ6VXNlcjM2NzIyNTkz",
"avatar_url": "https://avatars.githubusercontent.com/u/36722593?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/fayejf",
"html_url": "https://github.com/fayejf",
"followers_url": "https://api.github.com/users/fayejf/followers",
"following_url": "https://api.github.com/users/fayejf/following{/other_user}",
"gists_url": "https://api.github.com/users/fayejf/gists{/gist_id}",
"starred_url": "https://api.github.com/users/fayejf/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/fayejf/subscriptions",
"organizations_url": "https://api.github.com/users/fayejf/orgs",
"repos_url": "https://api.github.com/users/fayejf/repos",
"events_url": "https://api.github.com/users/fayejf/events{/privacy}",
"received_events_url": "https://api.github.com/users/fayejf/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2022-11-21T13:08:42 | 2024-04-17T07:05:18 | 2022-12-05T08:19:46 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
I am trying to fine-tune the pre-trained ambernet-based language identification using [Mozzila Common Voice dataset ](https://commonvoice.mozilla.org/) by freezing the encoder and fine-tune only the classifier module.
` /NeMo/NeMo_1.12.0/ambernet/16d76d1aa14af9b770fabdac12fcf8e0/ambernet.nemo`
I have `65226 files loaded accounting to # 4 labels` train dataset and `774 files loaded accounting to # 4 labels` validation dataset.
I am using the default training [configuration ](https://github.com/NVIDIA/NeMo/tree/main/examples/asr/conf/lang_id) with below changes:
```
init_from_nemo_model:
model:
path: ./.cache/torch/NeMo/NeMo_1.12.0/ambernet/16d76d1aa14af9b770fabdac12fcf8e0/ambernet.nemo
decoder:
_target_: nemo.collections.asr.modules.SpeakerDecoder
feat_in: *enc_feat_out
num_classes: 4
pool_mode: 'xvector' # 'attention'
emb_sizes: 512
```
**Expected behavior**
All the parameters are loaded correctly on the pre-trained model and the sanity check are passed. The training starts good and it is expected to proceed with iterations/epochs to fine-tune the model instead, I am getting the below error during the training.
```
GPU available: True (cuda), used: True
TPU available: False, using: 0 TPU cores
IPU available: False, using: 0 IPUs
HPU available: False, using: 0 HPUs
`Trainer(val_check_interval=1.0)` was configured so validation will run at the end of the training epoch..
[NeMo I 2022-11-21 14:03:57 collections:297] Filtered duration for loading collection is 0.000000.
[NeMo I 2022-11-21 14:03:57 collections:301] # 65226 files loaded accounting to # 4 labels
[NeMo W 2022-11-21 14:03:57 label_models:194] Total number of 4 found in all the manifest files.
[NeMo I 2022-11-21 14:03:57 collections:297] Filtered duration for loading collection is 0.000000.
[NeMo I 2022-11-21 14:03:57 collections:301] # 65226 files loaded accounting to # 4 labels
[NeMo I 2022-11-21 14:03:57 collections:297] Filtered duration for loading collection is 0.000000.
[NeMo I 2022-11-21 14:03:57 collections:301] # 774 files loaded accounting to # 4 labels
[NeMo I 2022-11-21 14:03:57 collections:297] Filtered duration for loading collection is 0.000000.
[NeMo I 2022-11-21 14:03:57 collections:301] # 774 files loaded accounting to # 4 labels
[NeMo I 2022-11-21 14:03:57 cross_entropy:55] Weighted Cross Entropy loss with weight tensor([0.7663, 1.0014, 1.3230, 1.0631])
[NeMo W 2022-11-21 14:03:57 nemo_logging:349] /home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/torchmetrics/utilities/prints.py:36: UserWarning: Torchmetrics v0.9 introduced a new argument class property called `full_state_update` that has
not been set for this class (TopKClassificationAccuracy). The property determines if `update` by
default needs access to the full metric state. If this is not the case, significant speedups can be
achieved and we recommend setting this to `False`.
We provide an checking function
`from torchmetrics.utilities import check_forward_full_state_property`
that can be used to check if the `full_state_update=True` (old and potential slower behaviour,
default for now) or if `full_state_update=False` can be used safely.
warnings.warn(*args, **kwargs)
[NeMo I 2022-11-21 14:03:57 features:225] PADDING: 16
Model encoder has been frozen
LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [7]
[NeMo I 2022-11-21 14:03:59 modelPT:597] Optimizer config = SGD (
Parameter Group 0
dampening: 0
differentiable: False
foreach: None
lr: 0.001
maximize: False
momentum: 0
nesterov: False
weight_decay: 0.001
)
[NeMo I 2022-11-21 14:03:59 lr_scheduler:910] Scheduler "<nemo.core.optim.lr_scheduler.CosineAnnealing object at 0x7fd62f330a90>"
will be used during training (effective maximum steps = 25500) -
Parameters :
(warmup_ratio: 0.1
min_lr: 0.0001
max_steps: 25500
)
| Name | Type | Params
------------------------------------------------------------------------
0 | loss | CrossEntropyLoss | 0
1 | eval_loss | CrossEntropyLoss | 0
2 | _accuracy | TopKClassificationAccuracy | 0
3 | preprocessor | AudioToMelSpectrogramPreprocessor | 0
4 | encoder | ConvASREncoder | 19.4 M
5 | decoder | SpeakerDecoder | 3.1 M
6 | _macro_accuracy | Accuracy | 0
7 | spec_augmentation | SpectrogramAugmentation | 0
------------------------------------------------------------------------
3.1 M Trainable params
19.4 M Non-trainable params
22.5 M Total params
90.009 Total estimated model params size (MB)
Sanity Checking: 0it [00:00, ?it/s]
Sanity Checking: 0%| | 0/2 [00:00<?, ?it/s]
Sanity Checking DataLoader 0: 0%| | 0/2 [00:00<?, ?it/s]
Sanity Checking DataLoader 0: 50%|█████ | 1/2 [00:01<00:01, 1.00s/it]
Sanity Checking DataLoader 0: 100%|██████████| 2/2 [00:01<00:00, 1.40it/s]
Training: 0it [00:00, ?it/s]
Training: 0%| | 0/517 [00:00<?, ?it/s]
Epoch 0: 0%| | 0/517 [00:00<?, ?it/s] [NeMo W 2022-11-21 14:04:03 nemo_logging:349] /home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/trainer/connectors/logger_connector/result.py:231: UserWarning: You called `self.log('global_step', ...)` in your `training_step` but the value needs to be floating point. Converting it to torch.float32.
warning_cache.warn(
Epoch 0: 0%| | 1/517 [00:01<16:14, 1.89s/it]
Epoch 0: 0%| | 1/517 [00:01<16:14, 1.89s/it, loss=1.67]
Epoch 0: 0%| | 2/517 [00:02<09:49, 1.14s/it, loss=1.67]
Epoch 0: 0%| | 2/517 [00:02<09:49, 1.14s/it, loss=1.63]
Epoch 0: 1%| | 3/517 [00:02<07:38, 1.12it/s, loss=1.63]
Epoch 0: 1%| | 3/517 [00:02<07:38, 1.12it/s, loss=1.66]
Epoch 0: 1%| | 4/517 [00:03<06:33, 1.30it/s, loss=1.66]
Epoch 0: 1%| | 4/517 [00:03<06:33, 1.30it/s, loss=1.63]
Epoch 0: 1%| | 5/517 [00:03<05:51, 1.46it/s, loss=1.63]
Epoch 0: 1%| | 5/517 [00:03<05:51, 1.46it/s, loss=1.63]
Epoch 0: 1%| | 6/517 [00:03<05:26, 1.57it/s, loss=1.63]
Epoch 0: 1%| | 6/517 [00:03<05:26, 1.57it/s, loss=1.63]
Epoch 0: 1%|▏ | 7/517 [00:04<05:07, 1.66it/s, loss=1.63]
Epoch 0: 1%|▏ | 7/517 [00:04<05:07, 1.66it/s, loss=1.62]
Epoch 0: 2%|▏ | 8/517 [00:04<04:52, 1.74it/s, loss=1.62]
Epoch 0: 2%|▏ | 8/517 [00:04<04:52, 1.74it/s, loss=1.62]
Epoch 0: 2%|▏ | 9/517 [00:04<04:41, 1.81it/s, loss=1.62]
Epoch 0: 2%|▏ | 9/517 [00:04<04:41, 1.81it/s, loss=1.63]
Epoch 0: 2%|▏ | 10/517 [00:05<04:30, 1.87it/s, loss=1.63]
Epoch 0: 2%|▏ | 10/517 [00:05<04:30, 1.87it/s, loss=1.63]
Epoch 0: 2%|▏ | 11/517 [00:05<04:22, 1.93it/s, loss=1.63]
Epoch 0: 2%|▏ | 11/517 [00:05<04:22, 1.93it/s, loss=1.63]
Epoch 0: 2%|▏ | 12/517 [00:06<04:31, 1.86it/s, loss=1.63]
Epoch 0: 2%|▏ | 12/517 [00:06<04:31, 1.86it/s, loss=1.64]
Epoch 0: 3%|▎ | 13/517 [00:06<04:23, 1.91it/s, loss=1.64]
Epoch 0: 3%|▎ | 13/517 [00:06<04:23, 1.91it/s, loss=1.63]
Epoch 0: 3%|▎ | 14/517 [00:07<04:18, 1.95it/s, loss=1.63]
Epoch 0: 3%|▎ | 14/517 [00:07<04:18, 1.95it/s, loss=1.64]
Epoch 0: 3%|▎ | 15/517 [00:07<04:12, 1.99it/s, loss=1.64]
Epoch 0: 3%|▎ | 15/517 [00:07<04:12, 1.99it/s, loss=1.63]
Epoch 0: 3%|▎ | 16/517 [00:07<04:07, 2.02it/s, loss=1.63]
Epoch 0: 3%|▎ | 16/517 [00:07<04:08, 2.02it/s, loss=1.63]
Epoch 0: 3%|▎ | 17/517 [00:08<04:05, 2.04it/s, loss=1.63]
Epoch 0: 3%|▎ | 17/517 [00:08<04:05, 2.04it/s, loss=1.63]
Epoch 0: 3%|▎ | 18/517 [00:08<04:02, 2.06it/s, loss=1.63]
Epoch 0: 3%|▎ | 18/517 [00:08<04:02, 2.06it/s, loss=1.63]
Epoch 0: 4%|▎ | 19/517 [00:10<04:30, 1.84it/s, loss=1.63]
Epoch 0: 4%|▎ | 19/517 [00:10<04:30, 1.84it/s, loss=1.64]
Epoch 0: 4%|▍ | 20/517 [00:10<04:26, 1.87it/s, loss=1.64]
Epoch 0: 4%|▍ | 20/517 [00:10<04:26, 1.87it/s, loss=1.64]
Epoch 0: 4%|▍ | 21/517 [00:11<04:27, 1.85it/s, loss=1.64]
Epoch 0: 4%|▍ | 21/517 [00:11<04:27, 1.85it/s, loss=1.63]
Epoch 0: 4%|▍ | 22/517 [00:11<04:27, 1.85it/s, loss=1.63]
Epoch 0: 4%|▍ | 22/517 [00:11<04:27, 1.85it/s, loss=1.63]
Epoch 0: 4%|▍ | 23/517 [00:12<04:23, 1.87it/s, loss=1.63]
Epoch 0: 4%|▍ | 23/517 [00:12<04:23, 1.87it/s, loss=1.62]
Epoch 0: 5%|▍ | 24/517 [00:12<04:19, 1.90it/s, loss=1.62]
Epoch 0: 5%|▍ | 24/517 [00:12<04:19, 1.90it/s, loss=1.62]
Epoch 0: 5%|▍ | 25/517 [00:13<04:15, 1.92it/s, loss=1.62]
Epoch 0: 5%|▍ | 25/517 [00:13<04:15, 1.92it/s, loss=1.62]
Epoch 0: 5%|▌ | 26/517 [00:13<04:12, 1.94it/s, loss=1.62]
Epoch 0: 5%|▌ | 26/517 [00:13<04:12, 1.94it/s, loss=1.62]
Epoch 0: 5%|▌ | 27/517 [00:13<04:09, 1.96it/s, loss=1.62]
Epoch 0: 5%|▌ | 27/517 [00:13<04:09, 1.96it/s, loss=1.63]
Epoch 0: 5%|▌ | 28/517 [00:14<04:07, 1.98it/s, loss=1.63]
Epoch 0: 5%|▌ | 28/517 [00:14<04:07, 1.98it/s, loss=1.63]
Epoch 0: 6%|▌ | 29/517 [00:14<04:04, 2.00it/s, loss=1.63]
Epoch 0: 6%|▌ | 29/517 [00:14<04:04, 2.00it/s, loss=1.63]
Epoch 0: 6%|▌ | 30/517 [00:14<04:01, 2.02it/s, loss=1.63]
Epoch 0: 6%|▌ | 30/517 [00:14<04:01, 2.02it/s, loss=1.63]
Epoch 0: 6%|▌ | 31/517 [00:15<03:59, 2.03it/s, loss=1.63]
Epoch 0: 6%|▌ | 31/517 [00:15<03:59, 2.03it/s, loss=1.62]
Epoch 0: 6%|▌ | 32/517 [00:15<03:56, 2.05it/s, loss=1.62]
Epoch 0: 6%|▌ | 32/517 [00:15<03:56, 2.05it/s, loss=1.62]
Epoch 0: 6%|▋ | 33/517 [00:15<03:54, 2.06it/s, loss=1.62]
Epoch 0: 6%|▋ | 33/517 [00:15<03:54, 2.06it/s, loss=1.62]
Epoch 0: 7%|▋ | 34/517 [00:16<03:52, 2.08it/s, loss=1.62]
Epoch 0: 7%|▋ | 34/517 [00:16<03:52, 2.08it/s, loss=1.62]
Epoch 0: 7%|▋ | 35/517 [00:16<03:51, 2.09it/s, loss=1.62]
Epoch 0: 7%|▋ | 35/517 [00:16<03:51, 2.09it/s, loss=1.62]
Epoch 0: 7%|▋ | 36/517 [00:17<03:49, 2.10it/s, loss=1.62]
Epoch 0: 7%|▋ | 36/517 [00:17<03:49, 2.10it/s, loss=1.62]
Epoch 0: 7%|▋ | 37/517 [00:17<03:48, 2.10it/s, loss=1.62]
Epoch 0: 7%|▋ | 37/517 [00:17<03:48, 2.10it/s, loss=1.62]
Epoch 0: 7%|▋ | 38/517 [00:17<03:46, 2.12it/s, loss=1.62]
Epoch 0: 7%|▋ | 38/517 [00:17<03:46, 2.12it/s, loss=1.62]
Epoch 0: 8%|▊ | 39/517 [00:18<03:44, 2.13it/s, loss=1.62]
Epoch 0: 8%|▊ | 39/517 [00:18<03:44, 2.13it/s, loss=1.63]
Traceback (most recent call last):
File "/home/thayabaran.kathiresan/nemo_LID/TL_Scripts/TL_Nemo_Training.py", line 32, in <module>
trainer.fit(ambernet_model)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/trainer/trainer.py", line 696, in fit
self._call_and_handle_interrupt(
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/trainer/trainer.py", line 650, in _call_and_handle_interrupt
return trainer_fn(*args, **kwargs)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/trainer/trainer.py", line 735, in _fit_impl
results = self._run(model, ckpt_path=self.ckpt_path)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/trainer/trainer.py", line 1166, in _run
results = self._run_stage()
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/trainer/trainer.py", line 1252, in _run_stage
return self._run_train()
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/trainer/trainer.py", line 1283, in _run_train
self.fit_loop.run()
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/loops/loop.py", line 200, in run
self.advance(*args, **kwargs)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/loops/fit_loop.py", line 271, in advance
self._outputs = self.epoch_loop.run(self._data_fetcher)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/loops/loop.py", line 200, in run
self.advance(*args, **kwargs)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/loops/epoch/training_epoch_loop.py", line 174, in advance
batch = next(data_fetcher)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/utilities/fetching.py", line 184, in __next__
return self.fetching_function()
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/utilities/fetching.py", line 263, in fetching_function
self._fetch_next_batch(self.dataloader_iter)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/utilities/fetching.py", line 277, in _fetch_next_batch
batch = next(iterator)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/trainer/supporters.py", line 557, in __next__
return self.request_next_batch(self.loader_iters)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/trainer/supporters.py", line 569, in request_next_batch
return apply_to_collection(loader_iters, Iterator, next)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/pytorch_lightning/utilities/apply_func.py", line 99, in apply_to_collection
return function(data, *args, **kwargs)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 628, in __next__
data = self._next_data()
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1333, in _next_data
return self._process_data(data)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1359, in _process_data
data.reraise()
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/torch/_utils.py", line 543, in reraise
raise exception
ZeroDivisionError: Caught ZeroDivisionError in DataLoader worker process 7.
Original Traceback (most recent call last):
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/torch/utils/data/_utils/worker.py", line 302, in _worker_loop
data = fetcher.fetch(index)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/torch/utils/data/_utils/fetch.py", line 61, in fetch
return self.collate_fn(data)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/nemo/collections/asr/data/audio_to_label.py", line 436, in fixed_seq_collate_fn
return _fixed_seq_collate_fn(self, batch)
File "/home/thayabaran.kathiresan/anaconda3/envs/nemo_LID/lib/python3.9/site-packages/nemo/collections/asr/data/audio_to_label.py", line 133, in _fixed_seq_collate_fn
repeat = fixed_length // sig_len
ZeroDivisionError: integer division or modulo by zero
```
**Environment overview**
- Method of NeMo install: [pip install or from source].
```
apt-get update && apt-get install -y libsndfile1 ffmpeg
pip install Cython
pip install nemo_toolkit[all]
```
**Additional Information**
- Train and Validation manifest have been checked for any bad-audio/corrupted files but, the data set is clean and according to the input standard
- The training is crashing at different iterations every time but, never completed 1 Epoch
- Tried running in CPU and it produces the same crash.
| {
"login": "Thayabaran",
"id": 46021275,
"node_id": "MDQ6VXNlcjQ2MDIxMjc1",
"avatar_url": "https://avatars.githubusercontent.com/u/46021275?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Thayabaran",
"html_url": "https://github.com/Thayabaran",
"followers_url": "https://api.github.com/users/Thayabaran/followers",
"following_url": "https://api.github.com/users/Thayabaran/following{/other_user}",
"gists_url": "https://api.github.com/users/Thayabaran/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Thayabaran/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Thayabaran/subscriptions",
"organizations_url": "https://api.github.com/users/Thayabaran/orgs",
"repos_url": "https://api.github.com/users/Thayabaran/repos",
"events_url": "https://api.github.com/users/Thayabaran/events{/privacy}",
"received_events_url": "https://api.github.com/users/Thayabaran/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5468/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5468/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/5467 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/5467/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/5467/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/5467/events | https://github.com/NVIDIA/NeMo/pull/5467 | 1,457,894,155 | PR_kwDOC_bI7s5DWpJ4 | 5,467 | ALiBi Positional Embeddings | {
"login": "michalivne",
"id": 7453913,
"node_id": "MDQ6VXNlcjc0NTM5MTM=",
"avatar_url": "https://avatars.githubusercontent.com/u/7453913?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/michalivne",
"html_url": "https://github.com/michalivne",
"followers_url": "https://api.github.com/users/michalivne/followers",
"following_url": "https://api.github.com/users/michalivne/following{/other_user}",
"gists_url": "https://api.github.com/users/michalivne/gists{/gist_id}",
"starred_url": "https://api.github.com/users/michalivne/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/michalivne/subscriptions",
"organizations_url": "https://api.github.com/users/michalivne/orgs",
"repos_url": "https://api.github.com/users/michalivne/repos",
"events_url": "https://api.github.com/users/michalivne/events{/privacy}",
"received_events_url": "https://api.github.com/users/michalivne/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1719393562,
"node_id": "MDU6TGFiZWwxNzE5MzkzNTYy",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/feature",
"name": "feature",
"color": "68F3F0",
"default": false,
"description": "request/PR for a new feature"
},
{
"id": 4811886210,
"node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP",
"name": "NLP",
"color": "CBF09A",
"default": false,
"description": ""
},
{
"id": 4840216855,
"node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI",
"name": "CI",
"color": "1D76DB",
"default": false,
"description": ""
}
] | closed | false | null | [] | null | null | 2022-11-21T12:48:33 | 2023-01-05T23:28:42 | 2023-01-04T23:03:15 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/5467",
"html_url": "https://github.com/NVIDIA/NeMo/pull/5467",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/5467.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/5467.patch",
"merged_at": "2023-01-04T23:03:15"
} | Adding support of ALiBi positional embeddings to Megatron
https://arxiv.org/abs/2108.12409
# What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
**Collection**: [Note which collection this PR will affect]
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
* You can potentially add a usage example below
```python
# Add a code snippet demonstrating how to use this
```
# Before your PR is "Ready for review"
**Pre checks**:
- [ ] Make sure you read and followed [Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md)
- [ ] Did you write any new necessary tests?
- [ ] Did you add or update any necessary documentation?
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
- [ ] Reviewer: Does the PR have correct import guards for all optional libraries?
**PR Type**:
- [ ] New Feature
- [ ] Bugfix
- [ ] Documentation
If you haven't finished some of the above items you can still open "Draft" PR.
## Who can review?
Anyone in the community is free to review the PR once the checks have passed.
[Contributor guidelines](https://github.com/NVIDIA/NeMo/blob/main/CONTRIBUTING.md) contains specific people who can review PRs to various areas.
# Additional Information
* Related to # (issue)
| {
"login": "michalivne",
"id": 7453913,
"node_id": "MDQ6VXNlcjc0NTM5MTM=",
"avatar_url": "https://avatars.githubusercontent.com/u/7453913?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/michalivne",
"html_url": "https://github.com/michalivne",
"followers_url": "https://api.github.com/users/michalivne/followers",
"following_url": "https://api.github.com/users/michalivne/following{/other_user}",
"gists_url": "https://api.github.com/users/michalivne/gists{/gist_id}",
"starred_url": "https://api.github.com/users/michalivne/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/michalivne/subscriptions",
"organizations_url": "https://api.github.com/users/michalivne/orgs",
"repos_url": "https://api.github.com/users/michalivne/repos",
"events_url": "https://api.github.com/users/michalivne/events{/privacy}",
"received_events_url": "https://api.github.com/users/michalivne/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/5467/reactions",
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/5467/timeline | null | null | true |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.