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/6979 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6979/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6979/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6979/events | https://github.com/NVIDIA/NeMo/pull/6979 | 1,790,020,919 | PR_kwDOC_bI7s5UuqPR | 6,979 | Remove `compute_on_step` from metrics | {
"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": 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": ""
},
{
"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": 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 | 2023-07-05T17:52:13 | 2023-07-05T21:11:54 | 2023-07-05T21:11:49 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6979",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6979",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6979.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6979.patch",
"merged_at": "2023-07-05T21:11:49"
} | # What does this PR do ?
Remove deprecated arg compute_on_step from torchmetrics
**Collection**: [ASR, NLP]
# 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/6979/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/6979/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6978 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6978/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6978/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6978/events | https://github.com/NVIDIA/NeMo/pull/6978 | 1,790,020,747 | PR_kwDOC_bI7s5UuqNG | 6,978 | Implemented generic kv-pair setting of export_config from args | {
"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": 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 | 2023-07-05T17:52:07 | 2023-07-06T21:30:24 | 2023-07-06T21:30:23 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6978",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6978",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6978.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6978.patch",
"merged_at": null
} | # What does this PR do ?
This will be a basis for several export extensions that involve user input.
As an example, I have changed handling of --cache_support arg in export.py to use the new method.
**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/6978/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/6978/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6977 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6977/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6977/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6977/events | https://github.com/NVIDIA/NeMo/issues/6977 | 1,789,946,298 | I_kwDOC_bI7s5qsGm6 | 6,977 | Unexpected keyword arguments: `compute_on_step` File "/home/muizz/nemo_workspace/Neural_net.py", line 9, in <module> asr_model = nemo_asr.models.EncDecCTCModel.from_pretrained(model_name="QuartzNet15x5Base-En") ValueError: Unexpected keyword arguments: `compute_on_step` | {
"login": "muizzhaque",
"id": 113048786,
"node_id": "U_kgDOBrz80g",
"avatar_url": "https://avatars.githubusercontent.com/u/113048786?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/muizzhaque",
"html_url": "https://github.com/muizzhaque",
"followers_url": "https://api.github.com/users/muizzhaque/followers",
"following_url": "https://api.github.com/users/muizzhaque/following{/other_user}",
"gists_url": "https://api.github.com/users/muizzhaque/gists{/gist_id}",
"starred_url": "https://api.github.com/users/muizzhaque/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/muizzhaque/subscriptions",
"organizations_url": "https://api.github.com/users/muizzhaque/orgs",
"repos_url": "https://api.github.com/users/muizzhaque/repos",
"events_url": "https://api.github.com/users/muizzhaque/events{/privacy}",
"received_events_url": "https://api.github.com/users/muizzhaque/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 | 2023-07-05T16:57:38 | 2024-07-13T05:52:25 | 2023-07-05T21:12:10 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null |
I've been trying to transcribe a wav file of mono channel and rated a 16khz but this error shows up all the time
below is my code:
import nemo.collections.asr as nemo_asr
asr_model = nemo_asr.models.EncDecCTCModel.from_pretrained(model_name="QuartzNet15x5Base-En")
emb = asr_model.transcribe('./right_channel.wav')
print(emb)
error:
Unexpected keyword arguments: `compute_on_step` File "/home/muizz/nemo_workspace/Neural_net.py", line 9, in <module> asr_model = nemo_asr.models.EncDecCTCModel.from_pretrained(model_name="QuartzNet15x5Base-En") ValueError: Unexpected keyword arguments: `compute_on_step` | {
"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/6977/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/6977/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6976 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6976/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6976/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6976/events | https://github.com/NVIDIA/NeMo/issues/6976 | 1,789,682,267 | I_kwDOC_bI7s5qrGJb | 6,976 | char_model = nemo_asr.models.ASRModel.from_pretrained("stt_en_quartznet15x5", map_location='cpu').And I have just been running the fine-tuning tutorial you just put out. I dont know how to fix this . And im running it on google colab .Please help | {
"login": "Hid-Gh",
"id": 78357724,
"node_id": "MDQ6VXNlcjc4MzU3NzI0",
"avatar_url": "https://avatars.githubusercontent.com/u/78357724?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Hid-Gh",
"html_url": "https://github.com/Hid-Gh",
"followers_url": "https://api.github.com/users/Hid-Gh/followers",
"following_url": "https://api.github.com/users/Hid-Gh/following{/other_user}",
"gists_url": "https://api.github.com/users/Hid-Gh/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Hid-Gh/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Hid-Gh/subscriptions",
"organizations_url": "https://api.github.com/users/Hid-Gh/orgs",
"repos_url": "https://api.github.com/users/Hid-Gh/repos",
"events_url": "https://api.github.com/users/Hid-Gh/events{/privacy}",
"received_events_url": "https://api.github.com/users/Hid-Gh/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 | 2023-07-05T14:25:11 | 2023-07-05T21:12:35 | 2023-07-05T21:12:35 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | [NeMo I 2023-07-05 14:14:49 cloud:68] Downloading from: https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_en_quartznet15x5/versions/1.0.0rc1/files/stt_en_quartznet15x5.nemo to /root/.cache/torch/NeMo/NeMo_1.20.0rc0/stt_en_quartznet15x5/16661021d16e679bdfd97a2a03944c49/stt_en_quartznet15x5.nemo
[NeMo I 2023-07-05 14:14:53 common:913] Instantiating model from pre-trained checkpoint
[NeMo W 2023-07-05 14:14:56 modelPT:161] If you intend to do training or fine-tuning, please call the ModelPT.setup_training_data() method and provide a valid configuration file to setup the train data loader.
Train config :
manifest_filepath: /data2/voices/train_1k.json
sample_rate: 16000
labels:
- ' '[NeMo I 2023-07-05 14:14:49 cloud:68] Downloading from: https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_en_quartznet15x5/versions/1.0.0rc1/files/stt_en_quartznet15x5.nemo to /root/.cache/torch/NeMo/NeMo_1.20.0rc0/stt_en_quartznet15x5/16661021d16e679bdfd97a2a03944c49/stt_en_quartznet15x5.nemo
[NeMo I 2023-07-05 14:14:53 common:913] Instantiating model from pre-trained checkpoint
[NeMo W 2023-07-05 14:14:56 modelPT:161] If you intend to do training or fine-tuning, please call the ModelPT.setup_training_data() method and provide a valid configuration file to setup the train data loader.
Train config :
manifest_filepath: /data2/voices/train_1k.json
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
- ''''
batch_size: 32
trim_silence: true
max_duration: 16.7
shuffle: true
is_tarred: false
tarred_audio_filepaths: /asr_set_1.2/train/train_{0..1023}.tar
num_workers: 20
[NeMo W 2023-07-05 14:14:56 modelPT:168] If you intend to do validation, please call the ModelPT.setup_validation_data() or ModelPT.setup_multiple_validation_data() method and provide a valid configuration file to setup the validation data loader(s).
Validation config :
manifest_filepath: /data2/voices/train_1k_samp.json
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
- ''''
batch_size: 32
shuffle: false
[NeMo I 2023-07-05 14:14:56 features:289] PADDING: 16
[NeMo W 2023-07-05 14:14:57 modelPT:161] If you intend to do training or fine-tuning, please call the ModelPT.setup_training_data() method and provide a valid configuration file to setup the train data loader.
Train config :
manifest_filepath: /data2/voices/train_1k.json
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
- ''''
batch_size: 32
trim_silence: true
max_duration: 16.7
shuffle: true
is_tarred: false
tarred_audio_filepaths: /asr_set_1.2/train/train_{0..1023}.tar
num_workers: 20
[NeMo W 2023-07-05 14:14:57 modelPT:168] If you intend to do validation, please call the ModelPT.setup_validation_data() or ModelPT.setup_multiple_validation_data() method and provide a valid configuration file to setup the validation data loader(s).
Validation config :
manifest_filepath: /data2/voices/train_1k_samp.json
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
- ''''
batch_size: 32
shuffle: false
[NeMo I 2023-07-05 14:14:57 features:289] PADDING: 16
[NeMo E 2023-07-05 14:14:58 common:506] Model instantiation failed!
Target class: nemo.collections.asr.models.ctc_models.EncDecCTCModel
Error(s): Unexpected keyword arguments: `compute_on_step`
Traceback (most recent call last):
File "/usr/local/lib/python3.10/dist-packages/nemo/core/classes/common.py", line 485, in from_config_dict
instance = imported_cls(cfg=config, trainer=trainer)
File "/usr/local/lib/python3.10/dist-packages/nemo/collections/asr/models/ctc_models.py", line 96, in __init__
self._wer = WER(
File "/usr/local/lib/python3.10/dist-packages/nemo/collections/asr/metrics/wer.py", line 1128, in __init__
super().__init__(dist_sync_on_step=dist_sync_on_step, compute_on_step=False)
File "/usr/local/lib/python3.10/dist-packages/torchmetrics/metric.py", line 145, in __init__
raise ValueError(f"Unexpected keyword arguments: {', '.join(kwargs_)}")
ValueError: Unexpected keyword arguments: `compute_on_step`
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
[<ipython-input-34-a0f5a44992b6>](https://0pzek2a4z9i-496ff2e9c6d22116-0-colab.googleusercontent.com/outputframe.html?vrz=colab-20230629-060050-RC00_544292008#) in <cell line: 1>()
----> 1 char_model = nemo_asr.models.ASRModel.from_pretrained("stt_en_quartznet15x5", map_location='cpu')
8 frames
[/usr/local/lib/python3.10/dist-packages/torchmetrics/metric.py](https://0pzek2a4z9i-496ff2e9c6d22116-0-colab.googleusercontent.com/outputframe.html?vrz=colab-20230629-060050-RC00_544292008#) in __init__(self, **kwargs)
143 if kwargs:
144 kwargs_ = [f"`{a}`" for a in sorted(kwargs)]
--> 145 raise ValueError(f"Unexpected keyword arguments: {', '.join(kwargs_)}")
146
147 # initialize
ValueError: Unexpected keyword arguments: `compute_on_step`
- 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
- ''''
batch_size: 32
trim_silence: true
max_duration: 16.7
shuffle: true
is_tarred: false
tarred_audio_filepaths: /asr_set_1.2/train/train_{0..1023}.tar
num_workers: 20
[NeMo W 2023-07-05 14:14:56 modelPT:168] If you intend to do validation, please call the ModelPT.setup_validation_data() or ModelPT.setup_multiple_validation_data() method and provide a valid configuration file to setup the validation data loader(s).
Validation config :
manifest_filepath: /data2/voices/train_1k_samp.json
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
- ''''
batch_size: 32
shuffle: false
[NeMo I 2023-07-05 14:14:56 features:289] PADDING: 16
[NeMo W 2023-07-05 14:14:57 modelPT:161] If you intend to do training or fine-tuning, please call the ModelPT.setup_training_data() method and provide a valid configuration file to setup the train data loader.
Train config :
manifest_filepath: /data2/voices/train_1k.json
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
- ''''
batch_size: 32
trim_silence: true
max_duration: 16.7
shuffle: true
is_tarred: false
tarred_audio_filepaths: /asr_set_1.2/train/train_{0..1023}.tar
num_workers: 20
[NeMo W 2023-07-05 14:14:57 modelPT:168] If you intend to do validation, please call the ModelPT.setup_validation_data() or ModelPT.setup_multiple_validation_data() method and provide a valid configuration file to setup the validation data loader(s).
Validation config :
manifest_filepath: /data2/voices/train_1k_samp.json
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
- ''''
batch_size: 32
shuffle: false
[NeMo I 2023-07-05 14:14:57 features:289] PADDING: 16
[NeMo E 2023-07-05 14:14:58 common:506] Model instantiation failed!
Target class: nemo.collections.asr.models.ctc_models.EncDecCTCModel
Error(s): Unexpected keyword arguments: `compute_on_step`
Traceback (most recent call last):
File "/usr/local/lib/python3.10/dist-packages/nemo/core/classes/common.py", line 485, in from_config_dict
instance = imported_cls(cfg=config, trainer=trainer)
File "/usr/local/lib/python3.10/dist-packages/nemo/collections/asr/models/ctc_models.py", line 96, in __init__
self._wer = WER(
File "/usr/local/lib/python3.10/dist-packages/nemo/collections/asr/metrics/wer.py", line 1128, in __init__
super().__init__(dist_sync_on_step=dist_sync_on_step, compute_on_step=False)
File "/usr/local/lib/python3.10/dist-packages/torchmetrics/metric.py", line 145, in __init__
raise ValueError(f"Unexpected keyword arguments: {', '.join(kwargs_)}")
ValueError: Unexpected keyword arguments: `compute_on_step`
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
[<ipython-input-34-a0f5a44992b6>](https://0pzek2a4z9i-496ff2e9c6d22116-0-colab.googleusercontent.com/outputframe.html?vrz=colab-20230629-060050-RC00_544292008#) in <cell line: 1>()
----> 1 char_model = nemo_asr.models.ASRModel.from_pretrained("stt_en_quartznet15x5", map_location='cpu')
8 frames
[/usr/local/lib/python3.10/dist-packages/torchmetrics/metric.py](https://0pzek2a4z9i-496ff2e9c6d22116-0-colab.googleusercontent.com/outputframe.html?vrz=colab-20230629-060050-RC00_544292008#) in __init__(self, **kwargs)
143 if kwargs:
144 kwargs_ = [f"`{a}`" for a in sorted(kwargs)]
--> 145 raise ValueError(f"Unexpected keyword arguments: {', '.join(kwargs_)}")
146
147 # initialize
ValueError: Unexpected keyword arguments: `compute_on_step` | {
"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/6976/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/6976/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6973 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6973/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6973/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6973/events | https://github.com/NVIDIA/NeMo/issues/6973 | 1,788,891,458 | I_kwDOC_bI7s5qoFFC | 6,973 | DER as the metric for the callback in MSDD train | {
"login": "SagyHarpazGong",
"id": 99274232,
"node_id": "U_kgDOBerN-A",
"avatar_url": "https://avatars.githubusercontent.com/u/99274232?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SagyHarpazGong",
"html_url": "https://github.com/SagyHarpazGong",
"followers_url": "https://api.github.com/users/SagyHarpazGong/followers",
"following_url": "https://api.github.com/users/SagyHarpazGong/following{/other_user}",
"gists_url": "https://api.github.com/users/SagyHarpazGong/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SagyHarpazGong/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SagyHarpazGong/subscriptions",
"organizations_url": "https://api.github.com/users/SagyHarpazGong/orgs",
"repos_url": "https://api.github.com/users/SagyHarpazGong/repos",
"events_url": "https://api.github.com/users/SagyHarpazGong/events{/privacy}",
"received_events_url": "https://api.github.com/users/SagyHarpazGong/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 | {
"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 | 2023-07-05T06:51:05 | 2023-09-01T01:45:35 | 2023-09-01T01:45:35 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | Hi.
I'm wondering if is possible to set DER as the saving metric, as I can see I only can set loss or acc_f1 as the saving metric.
please advise | {
"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/6973/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/6973/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6971 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6971/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6971/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6971/events | https://github.com/NVIDIA/NeMo/pull/6971 | 1,787,716,359 | PR_kwDOC_bI7s5Um1e7 | 6,971 | rnnt and char utils | {
"login": "karpnv",
"id": 1645775,
"node_id": "MDQ6VXNlcjE2NDU3NzU=",
"avatar_url": "https://avatars.githubusercontent.com/u/1645775?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/karpnv",
"html_url": "https://github.com/karpnv",
"followers_url": "https://api.github.com/users/karpnv/followers",
"following_url": "https://api.github.com/users/karpnv/following{/other_user}",
"gists_url": "https://api.github.com/users/karpnv/gists{/gist_id}",
"starred_url": "https://api.github.com/users/karpnv/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/karpnv/subscriptions",
"organizations_url": "https://api.github.com/users/karpnv/orgs",
"repos_url": "https://api.github.com/users/karpnv/repos",
"events_url": "https://api.github.com/users/karpnv/events{/privacy}",
"received_events_url": "https://api.github.com/users/karpnv/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2023-07-04T11:09:18 | 2023-07-14T09:24:15 | 2023-07-14T09:24:14 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6971",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6971",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6971.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6971.patch",
"merged_at": "2023-07-14T09:24:14"
} | # What does this PR do ?
fix ngram_merge using rnnt nemo model.
fix char level ngram creation.
**Collection**: ASR
# Changelog
- scripts/asr_language_modeling/ngram_lm/ngram_merge.py count perplexity of a test set using tokenizer from RNNT model
- scripts/asr_language_modeling/ngram_lm/kenlm_utils.py create word-level ngram model using char-level nemo model
```python
# Add a code snippet demonstrating how to use this
```
**PR Type**:
- [ V] Bugfix
| {
"login": "karpnv",
"id": 1645775,
"node_id": "MDQ6VXNlcjE2NDU3NzU=",
"avatar_url": "https://avatars.githubusercontent.com/u/1645775?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/karpnv",
"html_url": "https://github.com/karpnv",
"followers_url": "https://api.github.com/users/karpnv/followers",
"following_url": "https://api.github.com/users/karpnv/following{/other_user}",
"gists_url": "https://api.github.com/users/karpnv/gists{/gist_id}",
"starred_url": "https://api.github.com/users/karpnv/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/karpnv/subscriptions",
"organizations_url": "https://api.github.com/users/karpnv/orgs",
"repos_url": "https://api.github.com/users/karpnv/repos",
"events_url": "https://api.github.com/users/karpnv/events{/privacy}",
"received_events_url": "https://api.github.com/users/karpnv/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6971/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/6971/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6970 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6970/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6970/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6970/events | https://github.com/NVIDIA/NeMo/pull/6970 | 1,787,260,105 | PR_kwDOC_bI7s5UlS5l | 6,970 | [pre-commit.ci] pre-commit suggestions | {
"login": "pre-commit-ci[bot]",
"id": 66853113,
"node_id": "MDM6Qm90NjY4NTMxMTM=",
"avatar_url": "https://avatars.githubusercontent.com/in/68672?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/pre-commit-ci%5Bbot%5D",
"html_url": "https://github.com/apps/pre-commit-ci",
"followers_url": "https://api.github.com/users/pre-commit-ci%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pre-commit-ci%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/pre-commit-ci%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/pre-commit-ci%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/pre-commit-ci%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/pre-commit-ci%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/pre-commit-ci%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/pre-commit-ci%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/pre-commit-ci%5Bbot%5D/received_events",
"type": "Bot",
"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 | 2023-07-04T06:42:50 | 2023-07-27T01:47:04 | 2023-07-27T01:47:04 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6970",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6970",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6970.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6970.patch",
"merged_at": null
} | <!--pre-commit.ci start-->
updates:
- [github.com/pre-commit/pre-commit-hooks: v4.3.0 → v4.4.0](https://github.com/pre-commit/pre-commit-hooks/compare/v4.3.0...v4.4.0)
- [github.com/psf/black: 19.10b0 → 23.3.0](https://github.com/psf/black/compare/19.10b0...23.3.0)
<!--pre-commit.ci end--> | {
"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/6970/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/6970/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6969 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6969/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6969/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6969/events | https://github.com/NVIDIA/NeMo/pull/6969 | 1,786,944,344 | PR_kwDOC_bI7s5UkQps | 6,969 | st standalone model | {
"login": "AlexGrinch",
"id": 8689095,
"node_id": "MDQ6VXNlcjg2ODkwOTU=",
"avatar_url": "https://avatars.githubusercontent.com/u/8689095?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/AlexGrinch",
"html_url": "https://github.com/AlexGrinch",
"followers_url": "https://api.github.com/users/AlexGrinch/followers",
"following_url": "https://api.github.com/users/AlexGrinch/following{/other_user}",
"gists_url": "https://api.github.com/users/AlexGrinch/gists{/gist_id}",
"starred_url": "https://api.github.com/users/AlexGrinch/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/AlexGrinch/subscriptions",
"organizations_url": "https://api.github.com/users/AlexGrinch/orgs",
"repos_url": "https://api.github.com/users/AlexGrinch/repos",
"events_url": "https://api.github.com/users/AlexGrinch/events{/privacy}",
"received_events_url": "https://api.github.com/users/AlexGrinch/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": 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": "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 | 2023-07-04T00:02:48 | 2023-07-18T19:17:43 | 2023-07-18T19:17:42 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6969",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6969",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6969.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6969.patch",
"merged_at": "2023-07-18T19:17:42"
} | # What does this PR do ?
Adds support for ConformerEnc-TransformerDec speech translation (ST) model.
**Collection**: asr
# Changelog
- Added baseline ST model
- Tested to work with both yttm and SentencePiece (SP) tokenizers, SP is considered default
# 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/6969/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/6969/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6968 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6968/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6968/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6968/events | https://github.com/NVIDIA/NeMo/pull/6968 | 1,786,552,484 | PR_kwDOC_bI7s5Ui7sB | 6,968 | Stop at the stop token | {
"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 | 2023-07-03T17:14:22 | 2023-07-07T18:18:46 | 2023-07-07T18:18:45 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6968",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6968",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6968.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6968.patch",
"merged_at": "2023-07-07T18:18:45"
} | Detects the stop token and stop the generation once the stop token is detected. | {
"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
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6968/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/6968/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6967 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6967/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6967/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6967/events | https://github.com/NVIDIA/NeMo/issues/6967 | 1,786,043,441 | I_kwDOC_bI7s5qdNwx | 6,967 | Inverse Normalization has some issue | {
"login": "shreerakshaachlady",
"id": 126155170,
"node_id": "U_kgDOB4T5og",
"avatar_url": "https://avatars.githubusercontent.com/u/126155170?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/shreerakshaachlady",
"html_url": "https://github.com/shreerakshaachlady",
"followers_url": "https://api.github.com/users/shreerakshaachlady/followers",
"following_url": "https://api.github.com/users/shreerakshaachlady/following{/other_user}",
"gists_url": "https://api.github.com/users/shreerakshaachlady/gists{/gist_id}",
"starred_url": "https://api.github.com/users/shreerakshaachlady/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/shreerakshaachlady/subscriptions",
"organizations_url": "https://api.github.com/users/shreerakshaachlady/orgs",
"repos_url": "https://api.github.com/users/shreerakshaachlady/repos",
"events_url": "https://api.github.com/users/shreerakshaachlady/events{/privacy}",
"received_events_url": "https://api.github.com/users/shreerakshaachlady/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": "yzhang123",
"id": 4204271,
"node_id": "MDQ6VXNlcjQyMDQyNzE=",
"avatar_url": "https://avatars.githubusercontent.com/u/4204271?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/yzhang123",
"html_url": "https://github.com/yzhang123",
"followers_url": "https://api.github.com/users/yzhang123/followers",
"following_url": "https://api.github.com/users/yzhang123/following{/other_user}",
"gists_url": "https://api.github.com/users/yzhang123/gists{/gist_id}",
"starred_url": "https://api.github.com/users/yzhang123/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/yzhang123/subscriptions",
"organizations_url": "https://api.github.com/users/yzhang123/orgs",
"repos_url": "https://api.github.com/users/yzhang123/repos",
"events_url": "https://api.github.com/users/yzhang123/events{/privacy}",
"received_events_url": "https://api.github.com/users/yzhang123/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "yzhang123",
"id": 4204271,
"node_id": "MDQ6VXNlcjQyMDQyNzE=",
"avatar_url": "https://avatars.githubusercontent.com/u/4204271?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/yzhang123",
"html_url": "https://github.com/yzhang123",
"followers_url": "https://api.github.com/users/yzhang123/followers",
"following_url": "https://api.github.com/users/yzhang123/following{/other_user}",
"gists_url": "https://api.github.com/users/yzhang123/gists{/gist_id}",
"starred_url": "https://api.github.com/users/yzhang123/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/yzhang123/subscriptions",
"organizations_url": "https://api.github.com/users/yzhang123/orgs",
"repos_url": "https://api.github.com/users/yzhang123/repos",
"events_url": "https://api.github.com/users/yzhang123/events{/privacy}",
"received_events_url": "https://api.github.com/users/yzhang123/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2023-07-03T12:05:08 | 2023-09-07T01:44:17 | 2023-09-07T01:44:17 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | Hi Team,
I am trying latest itn model t5 for inference.(https://catalog.ngc.nvidia.com/orgs/nvidia/teams/nemo/models/itn_en_t5)
Getting error as the following:
Error executing job with overrides: ['lang=en', 'mode=itn', 'tagger_pretrained_model=itn_en_t5', 'decoder_pretrained_model=itn_en_t5', 'inference.interactive=False', 'inference.from_file=[sample.txt]']
Traceback (most recent call last):
File "NeMo/examples/nlp/duplex_text_normalization/duplex_text_normalization_infer.py", line 78, in main
normalizer_electronic = ElectronicNormalizer(input_case="cased", lang=lang, deterministic=True)
File "/home/ubuntu/NeMo/examples/nlp/duplex_text_normalization/nn_wfst/en/electronic/normalize.py", line 46, in __init__
self.tagger = ClassifyFst(
File "/home/ubuntu/NeMo/examples/nlp/duplex_text_normalization/nn_wfst/en/electronic/tokenize_and_classify.py", line 66, in __init__
electonic_graph = ElectronicFst(deterministic=deterministic).fst
TypeError: __init__() missing 1 required positional argument: 'cardinal'
Is it the issue with the "nemo_text_processing" package?or version mismatch of other packages?
| {
"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/6967/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/6967/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6965 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6965/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6965/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6965/events | https://github.com/NVIDIA/NeMo/issues/6965 | 1,785,474,513 | I_kwDOC_bI7s5qbC3R | 6,965 | Speaker Diarization - Extracting speaker embeddings for labels from files with multiple speakers from MSDD models | {
"login": "asusdisciple",
"id": 138434950,
"node_id": "U_kgDOCEBZhg",
"avatar_url": "https://avatars.githubusercontent.com/u/138434950?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/asusdisciple",
"html_url": "https://github.com/asusdisciple",
"followers_url": "https://api.github.com/users/asusdisciple/followers",
"following_url": "https://api.github.com/users/asusdisciple/following{/other_user}",
"gists_url": "https://api.github.com/users/asusdisciple/gists{/gist_id}",
"starred_url": "https://api.github.com/users/asusdisciple/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/asusdisciple/subscriptions",
"organizations_url": "https://api.github.com/users/asusdisciple/orgs",
"repos_url": "https://api.github.com/users/asusdisciple/repos",
"events_url": "https://api.github.com/users/asusdisciple/events{/privacy}",
"received_events_url": "https://api.github.com/users/asusdisciple/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 | {
"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 | 2023-07-03T06:58:57 | 2023-09-04T01:44:30 | 2023-09-04T01:44:30 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Is your feature request related to a problem? Please describe.**
One problem I come across often is, that I can't verify already known speakers from other files, since inference only yields labels for a **MSDD model** approach.
For example: I have file **A** with speakers Paul, Tom and Charlie and File **B** with speaker Bert, Charlie and Johann. To determine if **Charlie is in File A and B**, I have to get his speaker embedding. At the moment I have to infer speaker diarization on file A to do this, get the timestamps for Charlie, cut the file at the marks, stitch it together again and get the corresponding feature embedding which I can then compare to file B, by doing the same. However for a production setup I would need the embedding from the MSDD model, but from my understanding the MSDD model only produces a probability vector?
**Describe the solution you'd like**
A solution would be to extract the feature embeddings with some kind of method, after diarization and associate them with the labels. I found the get_cluster_avg_embs_model() method for MSDD models, but I am not sure how to use it and if it yields the correct results (to me its not clear which parameters to use here, since there are a lot of internal states in the model after diarization). So the question would, which parameter yields the best representation of a speaker to recognize him again?
| {
"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/6965/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/6965/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6963 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6963/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6963/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6963/events | https://github.com/NVIDIA/NeMo/pull/6963 | 1,785,360,933 | PR_kwDOC_bI7s5Ue5G6 | 6,963 | Allow `data_prefix` to be `None` when mocking dataset | {
"login": "yaox12",
"id": 3831900,
"node_id": "MDQ6VXNlcjM4MzE5MDA=",
"avatar_url": "https://avatars.githubusercontent.com/u/3831900?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/yaox12",
"html_url": "https://github.com/yaox12",
"followers_url": "https://api.github.com/users/yaox12/followers",
"following_url": "https://api.github.com/users/yaox12/following{/other_user}",
"gists_url": "https://api.github.com/users/yaox12/gists{/gist_id}",
"starred_url": "https://api.github.com/users/yaox12/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/yaox12/subscriptions",
"organizations_url": "https://api.github.com/users/yaox12/orgs",
"repos_url": "https://api.github.com/users/yaox12/repos",
"events_url": "https://api.github.com/users/yaox12/events{/privacy}",
"received_events_url": "https://api.github.com/users/yaox12/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 | 2023-07-03T05:24:10 | 2023-10-24T01:20:58 | 2023-10-20T18:28:04 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6963",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6963",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6963.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6963.patch",
"merged_at": "2023-10-20T18:28:04"
} | # What does this PR do ?
Allow `data_prefix` to be None when mocking dataset, i.e., allow the following config
```yaml
data_prefix: null
data_impl: mock
```
Without this change, it will throw en error saying that `None` type doesn't support the `len()` method.
**Collection**: [NLP]
# Changelog
- nemo/collections/nlp/data/language_modeling/megatron/gpt_dataset.py
# 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)
cc @titu1994 Could you kindly review it? Thanks. | {
"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/6963/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/6963/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6962 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6962/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6962/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6962/events | https://github.com/NVIDIA/NeMo/issues/6962 | 1,785,318,192 | I_kwDOC_bI7s5qacsw | 6,962 | RuntimeError: torch.cat(): expected a non-empty list of Tensors when sequence parallel is enabled | {
"login": "szhengac",
"id": 3960020,
"node_id": "MDQ6VXNlcjM5NjAwMjA=",
"avatar_url": "https://avatars.githubusercontent.com/u/3960020?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/szhengac",
"html_url": "https://github.com/szhengac",
"followers_url": "https://api.github.com/users/szhengac/followers",
"following_url": "https://api.github.com/users/szhengac/following{/other_user}",
"gists_url": "https://api.github.com/users/szhengac/gists{/gist_id}",
"starred_url": "https://api.github.com/users/szhengac/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/szhengac/subscriptions",
"organizations_url": "https://api.github.com/users/szhengac/orgs",
"repos_url": "https://api.github.com/users/szhengac/repos",
"events_url": "https://api.github.com/users/szhengac/events{/privacy}",
"received_events_url": "https://api.github.com/users/szhengac/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 | 2023-07-03T04:55:12 | 2023-08-10T01:50:07 | 2023-08-10T01:50:07 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | Hi,
I found there are some bugs when sequence parallel is enabled and transformer engine is not used. Specifically, I received
```
File "/fsx/NeMo/nemo/collections/nlp/models/language_modeling/megatron_gpt_model.py", line 551, in training_step
self.allreduce_sequence_parallel_gradients()
File "/fsx/NeMo/nemo/collections/nlp/models/language_modeling/megatron_gpt_model.py", line 654, in allreduce_sequence_parallel_gradients
coalesced = torch._utils._flatten_dense_tensors(grads)
File "/usr/local/lib/python3.8/dist-packages/torch/_utils.py", line 459, in _flatten_dense_tensors
return torch._C._nn.flatten_dense_tensors(tensors)
RuntimeError: torch.cat(): expected a non-empty list of Tensors
```
After some debugging, I found attribute `sequence_parallel` is not added to the layer norm because the apex layer norm implementation that NeMo uses does not support sequence parallel: https://github.com/NVIDIA/NeMo/blob/b0e5bf3627dbcfb3f4a72d73d3c5e92184d8b1f6/nemo/collections/nlp/modules/common/megatron/fused_layer_norm.py#L18-L19
Besides, is there any plan to add sequence parallel to `rmsnorm`?
| {
"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/6962/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/6962/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6961 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6961/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6961/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6961/events | https://github.com/NVIDIA/NeMo/issues/6961 | 1,784,695,087 | I_kwDOC_bI7s5qYEkv | 6,961 | Please help advise on multi-gpu training of marblenet with pytorch lightning 1.9.4 | {
"login": "PaulSZH95",
"id": 85733116,
"node_id": "MDQ6VXNlcjg1NzMzMTE2",
"avatar_url": "https://avatars.githubusercontent.com/u/85733116?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/PaulSZH95",
"html_url": "https://github.com/PaulSZH95",
"followers_url": "https://api.github.com/users/PaulSZH95/followers",
"following_url": "https://api.github.com/users/PaulSZH95/following{/other_user}",
"gists_url": "https://api.github.com/users/PaulSZH95/gists{/gist_id}",
"starred_url": "https://api.github.com/users/PaulSZH95/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/PaulSZH95/subscriptions",
"organizations_url": "https://api.github.com/users/PaulSZH95/orgs",
"repos_url": "https://api.github.com/users/PaulSZH95/repos",
"events_url": "https://api.github.com/users/PaulSZH95/events{/privacy}",
"received_events_url": "https://api.github.com/users/PaulSZH95/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | 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 | 2023-07-02T15:40:46 | 2023-07-29T08:06:42 | 2023-07-29T08:06:41 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | ### Discussed in https://github.com/NVIDIA/NeMo/discussions/6959
<div type='discussions-op-text'>
<sup>Originally posted by **PaulSZH95** July 2, 2023</sup>
Hi all,
1) May i ask if it is normal for single gpu to have lesser % utilisation than if ddp was used.
2) May I ask if it is normal that if i have 2 gpu and 1 gpu is at 100% the other would be at only 30 %
3) how do i track what is normal gpu utilisation with multigpu training
4) are these methods : nemo.collections.asr.modules.AudioToMFCCPreprocessor and nemo.collections.asr.modules.SpectrogramAugmentation using cpu or gpu
5) if the methods uses cpu, how will if affect parallel training. Details below
I followed the tutorial on VAD through https://docs.nvidia.com/deeplearning/nemo/user-guide/docs/en/stable/starthere/tutorials.html which used branch r1.18.0. I have set up all dependencies i.e. having megatron and apex. My pytorch compatibility are also resolved.
According to the tutorial I can speed up training with device = 2 if I have 2 device.
While the image I provide is a snapshot where both gpu has the same load, most of the time it would be either 1 gpu on 80 - 100% and the other at 0% - 30 % or the other way around. So both gpu are used but at any 1 time the load on either gpu would be vastly different. I am not sure if it is an expected behavior.
In addition, when i use the default marble_net config as given in tutorial, the gpu utilisation is at 14 - 20 % which is much diff than if ddp was used.
Using watch -n 0.5 nvidia-smi (please ignore the circled area)
Lastly, unlike other discussion where they found the pid to be allocated to just 1 gpu, it seems both of my gpu is running. Unfortunately I am still trying to get pid for gpu to be shown thus I can't say for sure either.
</div>
<div>
Could I also check how the number of batch is decided?
in the MarbleNet_3x2x64.yaml file, batch_size was clearly stated as 128. in log,
[NeMo I 2023-07-02 03:08:18 collections:298] Filtered duration for loading collection is 0.000000.
[NeMo I 2023-07-02 03:08:18 collections:301] Dataset loaded with 2164876 items, total duration of 378.85 hours.
[NeMo I 2023-07-02 03:08:18 collections:303] # 2164876 files loaded accounting to # 2 labels
[NeMo I 2023-07-02 03:08:22 collections:298] Filtered duration for loading collection is 0.000000.
[NeMo I 2023-07-02 03:08:22 collections:301] Dataset loaded with 234914 items, total duration of 41.11 hours.
[NeMo I 2023-07-02 03:08:22 collections:303] # 234914 files loaded accounting to # 2 labels
[NeMo I 2023-07-02 03:08:34 collections:298] Filtered duration for loading collection is 0.000000.
[NeMo I 2023-07-02 03:08:34 collections:301] Dataset loaded with 824574 items, total duration of 144.30 hours.
but in training:
Epoch 3: 100%|████████████████████████████████████████████████████| 18750/18750 [59:59<00:00, 5.21it/s, loss=0.322, v_num=8-42
if 18750 at batch_size of 128 I would have to have 2400000 which is severely greater than my 2164876 training items. And even if the batch size is the measurement of mfcc, each file give 64 x 64, then number of batch should be at 1082438. if purely by audio file then it should be 16913 or 16914 depending on rounding.
please someone help, thank you.
</div> | {
"login": "PaulSZH95",
"id": 85733116,
"node_id": "MDQ6VXNlcjg1NzMzMTE2",
"avatar_url": "https://avatars.githubusercontent.com/u/85733116?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/PaulSZH95",
"html_url": "https://github.com/PaulSZH95",
"followers_url": "https://api.github.com/users/PaulSZH95/followers",
"following_url": "https://api.github.com/users/PaulSZH95/following{/other_user}",
"gists_url": "https://api.github.com/users/PaulSZH95/gists{/gist_id}",
"starred_url": "https://api.github.com/users/PaulSZH95/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/PaulSZH95/subscriptions",
"organizations_url": "https://api.github.com/users/PaulSZH95/orgs",
"repos_url": "https://api.github.com/users/PaulSZH95/repos",
"events_url": "https://api.github.com/users/PaulSZH95/events{/privacy}",
"received_events_url": "https://api.github.com/users/PaulSZH95/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6961/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/6961/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6960 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6960/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6960/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6960/events | https://github.com/NVIDIA/NeMo/issues/6960 | 1,784,616,025 | I_kwDOC_bI7s5qXxRZ | 6,960 | No module named 'nemo.collections' | {
"login": "muizzhaque",
"id": 113048786,
"node_id": "U_kgDOBrz80g",
"avatar_url": "https://avatars.githubusercontent.com/u/113048786?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/muizzhaque",
"html_url": "https://github.com/muizzhaque",
"followers_url": "https://api.github.com/users/muizzhaque/followers",
"following_url": "https://api.github.com/users/muizzhaque/following{/other_user}",
"gists_url": "https://api.github.com/users/muizzhaque/gists{/gist_id}",
"starred_url": "https://api.github.com/users/muizzhaque/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/muizzhaque/subscriptions",
"organizations_url": "https://api.github.com/users/muizzhaque/orgs",
"repos_url": "https://api.github.com/users/muizzhaque/repos",
"events_url": "https://api.github.com/users/muizzhaque/events{/privacy}",
"received_events_url": "https://api.github.com/users/muizzhaque/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 | 2023-07-02T12:26:36 | 2023-07-05T00:26:49 | 2023-07-05T00:26:34 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | I have followed the steps as mentioned in git repo for nemo, but each time I'm installing nemo_toolkit['all'] two packages pynini and mecab-python3 is failing to build wheels, also tried nemo_toolkit['asr'] which was successful but as soon as i run the code, i got the same module not found error.
btw, I'm using windows 10. | {
"login": "muizzhaque",
"id": 113048786,
"node_id": "U_kgDOBrz80g",
"avatar_url": "https://avatars.githubusercontent.com/u/113048786?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/muizzhaque",
"html_url": "https://github.com/muizzhaque",
"followers_url": "https://api.github.com/users/muizzhaque/followers",
"following_url": "https://api.github.com/users/muizzhaque/following{/other_user}",
"gists_url": "https://api.github.com/users/muizzhaque/gists{/gist_id}",
"starred_url": "https://api.github.com/users/muizzhaque/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/muizzhaque/subscriptions",
"organizations_url": "https://api.github.com/users/muizzhaque/orgs",
"repos_url": "https://api.github.com/users/muizzhaque/repos",
"events_url": "https://api.github.com/users/muizzhaque/events{/privacy}",
"received_events_url": "https://api.github.com/users/muizzhaque/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6960/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/6960/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6958 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6958/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6958/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6958/events | https://github.com/NVIDIA/NeMo/pull/6958 | 1,783,311,898 | PR_kwDOC_bI7s5UX2FR | 6,958 | Fix incorrect embedding grads with distopt BF16 grad reductions | {
"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": 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": 4418253970,
"node_id": "LA_kwDOC_bI7s8AAAABB1k0kg",
"url": "https://api.github.com/repos/NVIDIA/NeMo/labels/TTS",
"name": "TTS",
"color": "49AF28",
"default": false,
"description": ""
},
{
"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": ""
},
{
"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": ""
},
{
"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": ""
},
{
"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 | 2023-07-01T01:06:04 | 2023-09-01T19:10:25 | 2023-08-02T14:59:16 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6958",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6958",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6958.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6958.patch",
"merged_at": "2023-08-02T14:59:16"
} | # What does this PR do ?
In the case where we are training GPT with distopt, BF16 grads, and pipeline parallelism, we are not properly synchronizing the embedding grads between the first and last pipeline stage. The embeddings were being accumulated into BF16 buffers and the pipeline-parallel reduction was on FP32 buffers (so just all-reducing zeros). Note that we synchronize these buffers before the data-parallel reduction, so that is still being done properly in FP32.
This PR fixes the issue by eagerly accumulating the embedding grads into FP32 buffers.
**Collection**: NLP
# Changelog
- Distopt accumulates grads for FP32 params into FP32 buffers.
# Usage
In the config file, set the optimizer with `name: distributed_fused_adam` and `grad_sync_dtype: bf16`.
# 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
* Depends/includes https://github.com/NVIDIA/NeMo/pull/6886
| {
"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/6958/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/6958/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6957 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6957/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6957/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6957/events | https://github.com/NVIDIA/NeMo/pull/6957 | 1,783,267,626 | PR_kwDOC_bI7s5UXsc9 | 6,957 | Stop at the stop token | {
"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 | 2023-07-01T00:01:31 | 2023-07-03T17:13:56 | 2023-07-03T17:13:55 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6957",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6957",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6957.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6957.patch",
"merged_at": "2023-07-03T17:13:55"
} | Detects the stop token and stop the generation once the stop token is detected. | {
"login": "aklife97",
"id": 16078071,
"node_id": "MDQ6VXNlcjE2MDc4MDcx",
"avatar_url": "https://avatars.githubusercontent.com/u/16078071?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/aklife97",
"html_url": "https://github.com/aklife97",
"followers_url": "https://api.github.com/users/aklife97/followers",
"following_url": "https://api.github.com/users/aklife97/following{/other_user}",
"gists_url": "https://api.github.com/users/aklife97/gists{/gist_id}",
"starred_url": "https://api.github.com/users/aklife97/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/aklife97/subscriptions",
"organizations_url": "https://api.github.com/users/aklife97/orgs",
"repos_url": "https://api.github.com/users/aklife97/repos",
"events_url": "https://api.github.com/users/aklife97/events{/privacy}",
"received_events_url": "https://api.github.com/users/aklife97/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6957/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/6957/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6956 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6956/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6956/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6956/events | https://github.com/NVIDIA/NeMo/pull/6956 | 1,783,195,200 | PR_kwDOC_bI7s5UXc4J | 6,956 | Fix requirements for pydantic + inflect | {
"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
} | [] | closed | false | null | [] | null | null | 2023-06-30T21:44:18 | 2023-07-05T17:07:18 | 2023-07-01T00:32:52 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6956",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6956",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6956.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6956.patch",
"merged_at": "2023-07-01T00:32:52"
} | # What does this PR do ?
Upperbound pydantic
**Collection**: [ASR,NLP,TTS]
# 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": "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/6956/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/6956/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6955 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6955/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6955/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6955/events | https://github.com/NVIDIA/NeMo/pull/6955 | 1,783,103,403 | PR_kwDOC_bI7s5UXIvo | 6,955 | Add ASR with TTS Tutorial. Fix enhancer usage. | {
"login": "artbataev",
"id": 9380560,
"node_id": "MDQ6VXNlcjkzODA1NjA=",
"avatar_url": "https://avatars.githubusercontent.com/u/9380560?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/artbataev",
"html_url": "https://github.com/artbataev",
"followers_url": "https://api.github.com/users/artbataev/followers",
"following_url": "https://api.github.com/users/artbataev/following{/other_user}",
"gists_url": "https://api.github.com/users/artbataev/gists{/gist_id}",
"starred_url": "https://api.github.com/users/artbataev/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/artbataev/subscriptions",
"organizations_url": "https://api.github.com/users/artbataev/orgs",
"repos_url": "https://api.github.com/users/artbataev/repos",
"events_url": "https://api.github.com/users/artbataev/events{/privacy}",
"received_events_url": "https://api.github.com/users/artbataev/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 | 2023-06-30T20:05:47 | 2023-07-12T17:04:37 | 2023-07-12T17:04:28 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6955",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6955",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6955.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6955.patch",
"merged_at": "2023-07-12T17:04:28"
} | # What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
**Collection**: [ASR]
# Changelog
- Tutorial for using hybrid ASR-TTS Models (`ASRWithTTSModel`)
- fix enhancer loading in `ASRWithTTSModel`
# 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
- [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": "artbataev",
"id": 9380560,
"node_id": "MDQ6VXNlcjkzODA1NjA=",
"avatar_url": "https://avatars.githubusercontent.com/u/9380560?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/artbataev",
"html_url": "https://github.com/artbataev",
"followers_url": "https://api.github.com/users/artbataev/followers",
"following_url": "https://api.github.com/users/artbataev/following{/other_user}",
"gists_url": "https://api.github.com/users/artbataev/gists{/gist_id}",
"starred_url": "https://api.github.com/users/artbataev/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/artbataev/subscriptions",
"organizations_url": "https://api.github.com/users/artbataev/orgs",
"repos_url": "https://api.github.com/users/artbataev/repos",
"events_url": "https://api.github.com/users/artbataev/events{/privacy}",
"received_events_url": "https://api.github.com/users/artbataev/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6955/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/6955/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6951 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6951/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6951/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6951/events | https://github.com/NVIDIA/NeMo/pull/6951 | 1,783,055,080 | PR_kwDOC_bI7s5UW-II | 6,951 | peft folder name | {
"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": 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": ""
},
{
"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 | 2023-06-30T19:23:30 | 2023-07-23T01:57:10 | 2023-07-23T01:57:10 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6951",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6951",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6951.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6951.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/6951/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/6951/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6949 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6949/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6949/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6949/events | https://github.com/NVIDIA/NeMo/pull/6949 | 1,781,583,059 | PR_kwDOC_bI7s5USC47 | 6,949 | Draft: Torch-native FSDP support | {
"login": "erhoo82",
"id": 15861259,
"node_id": "MDQ6VXNlcjE1ODYxMjU5",
"avatar_url": "https://avatars.githubusercontent.com/u/15861259?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/erhoo82",
"html_url": "https://github.com/erhoo82",
"followers_url": "https://api.github.com/users/erhoo82/followers",
"following_url": "https://api.github.com/users/erhoo82/following{/other_user}",
"gists_url": "https://api.github.com/users/erhoo82/gists{/gist_id}",
"starred_url": "https://api.github.com/users/erhoo82/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/erhoo82/subscriptions",
"organizations_url": "https://api.github.com/users/erhoo82/orgs",
"repos_url": "https://api.github.com/users/erhoo82/repos",
"events_url": "https://api.github.com/users/erhoo82/events{/privacy}",
"received_events_url": "https://api.github.com/users/erhoo82/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 | 2023-06-29T22:15:42 | 2023-10-20T22:22:35 | 2023-08-15T01:43:42 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6949",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6949",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6949.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6949.patch",
"merged_at": null
} | # What does this PR do ?
Add the support for PyTorch-native FSDP for GPT models.
# 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/6949/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/6949/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6948 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6948/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6948/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6948/events | https://github.com/NVIDIA/NeMo/pull/6948 | 1,781,568,472 | PR_kwDOC_bI7s5UR_qr | 6,948 | Merge release r1.19.0 into 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 | 2023-06-29T22:00:17 | 2023-06-29T22:07:41 | 2023-06-29T22:07:40 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6948",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6948",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6948.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6948.patch",
"merged_at": "2023-06-29T22:07:40"
} | # What does this PR do ?
Makes sure main and release have not diverged.
**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/6948/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/6948/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6947 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6947/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6947/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6947/events | https://github.com/NVIDIA/NeMo/pull/6947 | 1,781,535,936 | PR_kwDOC_bI7s5UR4mp | 6,947 | fix num_workers=0 issue for gpt sft | {
"login": "yzhang123",
"id": 4204271,
"node_id": "MDQ6VXNlcjQyMDQyNzE=",
"avatar_url": "https://avatars.githubusercontent.com/u/4204271?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/yzhang123",
"html_url": "https://github.com/yzhang123",
"followers_url": "https://api.github.com/users/yzhang123/followers",
"following_url": "https://api.github.com/users/yzhang123/following{/other_user}",
"gists_url": "https://api.github.com/users/yzhang123/gists{/gist_id}",
"starred_url": "https://api.github.com/users/yzhang123/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/yzhang123/subscriptions",
"organizations_url": "https://api.github.com/users/yzhang123/orgs",
"repos_url": "https://api.github.com/users/yzhang123/repos",
"events_url": "https://api.github.com/users/yzhang123/events{/privacy}",
"received_events_url": "https://api.github.com/users/yzhang123/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 | 2023-06-29T21:24:22 | 2023-08-20T01:44:05 | 2023-08-20T01:44:05 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6947",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6947",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6947.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6947.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/6947/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/6947/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6946 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6946/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6946/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6946/events | https://github.com/NVIDIA/NeMo/issues/6946 | 1,781,108,315 | I_kwDOC_bI7s5qKY5b | 6,946 | GPT eval script error out due to unknown restore_path | {
"login": "ZiyueXu77",
"id": 71786575,
"node_id": "MDQ6VXNlcjcxNzg2NTc1",
"avatar_url": "https://avatars.githubusercontent.com/u/71786575?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ZiyueXu77",
"html_url": "https://github.com/ZiyueXu77",
"followers_url": "https://api.github.com/users/ZiyueXu77/followers",
"following_url": "https://api.github.com/users/ZiyueXu77/following{/other_user}",
"gists_url": "https://api.github.com/users/ZiyueXu77/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ZiyueXu77/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ZiyueXu77/subscriptions",
"organizations_url": "https://api.github.com/users/ZiyueXu77/orgs",
"repos_url": "https://api.github.com/users/ZiyueXu77/repos",
"events_url": "https://api.github.com/users/ZiyueXu77/events{/privacy}",
"received_events_url": "https://api.github.com/users/ZiyueXu77/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 | 2023-06-29T15:50:35 | 2023-08-07T01:45:06 | 2023-08-07T01:45:06 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
megatron_gpt_eval.py script error out if using checkpoint+hparams instead of nemo
**Steps/Code to reproduce bug**
Following the instruction:
`python megatron_gpt_eval.py \ checkpoint_dir=PATH_TO_CHECKPOINT_FILE \ checkpoint_name=CHECKPOINT_FILE_NAME \ hparams_file=HPARAMS_FILE \ inference.greedy=True \ inference.add_BOS=True \ trainer.devices=1 \ trainer.num_nodes=1 \ tensor_model_parallel_size=-1 \ pipeline_model_parallel_size=-1 \ prompts=[prompt1,prompt2]`
The script will error out:
`Traceback (most recent call last): File "megatron_gpt_eval.py", line 176, in main model_config = MegatronGPTModel.restore_from( File "/workspace/NeMo-1.18.0/nemo/core/classes/modelPT.py", line 425, in restore_from restore_path = os.path.abspath(os.path.expanduser(restore_path)) File "/usr/lib/python3.8/posixpath.py", line 231, in expanduser path = os.fspath(path) TypeError: expected str, bytes or os.PathLike object, not NoneType`
**Expected behavior**
A clear and concise description of what you expected to happen.
The root cause is that "gpt_model_file=null" check is at the higher level, hence the current script will error out due to restore path unknown. Solution will be Move https://github.com/NVIDIA/NeMo/blob/main/examples/nlp/language_modeling/megatron_gpt_eval.py#L172-L189 to after https://github.com/NVIDIA/NeMo/blob/main/examples/nlp/language_modeling/megatron_gpt_eval.py#L191
**Environment overview (please complete the following information)**
23.02 docker
**Environment details**
If NVIDIA docker image is used you don't need to specify these.
Otherwise, please provide:
- OS version
- PyTorch version
- Python version
**Additional context**
Add any other context about the problem here.
Example: GPU model
| {
"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/6946/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/6946/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6945 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6945/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6945/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6945/events | https://github.com/NVIDIA/NeMo/issues/6945 | 1,781,096,948 | I_kwDOC_bI7s5qKWH0 | 6,945 | Enable SFT script to do multi-round/epoch training | {
"login": "ZiyueXu77",
"id": 71786575,
"node_id": "MDQ6VXNlcjcxNzg2NTc1",
"avatar_url": "https://avatars.githubusercontent.com/u/71786575?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ZiyueXu77",
"html_url": "https://github.com/ZiyueXu77",
"followers_url": "https://api.github.com/users/ZiyueXu77/followers",
"following_url": "https://api.github.com/users/ZiyueXu77/following{/other_user}",
"gists_url": "https://api.github.com/users/ZiyueXu77/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ZiyueXu77/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ZiyueXu77/subscriptions",
"organizations_url": "https://api.github.com/users/ZiyueXu77/orgs",
"repos_url": "https://api.github.com/users/ZiyueXu77/repos",
"events_url": "https://api.github.com/users/ZiyueXu77/events{/privacy}",
"received_events_url": "https://api.github.com/users/ZiyueXu77/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 | {
"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 | 2023-06-29T15:42:42 | 2023-08-07T01:45:07 | 2023-08-07T01:45:07 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Is your feature request related to a problem? Please describe.**
In the current SFT training script, it runs for 1 epoch with the given max_step number, the max_epoch does not matter as the script generates a temp data file according to the max_step and runs through it. But in some applications, it is desirable to build a multi-epoch baseline for comparison purpose (e.g. for comparing a 5-round (1 epoch per round) federated learning result against a 5-epoch centralized training with the same number of steps per epoch). Currently this cannot be done, since the data used are not the same: 1-epoch-1000-step is not equivalent to 5-epoch-200-step due to the difference in the actual data being sampled.
**Describe the solution you'd like**
I wonder if we could enable multi-epoch training for SFT: the per-epoch data will be specified by max_step (or call it step_per_epoch).
**Describe alternatives you've considered**
An "explicit" implementation of multi-round training will further be helpful: instead of calling .fit() once, something like below will really help easy-transition to FL:
for round in range(K):
#necessary state loading
trainer.fit(model)
#necessary state saving
**Additional context**
Add any other context or screenshots about the feature request here.
| {
"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/6945/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/6945/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6944 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6944/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6944/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6944/events | https://github.com/NVIDIA/NeMo/pull/6944 | 1,780,620,175 | PR_kwDOC_bI7s5UOyNd | 6,944 | Add more arithmetic resolvers | {
"login": "jbaczek",
"id": 45043825,
"node_id": "MDQ6VXNlcjQ1MDQzODI1",
"avatar_url": "https://avatars.githubusercontent.com/u/45043825?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jbaczek",
"html_url": "https://github.com/jbaczek",
"followers_url": "https://api.github.com/users/jbaczek/followers",
"following_url": "https://api.github.com/users/jbaczek/following{/other_user}",
"gists_url": "https://api.github.com/users/jbaczek/gists{/gist_id}",
"starred_url": "https://api.github.com/users/jbaczek/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/jbaczek/subscriptions",
"organizations_url": "https://api.github.com/users/jbaczek/orgs",
"repos_url": "https://api.github.com/users/jbaczek/repos",
"events_url": "https://api.github.com/users/jbaczek/events{/privacy}",
"received_events_url": "https://api.github.com/users/jbaczek/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": 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 | 2023-06-29T11:23:20 | 2023-07-22T01:52:10 | 2023-07-22T01:52:10 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6944",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6944",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6944.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6944.patch",
"merged_at": null
} | # What does this PR do ?
Add arithmetic resolvers other than `multiply`
# Changelog
- No high level changes
- This PR contains custom resolvers for arithmetic, comparison and ternary expression
# Usage
Now in hydra configs you can use more arithmetic and ternary expressions.
```yaml
hydra_config_field: ${if:${condition},0,${ceil_div:123400,${batch_size}}}}
```
# 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
## 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
These resolvers are solution to already implemented [here](https://github.com/NVIDIA/NeMo/blob/main/nemo/core/config/hydra_runner.py#L27). It would be beneficial to implement a single resolver for evaluating arithmetic. The approach could be to use ast module to make a custom parser that allows only literals and binary operators. | {
"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/6944/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/6944/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6943 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6943/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6943/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6943/events | https://github.com/NVIDIA/NeMo/issues/6943 | 1,780,508,702 | I_kwDOC_bI7s5qIGge | 6,943 | Interpretation of Conformer CTC training loss and WER | {
"login": "manchandasahil",
"id": 32937046,
"node_id": "MDQ6VXNlcjMyOTM3MDQ2",
"avatar_url": "https://avatars.githubusercontent.com/u/32937046?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/manchandasahil",
"html_url": "https://github.com/manchandasahil",
"followers_url": "https://api.github.com/users/manchandasahil/followers",
"following_url": "https://api.github.com/users/manchandasahil/following{/other_user}",
"gists_url": "https://api.github.com/users/manchandasahil/gists{/gist_id}",
"starred_url": "https://api.github.com/users/manchandasahil/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/manchandasahil/subscriptions",
"organizations_url": "https://api.github.com/users/manchandasahil/orgs",
"repos_url": "https://api.github.com/users/manchandasahil/repos",
"events_url": "https://api.github.com/users/manchandasahil/events{/privacy}",
"received_events_url": "https://api.github.com/users/manchandasahil/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 | 2023-06-29T10:05:46 | 2023-08-22T01:44:02 | 2023-08-22T01:44:02 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
The Val loss keeps increasing while the val_wer keeps getting better.
**Steps/Code to reproduce bug**
I am trying to fine tune conformer CTC large on a custom dataset and my val loss Keeps getting worse but val wer keeps getting better. Please have a look at graph images below.
Run command:
model.optim.lr=0.01 model.optim.weight_decay=1e-4 \
+init_from_nemo_model=stt_en_conformer_ctc_large.nemo \
model.train_ds.batch_size=8 \
trainer.devices=4 trainer.max_epochs=1000 model.optim.sched.warmup_steps=100 trainer.accumulate_grad_batches=4 ++exp_manager.use_datetime_version=False
Configuration:
```
name: "Conformer-CTC-BPE-finetune-OptumRX"
model:
sample_rate: 16000
log_prediction: true # enables logging sample predictions in the output during training
ctc_reduction: 'mean_batch'
skip_nan_grad: false
train_ds:
manifest_filepath: ???
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
use_start_end_token: false
trim_silence: false
max_duration: 21 # 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: ???
sample_rate: ${model.sample_rate}
batch_size: 2 # you may increase batch_size if your memory allows
shuffle: false
num_workers: 8
pin_memory: true
use_start_end_token: false
test_ds:
manifest_filepath: null
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
use_start_end_token: false
tokenizer:
dir: ??? # path to directory which contains either tokenizer.model (bpe) or vocab.txt (wpe)
type: bpe # Can be either bpe (SentencePiece tokenizer) or wpe (WordPiece tokenizer)
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: 18
d_model: 512
# 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 # -1 sets it to 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: []
optim:
name: adamw
lr: 0.1
# 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: 1000
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
accumulate_grad_batches: 1
gradient_clip_val: 0.0
precision: 16 # Should be set to 16 for O1 and O2 to enable the AMP.
log_every_n_steps: 20 # 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: null
name: ${name}-lr-${model.optim.lr}-wd-${model.optim.weight_decay}
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: 2
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: true
resume_ignore_no_checkpoint: true
# You may use this section to create a W&B logger
create_wandb_logger: false
wandb_logger_kwargs:
name: null
project: null
hydra:
run:
dir: .
sweep:
dir: .
subdir: .
```
**Expected behavior**
Successful fine tuning on our dataset.
**Environment overview (please complete the following information)**
- Environment location: Docker
- Method of NeMo install: official Memo docker file
**Environment details**
If NVIDIA docker image is used you don't need to specify these.
Otherwise, please provide:
- OS version
- PyTorch version
- Python version
**Additional context**
Add any other context about the problem here.
Example: GPU model
| {
"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/6943/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/6943/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6941 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6941/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6941/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6941/events | https://github.com/NVIDIA/NeMo/pull/6941 | 1,780,005,646 | PR_kwDOC_bI7s5UMtXV | 6,941 | Blank free transduction | {
"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": 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 | 2023-06-29T02:17:05 | 2023-07-23T01:57:12 | 2023-07-23T01:57:12 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | true | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6941",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6941",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6941.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6941.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/6941/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/6941/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6940 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6940/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6940/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6940/events | https://github.com/NVIDIA/NeMo/pull/6940 | 1,779,954,444 | PR_kwDOC_bI7s5UMi70 | 6,940 | Adding cache-aware streaming ASR checkpoints. | {
"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": 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 | 2023-06-29T00:58:48 | 2023-06-29T22:04:41 | 2023-06-29T22:04:40 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6940",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6940",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6940.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6940.patch",
"merged_at": "2023-06-29T22:04:40"
} | # What does this PR do ?
It adds the following cache-aware checkpoints for streaming ASR:
stt_en_fastconformer_hybrid_large_streaming_80ms
stt_en_fastconformer_hybrid_large_streaming_480ms
stt_en_fastconformer_hybrid_large_streaming_1040ms
| {
"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/6940/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/6940/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6939 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6939/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6939/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6939/events | https://github.com/NVIDIA/NeMo/pull/6939 | 1,779,599,382 | PR_kwDOC_bI7s5ULUjh | 6,939 | Update core commit for CI | {
"login": "aklife97",
"id": 16078071,
"node_id": "MDQ6VXNlcjE2MDc4MDcx",
"avatar_url": "https://avatars.githubusercontent.com/u/16078071?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/aklife97",
"html_url": "https://github.com/aklife97",
"followers_url": "https://api.github.com/users/aklife97/followers",
"following_url": "https://api.github.com/users/aklife97/following{/other_user}",
"gists_url": "https://api.github.com/users/aklife97/gists{/gist_id}",
"starred_url": "https://api.github.com/users/aklife97/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/aklife97/subscriptions",
"organizations_url": "https://api.github.com/users/aklife97/orgs",
"repos_url": "https://api.github.com/users/aklife97/repos",
"events_url": "https://api.github.com/users/aklife97/events{/privacy}",
"received_events_url": "https://api.github.com/users/aklife97/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 | 2023-06-28T19:21:58 | 2023-06-28T22:13:29 | 2023-06-28T22:13:28 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6939",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6939",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6939.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6939.patch",
"merged_at": "2023-06-28T22:13:28"
} | # What does this PR do ?
Update core commit for CI
**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/6939/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/6939/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6938 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6938/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6938/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6938/events | https://github.com/NVIDIA/NeMo/pull/6938 | 1,779,556,900 | PR_kwDOC_bI7s5ULKsN | 6,938 | Upgrade base pytorch container to 23.06 | {
"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": 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 | 2023-06-28T19:05:08 | 2023-06-30T17:35:58 | 2023-06-30T17:35:57 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6938",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6938",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6938.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6938.patch",
"merged_at": "2023-06-30T17:35:57"
} | # What does this PR do ?
Waiting for 23.06 container to be published.
**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/6938/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/6938/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6937 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6937/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6937/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6937/events | https://github.com/NVIDIA/NeMo/pull/6937 | 1,779,524,973 | PR_kwDOC_bI7s5ULDXQ | 6,937 | Fix confidence ensembles RNNT logprobs selection logic for exclude_blank scenario | {
"login": "KunalDhawan",
"id": 15515031,
"node_id": "MDQ6VXNlcjE1NTE1MDMx",
"avatar_url": "https://avatars.githubusercontent.com/u/15515031?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/KunalDhawan",
"html_url": "https://github.com/KunalDhawan",
"followers_url": "https://api.github.com/users/KunalDhawan/followers",
"following_url": "https://api.github.com/users/KunalDhawan/following{/other_user}",
"gists_url": "https://api.github.com/users/KunalDhawan/gists{/gist_id}",
"starred_url": "https://api.github.com/users/KunalDhawan/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/KunalDhawan/subscriptions",
"organizations_url": "https://api.github.com/users/KunalDhawan/orgs",
"repos_url": "https://api.github.com/users/KunalDhawan/repos",
"events_url": "https://api.github.com/users/KunalDhawan/events{/privacy}",
"received_events_url": "https://api.github.com/users/KunalDhawan/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 | 2023-06-28T18:49:28 | 2023-06-29T15:51:18 | 2023-06-29T15:51:02 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6937",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6937",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6937.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6937.patch",
"merged_at": "2023-06-29T15:51:02"
} | # What does this PR do ?
- Fixes the algorithm for filtering blank frames while building confidence ensembles
- Adds checks for confidence value calculation to handle the edge case of all blank frames
**Collection**: [Note which collection this PR will affect]
ASR
# Changelog
- changes to get_filtered_logprobs() and compute_confidence() functions in nemo/collections/asr/models/confidence_ensembles.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)
- [ ] 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.
| {
"login": "KunalDhawan",
"id": 15515031,
"node_id": "MDQ6VXNlcjE1NTE1MDMx",
"avatar_url": "https://avatars.githubusercontent.com/u/15515031?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/KunalDhawan",
"html_url": "https://github.com/KunalDhawan",
"followers_url": "https://api.github.com/users/KunalDhawan/followers",
"following_url": "https://api.github.com/users/KunalDhawan/following{/other_user}",
"gists_url": "https://api.github.com/users/KunalDhawan/gists{/gist_id}",
"starred_url": "https://api.github.com/users/KunalDhawan/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/KunalDhawan/subscriptions",
"organizations_url": "https://api.github.com/users/KunalDhawan/orgs",
"repos_url": "https://api.github.com/users/KunalDhawan/repos",
"events_url": "https://api.github.com/users/KunalDhawan/events{/privacy}",
"received_events_url": "https://api.github.com/users/KunalDhawan/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6937/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/6937/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6936 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6936/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6936/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6936/events | https://github.com/NVIDIA/NeMo/issues/6936 | 1,779,006,760 | I_kwDOC_bI7s5qCX0o | 6,936 | Frame VAD eval not working? | {
"login": "pehonnet",
"id": 4670720,
"node_id": "MDQ6VXNlcjQ2NzA3MjA=",
"avatar_url": "https://avatars.githubusercontent.com/u/4670720?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/pehonnet",
"html_url": "https://github.com/pehonnet",
"followers_url": "https://api.github.com/users/pehonnet/followers",
"following_url": "https://api.github.com/users/pehonnet/following{/other_user}",
"gists_url": "https://api.github.com/users/pehonnet/gists{/gist_id}",
"starred_url": "https://api.github.com/users/pehonnet/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/pehonnet/subscriptions",
"organizations_url": "https://api.github.com/users/pehonnet/orgs",
"repos_url": "https://api.github.com/users/pehonnet/repos",
"events_url": "https://api.github.com/users/pehonnet/events{/privacy}",
"received_events_url": "https://api.github.com/users/pehonnet/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 | 2023-06-28T13:56:18 | 2023-07-17T12:43:33 | 2023-07-17T12:43:33 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
I prepared a test manifest in the required format (i.e. with rttm label, because it doesn't work with the format indicated in the documentation), and tried to run the inference with the default parameters of the frame vad inference. I changed `evaluate` to `True` and here is my output:
```
[NeMo I 2023-06-28 13:40:46 frame_vad_infer:181] Evaluating VAD results
Evaluating VAD results: 0%| | 0/1508 [00:00<?, ?it/s][NeMo W 2023-06-28 13:40:46 nemo_logging:349] .../Tools/miniconda3/envs/nemo_official/lib/python3.8/site-packages/pyannote/metrics/utils.py:200: UserWarning: 'uem' was approximated by the union of 'reference' and 'hypothesis' extents.
warnings.warn(
Evaluating VAD results: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1508/1508 [00:08<00:00, 173.89it/s]
[NeMo I 2023-06-28 13:40:58 frame_vad_infer:192] AUROC: 0.9776
[NeMo I 2023-06-28 13:40:58 frame_vad_infer:193] DetER=100.0000, False Alarm=0.0000, Miss=100.0000
[NeMo I 2023-06-28 13:40:58 frame_vad_infer:194] with params: {'onset': 0.3, 'offset': 0.3, 'pad_onset': 0.2, 'pad_offset': 0.2, 'min_duration_on': 0.2, 'min_duration_off': 0.2, 'filter_speech_first': True}
[NeMo I 2023-06-28 13:40:58 frame_vad_infer:195] Done!
```
When I look at the output in:
```
vad_frame_outputs/seg_output_onset0.3-offset0.3-pad_onset0.2-pad_offset0.2-min_duration_on0.2-min_duration_off0.2-filter_speech_firstTrue-/file_id_1.rttm
```
there is definitely some overlap with the reference segments. So Miss=100 is unlikely. There are also some segments detected which are not in the reference, so False Alarm=0 is unlikely.
Could it be that the evaluation is missing something?
Thanks for your help!
| {
"login": "pehonnet",
"id": 4670720,
"node_id": "MDQ6VXNlcjQ2NzA3MjA=",
"avatar_url": "https://avatars.githubusercontent.com/u/4670720?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/pehonnet",
"html_url": "https://github.com/pehonnet",
"followers_url": "https://api.github.com/users/pehonnet/followers",
"following_url": "https://api.github.com/users/pehonnet/following{/other_user}",
"gists_url": "https://api.github.com/users/pehonnet/gists{/gist_id}",
"starred_url": "https://api.github.com/users/pehonnet/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/pehonnet/subscriptions",
"organizations_url": "https://api.github.com/users/pehonnet/orgs",
"repos_url": "https://api.github.com/users/pehonnet/repos",
"events_url": "https://api.github.com/users/pehonnet/events{/privacy}",
"received_events_url": "https://api.github.com/users/pehonnet/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6936/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/6936/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6935 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6935/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6935/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6935/events | https://github.com/NVIDIA/NeMo/pull/6935 | 1,778,090,965 | PR_kwDOC_bI7s5UGLGA | 6,935 | Add missing save restore connector to eval 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": 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 | 2023-06-28T03:02:25 | 2023-06-29T21:58:17 | 2023-06-29T21:58:14 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6935",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6935",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6935.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6935.patch",
"merged_at": "2023-06-29T21:58:14"
} | # What does this PR do ?
Fix missing save restore connector to eval scripts to allow dynamically determining the values of model parallelism
**Collection**: [NLP]
# 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/6935/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/6935/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6934 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6934/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6934/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6934/events | https://github.com/NVIDIA/NeMo/pull/6934 | 1,778,036,432 | PR_kwDOC_bI7s5UF_YK | 6,934 | support value attribution condition | {
"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 | 2023-06-28T01:58:15 | 2023-06-29T02:30:48 | 2023-06-29T02:30:47 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6934",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6934",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6934.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6934.patch",
"merged_at": "2023-06-29T02:30:47"
} | 1. support value attribute string prediction
2. support value attribute sft
3. update UI to allow user enter values. | {
"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
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6934/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/6934/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6933 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6933/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6933/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6933/events | https://github.com/NVIDIA/NeMo/pull/6933 | 1,778,007,943 | PR_kwDOC_bI7s5UF5EY | 6,933 | Removed optional optimize_for_inference | {
"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": 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"
}
] | closed | false | null | [] | null | null | 2023-06-28T01:29:57 | 2023-06-28T16:26:42 | 2023-06-28T16:26:40 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6933",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6933",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6933.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6933.patch",
"merged_at": "2023-06-28T16:26:40"
} | # What does this PR do ?
Removed optional optimize_for_inference call - causes problems with some nets and does not really help otherwise.
Fixed verbose flag passing.
**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/6933/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/6933/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6932 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6932/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6932/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6932/events | https://github.com/NVIDIA/NeMo/pull/6932 | 1,777,980,663 | PR_kwDOC_bI7s5UFy-j | 6,932 | Adding tutorial for confidence ensembles | {
"login": "Kipok",
"id": 2354422,
"node_id": "MDQ6VXNlcjIzNTQ0MjI=",
"avatar_url": "https://avatars.githubusercontent.com/u/2354422?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Kipok",
"html_url": "https://github.com/Kipok",
"followers_url": "https://api.github.com/users/Kipok/followers",
"following_url": "https://api.github.com/users/Kipok/following{/other_user}",
"gists_url": "https://api.github.com/users/Kipok/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Kipok/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Kipok/subscriptions",
"organizations_url": "https://api.github.com/users/Kipok/orgs",
"repos_url": "https://api.github.com/users/Kipok/repos",
"events_url": "https://api.github.com/users/Kipok/events{/privacy}",
"received_events_url": "https://api.github.com/users/Kipok/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 | 2023-06-28T01:08:01 | 2023-07-12T16:25:38 | 2023-07-12T16:25:23 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6932",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6932",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6932.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6932.patch",
"merged_at": "2023-07-12T16:25:23"
} | # What does this PR do?
There are a couple of bugfixes to the confidence computation and also a tutorial.
It currently misses a few links:
- to the paper (already submitted to arxiv, waiting to get the link)
- to the confidence tutorial from Aleksandr, which is not merged yet
- to the documentation of the confidence ensembles, which I hope to update in the next few days either in this PR or in the follow-up
**Collection**: 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": "Kipok",
"id": 2354422,
"node_id": "MDQ6VXNlcjIzNTQ0MjI=",
"avatar_url": "https://avatars.githubusercontent.com/u/2354422?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Kipok",
"html_url": "https://github.com/Kipok",
"followers_url": "https://api.github.com/users/Kipok/followers",
"following_url": "https://api.github.com/users/Kipok/following{/other_user}",
"gists_url": "https://api.github.com/users/Kipok/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Kipok/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Kipok/subscriptions",
"organizations_url": "https://api.github.com/users/Kipok/orgs",
"repos_url": "https://api.github.com/users/Kipok/repos",
"events_url": "https://api.github.com/users/Kipok/events{/privacy}",
"received_events_url": "https://api.github.com/users/Kipok/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6932/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/6932/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6931 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6931/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6931/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6931/events | https://github.com/NVIDIA/NeMo/pull/6931 | 1,777,932,385 | PR_kwDOC_bI7s5UFok3 | 6,931 | lora inference ci | {
"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": 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 | 2023-06-28T00:10:48 | 2023-06-28T23:25:47 | 2023-06-28T23:25:46 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6931",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6931",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6931.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6931.patch",
"merged_at": "2023-06-28T23:25:46"
} | # What does this PR do ?
adds a CI test to check inference with tp2 lora models.
**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": "aklife97",
"id": 16078071,
"node_id": "MDQ6VXNlcjE2MDc4MDcx",
"avatar_url": "https://avatars.githubusercontent.com/u/16078071?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/aklife97",
"html_url": "https://github.com/aklife97",
"followers_url": "https://api.github.com/users/aklife97/followers",
"following_url": "https://api.github.com/users/aklife97/following{/other_user}",
"gists_url": "https://api.github.com/users/aklife97/gists{/gist_id}",
"starred_url": "https://api.github.com/users/aklife97/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/aklife97/subscriptions",
"organizations_url": "https://api.github.com/users/aklife97/orgs",
"repos_url": "https://api.github.com/users/aklife97/repos",
"events_url": "https://api.github.com/users/aklife97/events{/privacy}",
"received_events_url": "https://api.github.com/users/aklife97/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6931/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/6931/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6930 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6930/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6930/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6930/events | https://github.com/NVIDIA/NeMo/pull/6930 | 1,777,930,388 | PR_kwDOC_bI7s5UFoJK | 6,930 | Fix require_grad typos | {
"login": "kit1980",
"id": 420184,
"node_id": "MDQ6VXNlcjQyMDE4NA==",
"avatar_url": "https://avatars.githubusercontent.com/u/420184?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/kit1980",
"html_url": "https://github.com/kit1980",
"followers_url": "https://api.github.com/users/kit1980/followers",
"following_url": "https://api.github.com/users/kit1980/following{/other_user}",
"gists_url": "https://api.github.com/users/kit1980/gists{/gist_id}",
"starred_url": "https://api.github.com/users/kit1980/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/kit1980/subscriptions",
"organizations_url": "https://api.github.com/users/kit1980/orgs",
"repos_url": "https://api.github.com/users/kit1980/repos",
"events_url": "https://api.github.com/users/kit1980/events{/privacy}",
"received_events_url": "https://api.github.com/users/kit1980/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 | 2023-06-28T00:08:32 | 2023-07-07T02:40:30 | 2023-07-07T02:40:30 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6930",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6930",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6930.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6930.patch",
"merged_at": "2023-07-07T02:40:30"
} | # What does this PR do ?
Fix `require_grad` typos (should be `requires_grad`).
Before the fix, the code doesn't cause any errors but doesn't do what it's supposed to do.
**Collection**: ASR, NLP
# 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
| {
"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/6930/reactions",
"total_count": 1,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 1,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/NVIDIA/NeMo/issues/6930/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6929 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6929/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6929/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6929/events | https://github.com/NVIDIA/NeMo/pull/6929 | 1,777,884,408 | PR_kwDOC_bI7s5UFeD6 | 6,929 | Fix FP8 scales conversion issue from non-nemo checkpoint to nemo checkpoint (try 2) (failed try 1: #6911) | {
"login": "sudhakarsingh27",
"id": 4879686,
"node_id": "MDQ6VXNlcjQ4Nzk2ODY=",
"avatar_url": "https://avatars.githubusercontent.com/u/4879686?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sudhakarsingh27",
"html_url": "https://github.com/sudhakarsingh27",
"followers_url": "https://api.github.com/users/sudhakarsingh27/followers",
"following_url": "https://api.github.com/users/sudhakarsingh27/following{/other_user}",
"gists_url": "https://api.github.com/users/sudhakarsingh27/gists{/gist_id}",
"starred_url": "https://api.github.com/users/sudhakarsingh27/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/sudhakarsingh27/subscriptions",
"organizations_url": "https://api.github.com/users/sudhakarsingh27/orgs",
"repos_url": "https://api.github.com/users/sudhakarsingh27/repos",
"events_url": "https://api.github.com/users/sudhakarsingh27/events{/privacy}",
"received_events_url": "https://api.github.com/users/sudhakarsingh27/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 | 2023-06-27T23:13:42 | 2023-08-12T01:42:55 | 2023-08-12T01:42:55 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6929",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6929",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6929.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6929.patch",
"merged_at": null
} | # What does this PR do ?
TLDR;
Converting a non-nemo checkpoint to NeMo with Transformer Engine modules results in truncating of scales in _extra_state. This WAR temporarily fixes the issue by manually toggling the fp8 attribute in Transformer Engine modules.
More context:
1. When exporting TE module data for checkpointing (fp8 scale info, etc.), `fp8` attribute needs to be set (see here: https://github.com/NVIDIA/TransformerEngine/blob/main/transformer_engine/pytorch/module/base.py#L344) but this attribute is only set during the forward (and when using `fp8_autocast` context manager). If the `fp8` attribute isn't set, the relevant scale information isn't included in the pickled `_extra_state` tensor (and this is the "truncation of scales" issue referred in the TLDR, basically an empty tensor is pickled which makes it seem like the scale information or `_extra_state` tensor is truncated). (Note that I tried adding the `fp8_autocast` context manager around the `model.save_to()` call and it didn't work as expected)
2. Until, there's a cleaner way to enable the `fp8` atribute for the Transformer Engine modules, this WAR allows externally enabling it before the `model.save_to()` call in the checkpoint converter script.
**Collection**: NLP
# Changelog
- Toggle the `fp8` attribute in a module if it exists (if it exists, its probably a TE module)
# 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/6929/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/6929/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6928 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6928/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6928/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6928/events | https://github.com/NVIDIA/NeMo/pull/6928 | 1,777,789,028 | PR_kwDOC_bI7s5UFI-- | 6,928 | tied weights for adapters | {
"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": ""
}
] | closed | false | null | [] | null | null | 2023-06-27T21:36:47 | 2023-08-18T21:13:43 | 2023-08-18T21:13:42 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6928",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6928",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6928.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6928.patch",
"merged_at": "2023-08-18T21:13:42"
} | # 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": "aklife97",
"id": 16078071,
"node_id": "MDQ6VXNlcjE2MDc4MDcx",
"avatar_url": "https://avatars.githubusercontent.com/u/16078071?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/aklife97",
"html_url": "https://github.com/aklife97",
"followers_url": "https://api.github.com/users/aklife97/followers",
"following_url": "https://api.github.com/users/aklife97/following{/other_user}",
"gists_url": "https://api.github.com/users/aklife97/gists{/gist_id}",
"starred_url": "https://api.github.com/users/aklife97/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/aklife97/subscriptions",
"organizations_url": "https://api.github.com/users/aklife97/orgs",
"repos_url": "https://api.github.com/users/aklife97/repos",
"events_url": "https://api.github.com/users/aklife97/events{/privacy}",
"received_events_url": "https://api.github.com/users/aklife97/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6928/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/6928/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6927 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6927/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6927/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6927/events | https://github.com/NVIDIA/NeMo/pull/6927 | 1,777,682,200 | PR_kwDOC_bI7s5UExnA | 6,927 | add rampup bs assertion | {
"login": "dimapihtar",
"id": 37850217,
"node_id": "MDQ6VXNlcjM3ODUwMjE3",
"avatar_url": "https://avatars.githubusercontent.com/u/37850217?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/dimapihtar",
"html_url": "https://github.com/dimapihtar",
"followers_url": "https://api.github.com/users/dimapihtar/followers",
"following_url": "https://api.github.com/users/dimapihtar/following{/other_user}",
"gists_url": "https://api.github.com/users/dimapihtar/gists{/gist_id}",
"starred_url": "https://api.github.com/users/dimapihtar/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/dimapihtar/subscriptions",
"organizations_url": "https://api.github.com/users/dimapihtar/orgs",
"repos_url": "https://api.github.com/users/dimapihtar/repos",
"events_url": "https://api.github.com/users/dimapihtar/events{/privacy}",
"received_events_url": "https://api.github.com/users/dimapihtar/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 | 2023-06-27T20:10:16 | 2023-06-27T22:29:27 | 2023-06-27T22:29:26 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6927",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6927",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6927.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6927.patch",
"merged_at": "2023-06-27T22:29:26"
} | # 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": "dimapihtar",
"id": 37850217,
"node_id": "MDQ6VXNlcjM3ODUwMjE3",
"avatar_url": "https://avatars.githubusercontent.com/u/37850217?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/dimapihtar",
"html_url": "https://github.com/dimapihtar",
"followers_url": "https://api.github.com/users/dimapihtar/followers",
"following_url": "https://api.github.com/users/dimapihtar/following{/other_user}",
"gists_url": "https://api.github.com/users/dimapihtar/gists{/gist_id}",
"starred_url": "https://api.github.com/users/dimapihtar/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/dimapihtar/subscriptions",
"organizations_url": "https://api.github.com/users/dimapihtar/orgs",
"repos_url": "https://api.github.com/users/dimapihtar/repos",
"events_url": "https://api.github.com/users/dimapihtar/events{/privacy}",
"received_events_url": "https://api.github.com/users/dimapihtar/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6927/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/6927/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6926 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6926/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6926/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6926/events | https://github.com/NVIDIA/NeMo/pull/6926 | 1,777,679,841 | PR_kwDOC_bI7s5UExFt | 6,926 | Fix AN4 dataset links | {
"login": "artbataev",
"id": 9380560,
"node_id": "MDQ6VXNlcjkzODA1NjA=",
"avatar_url": "https://avatars.githubusercontent.com/u/9380560?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/artbataev",
"html_url": "https://github.com/artbataev",
"followers_url": "https://api.github.com/users/artbataev/followers",
"following_url": "https://api.github.com/users/artbataev/following{/other_user}",
"gists_url": "https://api.github.com/users/artbataev/gists{/gist_id}",
"starred_url": "https://api.github.com/users/artbataev/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/artbataev/subscriptions",
"organizations_url": "https://api.github.com/users/artbataev/orgs",
"repos_url": "https://api.github.com/users/artbataev/repos",
"events_url": "https://api.github.com/users/artbataev/events{/privacy}",
"received_events_url": "https://api.github.com/users/artbataev/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": 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 | 2023-06-27T20:08:28 | 2023-06-28T21:41:54 | 2023-06-28T21:41:53 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6926",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6926",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6926.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6926.patch",
"merged_at": "2023-06-28T21:41:53"
} | # What does this PR do ?
Fix AN4 dataset link in docs. Remove broken links to AN4 data (no longer available).
**Collection**: [Docs, Tutorials]
# Changelog
- fix AN4 dataset link in documentation
- remove broken AN4 links from tutorials
# 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?
- [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.
# Additional Information
* Related to # (issue)
| {
"login": "artbataev",
"id": 9380560,
"node_id": "MDQ6VXNlcjkzODA1NjA=",
"avatar_url": "https://avatars.githubusercontent.com/u/9380560?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/artbataev",
"html_url": "https://github.com/artbataev",
"followers_url": "https://api.github.com/users/artbataev/followers",
"following_url": "https://api.github.com/users/artbataev/following{/other_user}",
"gists_url": "https://api.github.com/users/artbataev/gists{/gist_id}",
"starred_url": "https://api.github.com/users/artbataev/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/artbataev/subscriptions",
"organizations_url": "https://api.github.com/users/artbataev/orgs",
"repos_url": "https://api.github.com/users/artbataev/repos",
"events_url": "https://api.github.com/users/artbataev/events{/privacy}",
"received_events_url": "https://api.github.com/users/artbataev/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6926/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/6926/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6923 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6923/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6923/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6923/events | https://github.com/NVIDIA/NeMo/issues/6923 | 1,775,844,901 | I_kwDOC_bI7s5p2T4l | 6,923 | Is it a bug in rotary embedding implementation? | {
"login": "reminisce",
"id": 4978794,
"node_id": "MDQ6VXNlcjQ5Nzg3OTQ=",
"avatar_url": "https://avatars.githubusercontent.com/u/4978794?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/reminisce",
"html_url": "https://github.com/reminisce",
"followers_url": "https://api.github.com/users/reminisce/followers",
"following_url": "https://api.github.com/users/reminisce/following{/other_user}",
"gists_url": "https://api.github.com/users/reminisce/gists{/gist_id}",
"starred_url": "https://api.github.com/users/reminisce/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/reminisce/subscriptions",
"organizations_url": "https://api.github.com/users/reminisce/orgs",
"repos_url": "https://api.github.com/users/reminisce/repos",
"events_url": "https://api.github.com/users/reminisce/events{/privacy}",
"received_events_url": "https://api.github.com/users/reminisce/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": "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 | 2023-06-26T23:41:05 | 2023-06-27T14:05:54 | 2023-06-27T14:05:54 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
The figure below is how rotary embedding is applied to a query vector as described in the RoFormer paper, but the current implementation in NeMo seems not implementing this formula. Specifically in the [`_rotate_half` function](https://github.com/NVIDIA/NeMo/blob/82044837a7bcfa1cfd3a91f6b0014e052bdfcaf0/nemo/collections/nlp/modules/common/megatron/position_embedding/rotary_position_embedding.py#L39-L47), for an input `[A, B, C, D] -> [-C, -D, A, B]`, shouldn't the output be `[-B, A, -D, C]` instead (as the [RoFormer](https://github.com/huggingface/transformers/blob/8e164c5400b7b413c7b8fb32e35132001effc970/src/transformers/models/roformer/modeling_roformer.py#L328-L331) implemented in the Transformers library)?
| {
"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
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6923/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/6923/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6922 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6922/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6922/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6922/events | https://github.com/NVIDIA/NeMo/pull/6922 | 1,775,653,468 | PR_kwDOC_bI7s5T97GJ | 6,922 | [TTS] Add tutorial for TTS data prep scripts | {
"login": "rlangman",
"id": 9942053,
"node_id": "MDQ6VXNlcjk5NDIwNTM=",
"avatar_url": "https://avatars.githubusercontent.com/u/9942053?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/rlangman",
"html_url": "https://github.com/rlangman",
"followers_url": "https://api.github.com/users/rlangman/followers",
"following_url": "https://api.github.com/users/rlangman/following{/other_user}",
"gists_url": "https://api.github.com/users/rlangman/gists{/gist_id}",
"starred_url": "https://api.github.com/users/rlangman/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/rlangman/subscriptions",
"organizations_url": "https://api.github.com/users/rlangman/orgs",
"repos_url": "https://api.github.com/users/rlangman/repos",
"events_url": "https://api.github.com/users/rlangman/events{/privacy}",
"received_events_url": "https://api.github.com/users/rlangman/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 | 2023-06-26T21:29:28 | 2023-07-12T21:08:58 | 2023-07-12T21:08:57 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6922",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6922",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6922.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6922.patch",
"merged_at": "2023-07-12T21:08:57"
} | # What does this PR do ?
Add a tutorial demonstrating how to do the end to end data preparation and training with the new TTS preprocessing scripts and data loader.
**Collection**: [TTS]
# Changelog
- Create tutorial
# 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
- [ ] Bugfix
- [x] Documentation | {
"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/6922/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/6922/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6921 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6921/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6921/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6921/events | https://github.com/NVIDIA/NeMo/pull/6921 | 1,775,579,390 | PR_kwDOC_bI7s5T9qZJ | 6,921 | Fastpitch e2e | {
"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": ""
},
{
"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 | 2023-06-26T20:42:52 | 2023-09-09T01:43:43 | 2023-09-09T01:43:43 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6921",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6921",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6921.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6921.patch",
"merged_at": null
} | # What does this PR do ?
FastPitch + HiFi-GAN end-2-end model
**Collection**: [TTS]
# Changelog
- Add specific line by line info of high level changes in this PR.
# 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.
| {
"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/6921/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/6921/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6920 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6920/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6920/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6920/events | https://github.com/NVIDIA/NeMo/pull/6920 | 1,775,344,489 | PR_kwDOC_bI7s5T82q6 | 6,920 | Update fvad doc | {
"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": 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 | 2023-06-26T18:30:42 | 2023-06-26T21:23:36 | 2023-06-26T21:23:35 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6920",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6920",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6920.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6920.patch",
"merged_at": "2023-06-26T21:23:35"
} | # What does this PR do ?
Updated frame-VAD doc and fixed small bug | {
"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/6920/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/6920/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6919 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6919/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6919/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6919/events | https://github.com/NVIDIA/NeMo/issues/6919 | 1,774,687,593 | I_kwDOC_bI7s5px5Vp | 6,919 | libNVVM cannot be found when trying to train ASR transducer using GPU | {
"login": "gabitza-tech",
"id": 53811655,
"node_id": "MDQ6VXNlcjUzODExNjU1",
"avatar_url": "https://avatars.githubusercontent.com/u/53811655?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/gabitza-tech",
"html_url": "https://github.com/gabitza-tech",
"followers_url": "https://api.github.com/users/gabitza-tech/followers",
"following_url": "https://api.github.com/users/gabitza-tech/following{/other_user}",
"gists_url": "https://api.github.com/users/gabitza-tech/gists{/gist_id}",
"starred_url": "https://api.github.com/users/gabitza-tech/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/gabitza-tech/subscriptions",
"organizations_url": "https://api.github.com/users/gabitza-tech/orgs",
"repos_url": "https://api.github.com/users/gabitza-tech/repos",
"events_url": "https://api.github.com/users/gabitza-tech/events{/privacy}",
"received_events_url": "https://api.github.com/users/gabitza-tech/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 | 2023-06-26T12:38:26 | 2023-07-07T11:23:38 | 2023-07-07T11:23:38 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | I am trying to train the stt_en_conformer_transducer_medium on another language. I am running the train script inside a docker environment that sees the gpus and I was able to run simple pytorch lightning script for training in the same docker using gpu(on the mnist dataset, for testing). However when I run the following script:
```python
mport pytorch_lightning as pl
from omegaconf import OmegaConf
from nemo.collections.asr.models import EncDecRNNTBPEModel
from nemo.core.config import hydra_runner
from nemo.utils import logging
from nemo.utils.exp_manager import exp_manager
@hydra_runner(config_path="config/", config_name="conformer_transducer_bpe_medium.yaml")
def main(cfg):
logging.info(f'Hydra config: {OmegaConf.to_yaml(cfg)}')
trainer = pl.Trainer(**cfg.trainer)
exp_manager(trainer, cfg.get("exp_manager", None))
asr_model = EncDecRNNTBPEModel(cfg=cfg.model, trainer=trainer)
print(asr_model)
# Initialize the weights of the model from another model, if provided via config
asr_model.maybe_init_from_pretrained_checkpoint(cfg)
print(asr_model)
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
```
I get the following error when the first epoch is being started:
```
numba.core.errors.TypingError: Failed in cuda mode pipeline (step: nopython frontend)
Internal error at <numba.core.typeinfer.CallConstraint object at 0x7f7e7aadeef0>.
libNVVM cannot be found. Do `conda install cudatoolkit`:
libnvvm.so: cannot open shared object file: No such file or directory
During: resolving callee type: type(CUDADispatcher(<function exponential at 0x7f7dfacbf1c0>))
During: typing of call at /mnt/zfs-mirror-02/homes/gpirlogeanu/git_repos/nemoasr/nemo/collections/asr/parts/numba/rnnt_loss/utils/cuda_utils/reduce.py (158)
Enable logging at debug level for details.
File "../../../nemo/collections/asr/parts/numba/rnnt_loss/utils/cuda_utils/reduce.py", line 158:
def _reduce_rows(I_opid: int, R_opid: int, acts, output, num_rows: int):
<source elided>
if I_opid == 0:
curr = rnnt_helper.exponential(curr)
^
```
My gpus are seen inside the docker and by the trainer:
```
Initializing distributed: GLOBAL_RANK: 1, MEMBER: 2/2
Added key: store_based_barrier_key:1 to store for rank: 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 2 nodes.
Rank 1: Completed store-based barrier for key:store_based_barrier_key:1 with 2 nodes.
----------------------------------------------------------------------------------------------------
distributed_backend=nccl
All distributed processes registered. Starting with 2 processes
----------------------------------------------------------------------------------------------------
LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0,1]
LOCAL_RANK: 1 - CUDA_VISIBLE_DEVICES: [0,1]
[NeMo I 2023-06-26 12:32:50 modelPT:721] Optimizer config = AdamW (
Parameter Group 0
amsgrad: False
betas: [0.9, 0.98]
capturable: False
differentiable: False
eps: 1e-08
foreach: None
fused: None
lr: 5.0
maximize: False
weight_decay: 0.001
)
[NeMo I 2023-06-26 12:32:50 lr_scheduler:910] Scheduler "<nemo.core.optim.lr_scheduler.NoamAnnealing object at 0x7f741c2202e0>"
will be used during training (effective maximum steps = 2510000) -
Parameters :
(d_model: 256
warmup_steps: 10000
warmup_ratio: null
min_lr: 1.0e-06
max_steps: 2510000
)
| Name | Type | Params
------------------------------------------------------------------------
0 | preprocessor | AudioToMelSpectrogramPreprocessor | 0
1 | encoder | ConformerEncoder | 27.3 M
2 | decoder | RNNTDecoder | 3.9 M
3 | joint | RNNTJoint | 1.2 M
4 | loss | RNNTLoss | 0
5 | spec_augmentation | SpectrogramAugmentation | 0
6 | wer | RNNTBPEWER | 0
------------------------------------------------------------------------
32.5 M Trainable params
0 Non-trainable params
32.5 M Total params
64.994 Total estimated model params size (MB)
Epoch 0: 0%| | 0/6136 [00:00<?, ?it/s]Error executing job with overrides: []
Error executing job with overrides: []
```
I would greatly appreciate any help, as I have tried several scenarios, such as exporting the path to nvvm and libnvvm to NUMBAPRO_NVVM and NUMBAPRO_LIBDEVICE.
Thanks in advance!! | {
"login": "gabitza-tech",
"id": 53811655,
"node_id": "MDQ6VXNlcjUzODExNjU1",
"avatar_url": "https://avatars.githubusercontent.com/u/53811655?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/gabitza-tech",
"html_url": "https://github.com/gabitza-tech",
"followers_url": "https://api.github.com/users/gabitza-tech/followers",
"following_url": "https://api.github.com/users/gabitza-tech/following{/other_user}",
"gists_url": "https://api.github.com/users/gabitza-tech/gists{/gist_id}",
"starred_url": "https://api.github.com/users/gabitza-tech/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/gabitza-tech/subscriptions",
"organizations_url": "https://api.github.com/users/gabitza-tech/orgs",
"repos_url": "https://api.github.com/users/gabitza-tech/repos",
"events_url": "https://api.github.com/users/gabitza-tech/events{/privacy}",
"received_events_url": "https://api.github.com/users/gabitza-tech/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6919/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/6919/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6918 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6918/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6918/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6918/events | https://github.com/NVIDIA/NeMo/pull/6918 | 1,773,575,562 | PR_kwDOC_bI7s5T2-_- | 6,918 | Caching conformer fix to not mutate input tensors | {
"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": 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 | 2023-06-25T23:58:57 | 2023-07-05T18:12:54 | 2023-07-05T18:12:54 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6918",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6918",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6918.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6918.patch",
"merged_at": null
} | # What does this PR do ?
Mutating inputs are, strictly speaking, not allowed by ONNX spec.
Fixes ONNXruntime issue:
https://github.com/microsoft/onnxruntime/issues/14956
**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/6918/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/6918/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6917 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6917/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6917/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6917/events | https://github.com/NVIDIA/NeMo/issues/6917 | 1,773,323,796 | I_kwDOC_bI7s5pssYU | 6,917 | Cache aware streaming infer error with size mismatched | {
"login": "nullscc",
"id": 17700282,
"node_id": "MDQ6VXNlcjE3NzAwMjgy",
"avatar_url": "https://avatars.githubusercontent.com/u/17700282?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/nullscc",
"html_url": "https://github.com/nullscc",
"followers_url": "https://api.github.com/users/nullscc/followers",
"following_url": "https://api.github.com/users/nullscc/following{/other_user}",
"gists_url": "https://api.github.com/users/nullscc/gists{/gist_id}",
"starred_url": "https://api.github.com/users/nullscc/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/nullscc/subscriptions",
"organizations_url": "https://api.github.com/users/nullscc/orgs",
"repos_url": "https://api.github.com/users/nullscc/repos",
"events_url": "https://api.github.com/users/nullscc/events{/privacy}",
"received_events_url": "https://api.github.com/users/nullscc/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 | 2023-06-25T14:41:08 | 2023-06-26T15:01:06 | 2023-06-26T15:01:06 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
I trained a Cache aware streaming Conformer CTC model, and decode with
```
CUDA_VISIBLE_DEVICES="0" python examples/asr/asr_cache_aware_streaming/speech_to_text_cache_aware_streaming_infer.py \
--asr_model=/home/xx/NeMo/exp/libri_mini_conformer_ctc_small_streaming/Conformer-CTC-BPE/2023-06-25_01-05-04/checkpoints/Conformer-CTC-BPE.nemo \
--manifest_file=/home/xx/NeMo/data/libri_960h/dev_clean_2.json \
--batch_size=1 \
--use_amp \
--debug_mode
```
but got a error:
```
File "/home/xx/NeMo/nemo/collections/asr/parts/submodules/multi_head_attention.py", line 211, in forward
key, value, query = self.update_cache(key=key, value=value, query=query, cache=cache, cache_next=cache_next)
File "/home/xx/NeMo/nemo/collections/asr/parts/submodules/multi_head_attention.py", line 154, in update_cache
cache_next[self._cache_id, :, -q_keep_size:, :] = query[:, :q_keep_size, :]
RuntimeError: The expanded size of the tensor (9988) must match the existing size (200) at non-singleton dimension 1. Target sizes: [1, 9988, 176]. Tensor sizes: [200, 176]
```
**Steps/Code to reproduce bug**
Use lated NeMo code and use code like below to train a streaming model:
```
python examples/asr/asr_ctc/speech_to_text_ctc_bpe.py \
--config-path /home/xx/NeMo/examples/asr/conf/conformer \
--config-name conformer_ctc_bpe_small.yaml \
model.train_ds.manifest_filepath=/home/xx/NeMo/data/libri_960h/train_clean_100
model.validation_ds.manifest_filepath=/home/xx/NeMo/data/libri_960h/dev_clean.
model.tokenizer.dir=/home/xx/NeMo/data/eng_bpe_128 \ # extracted from pretrained model stt_en_conformer_ctc_small_ls
model.tokenizer.type=bpe \
model.optim.lr=0.1 \
model.optim.sched.warmup_steps=20000 \
trainer.devices=2 \
trainer.accelerator="gpu" \
trainer.strategy="ddp" \
trainer.max_epochs=100 \
exp_manager.exp_dir=/home/xx/NeMo/exp/libri_mini_conformer_ctc_small_streaming
```
and then use code like below to decode:
```
CUDA_VISIBLE_DEVICES="0" python examples/asr/asr_cache_aware_streaming/speech_to_text_cache_aware_streaming_infer.py \
--asr_model=/home/xx/NeMo/exp/libri_mini_conformer_ctc_small_streaming/Conformer-CTC-BPE/2023-06-25_01-05-04/checkpoints/Conformer-CTC-BPE.nemo \
--manifest_file=/home/xx/NeMo/data/libri_960h/dev_clean_2.json \
--batch_size=1 \
--use_amp \
--debug_mode
```
**Expected behavior**
Normal decoding.
**Environment details**
If NVIDIA docker image is used you don't need to specify these.
Otherwise, please provide:
- Red Hat 8.4.1-1
- PyTorch 1.11.0
- Python 3.8.17
**Additional context**
full log:
[xx.log](https://github.com/NVIDIA/NeMo/files/11859985/xx.log)
| {
"login": "nullscc",
"id": 17700282,
"node_id": "MDQ6VXNlcjE3NzAwMjgy",
"avatar_url": "https://avatars.githubusercontent.com/u/17700282?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/nullscc",
"html_url": "https://github.com/nullscc",
"followers_url": "https://api.github.com/users/nullscc/followers",
"following_url": "https://api.github.com/users/nullscc/following{/other_user}",
"gists_url": "https://api.github.com/users/nullscc/gists{/gist_id}",
"starred_url": "https://api.github.com/users/nullscc/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/nullscc/subscriptions",
"organizations_url": "https://api.github.com/users/nullscc/orgs",
"repos_url": "https://api.github.com/users/nullscc/repos",
"events_url": "https://api.github.com/users/nullscc/events{/privacy}",
"received_events_url": "https://api.github.com/users/nullscc/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6917/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/6917/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6916 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6916/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6916/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6916/events | https://github.com/NVIDIA/NeMo/pull/6916 | 1,772,404,239 | PR_kwDOC_bI7s5TzGXe | 6,916 | ptuning oom fix | {
"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": ""
}
] | closed | false | null | [] | null | null | 2023-06-24T02:31:25 | 2023-06-27T22:16:35 | 2023-06-27T22:16:34 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6916",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6916",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6916.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6916.patch",
"merged_at": "2023-06-27T22:16:34"
} | # What does this PR do ?
The prompt learning dataset was not padding sequences to a multiple of 8 which forced a non-optimized version of softmax to be used resulting in more memory use and OOM errors. This PR pads to a multiple of 8 to reduce memory.
**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": "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/6916/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/6916/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6915 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6915/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6915/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6915/events | https://github.com/NVIDIA/NeMo/pull/6915 | 1,772,384,964 | PR_kwDOC_bI7s5TzCP4 | 6,915 | Fix gpt trainer test | {
"login": "hsiehjackson",
"id": 37269846,
"node_id": "MDQ6VXNlcjM3MjY5ODQ2",
"avatar_url": "https://avatars.githubusercontent.com/u/37269846?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/hsiehjackson",
"html_url": "https://github.com/hsiehjackson",
"followers_url": "https://api.github.com/users/hsiehjackson/followers",
"following_url": "https://api.github.com/users/hsiehjackson/following{/other_user}",
"gists_url": "https://api.github.com/users/hsiehjackson/gists{/gist_id}",
"starred_url": "https://api.github.com/users/hsiehjackson/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/hsiehjackson/subscriptions",
"organizations_url": "https://api.github.com/users/hsiehjackson/orgs",
"repos_url": "https://api.github.com/users/hsiehjackson/repos",
"events_url": "https://api.github.com/users/hsiehjackson/events{/privacy}",
"received_events_url": "https://api.github.com/users/hsiehjackson/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": ""
},
{
"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 | 2023-06-24T01:34:25 | 2023-08-10T19:44:01 | 2023-08-10T19:44:00 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6915",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6915",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6915.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6915.patch",
"merged_at": "2023-08-10T19:44:00"
} | # What does this PR do ?
- Fix gpt trainer inference_step to compute generation results during validation and test
- Add dataset with metadata, to output in prediction file
- Add token_f1 metrics for Scrolls
**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**:
- [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
- [ ] 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": "hsiehjackson",
"id": 37269846,
"node_id": "MDQ6VXNlcjM3MjY5ODQ2",
"avatar_url": "https://avatars.githubusercontent.com/u/37269846?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/hsiehjackson",
"html_url": "https://github.com/hsiehjackson",
"followers_url": "https://api.github.com/users/hsiehjackson/followers",
"following_url": "https://api.github.com/users/hsiehjackson/following{/other_user}",
"gists_url": "https://api.github.com/users/hsiehjackson/gists{/gist_id}",
"starred_url": "https://api.github.com/users/hsiehjackson/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/hsiehjackson/subscriptions",
"organizations_url": "https://api.github.com/users/hsiehjackson/orgs",
"repos_url": "https://api.github.com/users/hsiehjackson/repos",
"events_url": "https://api.github.com/users/hsiehjackson/events{/privacy}",
"received_events_url": "https://api.github.com/users/hsiehjackson/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6915/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/6915/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6914 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6914/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6914/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6914/events | https://github.com/NVIDIA/NeMo/pull/6914 | 1,771,779,639 | PR_kwDOC_bI7s5Tw_VN | 6,914 | Documentation for ONNX export of Megatron Models | {
"login": "asfiyab-nvidia",
"id": 117682710,
"node_id": "U_kgDOBwOyFg",
"avatar_url": "https://avatars.githubusercontent.com/u/117682710?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/asfiyab-nvidia",
"html_url": "https://github.com/asfiyab-nvidia",
"followers_url": "https://api.github.com/users/asfiyab-nvidia/followers",
"following_url": "https://api.github.com/users/asfiyab-nvidia/following{/other_user}",
"gists_url": "https://api.github.com/users/asfiyab-nvidia/gists{/gist_id}",
"starred_url": "https://api.github.com/users/asfiyab-nvidia/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/asfiyab-nvidia/subscriptions",
"organizations_url": "https://api.github.com/users/asfiyab-nvidia/orgs",
"repos_url": "https://api.github.com/users/asfiyab-nvidia/repos",
"events_url": "https://api.github.com/users/asfiyab-nvidia/events{/privacy}",
"received_events_url": "https://api.github.com/users/asfiyab-nvidia/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 | 2023-06-23T17:03:41 | 2023-06-25T21:20:00 | 2023-06-25T21:20:00 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6914",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6914",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6914.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6914.patch",
"merged_at": "2023-06-25T21:20:00"
} | # What does this PR do ?
Adds documentation for ONNX export of Megatron Models
| {
"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/6914/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/6914/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6913 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6913/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6913/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6913/events | https://github.com/NVIDIA/NeMo/pull/6913 | 1,771,746,244 | PR_kwDOC_bI7s5Tw35G | 6,913 | Update container info in README.rst | {
"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
} | [] | closed | false | null | [] | null | null | 2023-06-23T16:40:19 | 2023-06-23T17:49:28 | 2023-06-23T17:49:27 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6913",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6913",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6913.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6913.patch",
"merged_at": "2023-06-23T17:49:27"
} | # 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": "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/6913/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/6913/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6911 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6911/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6911/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6911/events | https://github.com/NVIDIA/NeMo/pull/6911 | 1,770,895,480 | PR_kwDOC_bI7s5TuC_X | 6,911 | Fix FP8 scales conversion issue from non-nemo checkpoint to nemo checkpoint | {
"login": "sudhakarsingh27",
"id": 4879686,
"node_id": "MDQ6VXNlcjQ4Nzk2ODY=",
"avatar_url": "https://avatars.githubusercontent.com/u/4879686?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sudhakarsingh27",
"html_url": "https://github.com/sudhakarsingh27",
"followers_url": "https://api.github.com/users/sudhakarsingh27/followers",
"following_url": "https://api.github.com/users/sudhakarsingh27/following{/other_user}",
"gists_url": "https://api.github.com/users/sudhakarsingh27/gists{/gist_id}",
"starred_url": "https://api.github.com/users/sudhakarsingh27/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/sudhakarsingh27/subscriptions",
"organizations_url": "https://api.github.com/users/sudhakarsingh27/orgs",
"repos_url": "https://api.github.com/users/sudhakarsingh27/repos",
"events_url": "https://api.github.com/users/sudhakarsingh27/events{/privacy}",
"received_events_url": "https://api.github.com/users/sudhakarsingh27/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 | 2023-06-23T06:58:32 | 2023-06-27T23:15:25 | 2023-06-27T23:15:24 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6911",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6911",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6911.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6911.patch",
"merged_at": null
} | # What does this PR do ?
Converting a non-nemo checkpoint to NeMo with Transformer Engine modules results in truncating of scales in `_extra_state`. This WAR temporarily fixes the issue by manually toggling the `fp8` attribute in TE modules.
**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": "sudhakarsingh27",
"id": 4879686,
"node_id": "MDQ6VXNlcjQ4Nzk2ODY=",
"avatar_url": "https://avatars.githubusercontent.com/u/4879686?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sudhakarsingh27",
"html_url": "https://github.com/sudhakarsingh27",
"followers_url": "https://api.github.com/users/sudhakarsingh27/followers",
"following_url": "https://api.github.com/users/sudhakarsingh27/following{/other_user}",
"gists_url": "https://api.github.com/users/sudhakarsingh27/gists{/gist_id}",
"starred_url": "https://api.github.com/users/sudhakarsingh27/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/sudhakarsingh27/subscriptions",
"organizations_url": "https://api.github.com/users/sudhakarsingh27/orgs",
"repos_url": "https://api.github.com/users/sudhakarsingh27/repos",
"events_url": "https://api.github.com/users/sudhakarsingh27/events{/privacy}",
"received_events_url": "https://api.github.com/users/sudhakarsingh27/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6911/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/6911/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6910 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6910/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6910/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6910/events | https://github.com/NVIDIA/NeMo/issues/6910 | 1,770,884,879 | I_kwDOC_bI7s5pjY8P | 6,910 | Creating an inference server for NeMo GPT-5B leads to a bus error during model conversion | {
"login": "rohitdwivedula",
"id": 25080952,
"node_id": "MDQ6VXNlcjI1MDgwOTUy",
"avatar_url": "https://avatars.githubusercontent.com/u/25080952?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/rohitdwivedula",
"html_url": "https://github.com/rohitdwivedula",
"followers_url": "https://api.github.com/users/rohitdwivedula/followers",
"following_url": "https://api.github.com/users/rohitdwivedula/following{/other_user}",
"gists_url": "https://api.github.com/users/rohitdwivedula/gists{/gist_id}",
"starred_url": "https://api.github.com/users/rohitdwivedula/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/rohitdwivedula/subscriptions",
"organizations_url": "https://api.github.com/users/rohitdwivedula/orgs",
"repos_url": "https://api.github.com/users/rohitdwivedula/repos",
"events_url": "https://api.github.com/users/rohitdwivedula/events{/privacy}",
"received_events_url": "https://api.github.com/users/rohitdwivedula/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 | 2023-06-23T06:49:04 | 2023-09-04T01:44:33 | 2023-09-04T01:44:33 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
Attempting to follow instructions from a post on the Nvidia developer blog: [Deploying a 1.3B GPT-3 Model with NVIDIA NeMo Framework](https://developer.nvidia.com/blog/deploying-a-1-3b-gpt-3-model-with-nvidia-nemo-megatron/). The instructions in the article work perfectly when using the 1.3B param pretrained model. However, using the same instructions for 5B leads to an error.
**Steps/Code to reproduce bug**
1. Download the 5B parameter pretrained model using:
```mkdir ~/nemo && cd ~/nemo && wget https://huggingface.co/nvidia/nemo-megatron-gpt-5B/resolve/main/nemo_gpt5B_fp16_tp1.nemo```
2. Run the following command:
```
docker run --rm \
--gpus all \
--shm-size=16GB \
-v /home/$USER/nemo/:/checkpoints \
-v /home/$USER/nemo/outputs-5B/:/model_repository \
nvcr.io/ea-bignlp/bignlp-training:22.08.01-py3 \
bash -c 'export PYTHONPATH=/opt/bignlp/FasterTransformer:${PYTHONPATH} && \
cd /opt/bignlp && \
python3 FasterTransformer/examples/pytorch/gpt/utils/nemo_ckpt_convert.py \
--in-file /checkpoints/nemo_gpt5B_fp16_tp1.nemo \
--infer-gpu-num 1 \
--saved-dir /model_repository/gpt3_5b \
--weight-data-type fp16 \
--load-checkpoints-to-cpu 0 && \
python3 /opt/bignlp/bignlp-scripts/bignlp/collections/export_scripts/prepare_triton_model_config.py \
--model-train-name gpt3_5b \
--template-path /opt/bignlp/fastertransformer_backend/all_models/gpt/fastertransformer/config.pbtxt \
--ft-checkpoint /model_repository/gpt3_5b/1-gpu \
--config-path /model_repository/gpt3_5b/config.pbtxt \
--max-batch-size 256 \
--pipeline-model-parallel-size 1 \
--tensor-model-parallel-size 1 \
--data-type bf16'
```
This command fails with the following output:
```
This container image and its contents are governed by the NVIDIA Deep Learning Container License.
By pulling and using the container, you accept the terms and conditions of this license:
https://developer.nvidia.com/ngc/nvidia-deep-learning-container-license
NOTE: CUDA Forward Compatibility mode ENABLED.
Using CUDA 11.7 driver version 515.48.08 with kernel driver version 470.182.03.
See https://docs.nvidia.com/deploy/cuda-compatibility/ for details.
2023-06-23 06:31:41,635 __main__ [INFO] Spent 0:01:16.565989 (h:m:s) to unpack NeMo archive
=============== Argument ===============
saved_dir: /model_repository/gpt3_5b
in_file: /checkpoints/nemo_gpt5B_fp16_tp1.nemo
prompt_in_file: None
prompt_saved_dir: None
infer_gpu_num: 1
fused_qkv: 1
processes: 16
weight_data_type: fp16
load_checkpoints_to_cpu: 0
vocab_path: None
merges_path: None
tokenizer_model_path: None
verbose: False
========================================
bash: line 16: 102 Bus error (core dumped) python3 FasterTransformer/examples/pytorch/gpt/utils/nemo_ckpt_convert.py --in-file /checkpoints/nemo_gpt5B_fp16_tp1.nemo --infer-gpu-num 1 --saved-dir /model_repository/gpt3_5b --weight-data-type fp16 --load-checkpoints-to-cpu 0
```
**Expected behavior**
An inference model should have been saved to `~/nemo/outputs-5B`
**Environment overview (please complete the following information)**
- Environment location: Docker
- Docker images used:
```
docker login nvcr.io
docker pull nvcr.io/ea-bignlp/bignlp-inference:22.08-py3
docker pull nvcr.io/ea-bignlp/bignlp-training:22.08.01-py3
```
**Additional context**
Attempting to run this on a machine that has ~200+GB RAM and an A100 GPU. | {
"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/6910/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/6910/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6909 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6909/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6909/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6909/events | https://github.com/NVIDIA/NeMo/pull/6909 | 1,770,715,590 | PR_kwDOC_bI7s5TtcXd | 6,909 | Fix fast-glu activation in change partitions | {
"login": "hsiehjackson",
"id": 37269846,
"node_id": "MDQ6VXNlcjM3MjY5ODQ2",
"avatar_url": "https://avatars.githubusercontent.com/u/37269846?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/hsiehjackson",
"html_url": "https://github.com/hsiehjackson",
"followers_url": "https://api.github.com/users/hsiehjackson/followers",
"following_url": "https://api.github.com/users/hsiehjackson/following{/other_user}",
"gists_url": "https://api.github.com/users/hsiehjackson/gists{/gist_id}",
"starred_url": "https://api.github.com/users/hsiehjackson/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/hsiehjackson/subscriptions",
"organizations_url": "https://api.github.com/users/hsiehjackson/orgs",
"repos_url": "https://api.github.com/users/hsiehjackson/repos",
"events_url": "https://api.github.com/users/hsiehjackson/events{/privacy}",
"received_events_url": "https://api.github.com/users/hsiehjackson/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 | 2023-06-23T03:24:08 | 2023-06-23T19:19:21 | 2023-06-23T19:19:19 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6909",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6909",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6909.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6909.patch",
"merged_at": "2023-06-23T19:19:19"
} | # What does this PR do ?
Change `swiglu` to all fast glu activation for partition conversion.
**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**:
- [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": "hsiehjackson",
"id": 37269846,
"node_id": "MDQ6VXNlcjM3MjY5ODQ2",
"avatar_url": "https://avatars.githubusercontent.com/u/37269846?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/hsiehjackson",
"html_url": "https://github.com/hsiehjackson",
"followers_url": "https://api.github.com/users/hsiehjackson/followers",
"following_url": "https://api.github.com/users/hsiehjackson/following{/other_user}",
"gists_url": "https://api.github.com/users/hsiehjackson/gists{/gist_id}",
"starred_url": "https://api.github.com/users/hsiehjackson/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/hsiehjackson/subscriptions",
"organizations_url": "https://api.github.com/users/hsiehjackson/orgs",
"repos_url": "https://api.github.com/users/hsiehjackson/repos",
"events_url": "https://api.github.com/users/hsiehjackson/events{/privacy}",
"received_events_url": "https://api.github.com/users/hsiehjackson/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6909/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/6909/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6908 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6908/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6908/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6908/events | https://github.com/NVIDIA/NeMo/pull/6908 | 1,770,412,529 | PR_kwDOC_bI7s5TsgBQ | 6,908 | Make sure asr_model.change_attention_model is run if either cfg.model_path or cfg.pretrained_name is specified | {
"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
} | [
{
"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 | 2023-06-22T21:21:43 | 2023-06-26T16:23:59 | 2023-06-26T16:23:58 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6908",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6908",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6908.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6908.patch",
"merged_at": "2023-06-26T16:23:58"
} | # What does this PR do ?
Makes sure `asr_model.change_attention_model` is run regardless of whether `cfg.model_path` or `cfg.pretrained_name` is specified.
**Collection**: ASR
# Changelog
- Moved `asr_model.change_attention_model` call outside of if-branch.
# Usage
* You can potentially add a usage example below
```bash
python NeMo/examples/asr/transcribe_speech.py \
pretrained_name="stt_en_fastconformer_ctc_large" \
dataset_manifest=$src_manifest \
++model_change.conformer.self_attention_model="rel_pos_local_attn" \
++model_change.conformer.att_context_size="[128, 128]"
```
# 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?
- [N/A] 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
- [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": "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
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6908/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/6908/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6907 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6907/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6907/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6907/events | https://github.com/NVIDIA/NeMo/pull/6907 | 1,770,294,823 | PR_kwDOC_bI7s5TsGZ3 | 6,907 | Reduce number of nodes produced in ConformerEncoder export | {
"login": "ilyasher",
"id": 46343317,
"node_id": "MDQ6VXNlcjQ2MzQzMzE3",
"avatar_url": "https://avatars.githubusercontent.com/u/46343317?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ilyasher",
"html_url": "https://github.com/ilyasher",
"followers_url": "https://api.github.com/users/ilyasher/followers",
"following_url": "https://api.github.com/users/ilyasher/following{/other_user}",
"gists_url": "https://api.github.com/users/ilyasher/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ilyasher/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ilyasher/subscriptions",
"organizations_url": "https://api.github.com/users/ilyasher/orgs",
"repos_url": "https://api.github.com/users/ilyasher/repos",
"events_url": "https://api.github.com/users/ilyasher/events{/privacy}",
"received_events_url": "https://api.github.com/users/ilyasher/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 | 2023-06-22T19:45:41 | 2023-06-26T22:02:59 | 2023-06-26T22:02:58 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6907",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6907",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6907.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6907.patch",
"merged_at": null
} | This commit breaks up the output cache tensor into a list of tensors. Each layer which writes to the output cache now writes to its own tensor. The output cache tensors are then stacked into one tensor before being returned, so ConformerEncoder still has the same inputs and outputs as before.
The reason for this change is from an ONNX perspective, updating a tensor creates a new copy with the changes. So minimizing the sizes of the tensors that get updated makes the exported ONNX model more efficient.
The number of nodes produced in the ONNX export is also reduced. In the example I tested, the exported ONNX model had 11.484 nodes before this change, and only 4,892 nodes after this change.
Signed off by: Ilya Sherstyuk <[email protected]>
| {
"login": "ilyasher",
"id": 46343317,
"node_id": "MDQ6VXNlcjQ2MzQzMzE3",
"avatar_url": "https://avatars.githubusercontent.com/u/46343317?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ilyasher",
"html_url": "https://github.com/ilyasher",
"followers_url": "https://api.github.com/users/ilyasher/followers",
"following_url": "https://api.github.com/users/ilyasher/following{/other_user}",
"gists_url": "https://api.github.com/users/ilyasher/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ilyasher/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ilyasher/subscriptions",
"organizations_url": "https://api.github.com/users/ilyasher/orgs",
"repos_url": "https://api.github.com/users/ilyasher/repos",
"events_url": "https://api.github.com/users/ilyasher/events{/privacy}",
"received_events_url": "https://api.github.com/users/ilyasher/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6907/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/6907/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6906 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6906/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6906/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6906/events | https://github.com/NVIDIA/NeMo/pull/6906 | 1,769,929,568 | PR_kwDOC_bI7s5Tq2fP | 6,906 | Add hybrid model support to transcribe_speech_parallel.py | {
"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": 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 | 2023-06-22T15:27:55 | 2023-06-22T18:37:09 | 2023-06-22T18:37:08 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6906",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6906",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6906.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6906.patch",
"merged_at": "2023-06-22T18:37:08"
} | # What does this PR do ?
Add hybrid model support to transcribe_speech_parallel.py
**Collection**: [ASR]
| {
"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/6906/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/6906/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6905 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6905/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6905/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6905/events | https://github.com/NVIDIA/NeMo/issues/6905 | 1,769,829,824 | I_kwDOC_bI7s5pfXXA | 6,905 | Support about Streaming inference of Conformer Transducer? | {
"login": "nullscc",
"id": 17700282,
"node_id": "MDQ6VXNlcjE3NzAwMjgy",
"avatar_url": "https://avatars.githubusercontent.com/u/17700282?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/nullscc",
"html_url": "https://github.com/nullscc",
"followers_url": "https://api.github.com/users/nullscc/followers",
"following_url": "https://api.github.com/users/nullscc/following{/other_user}",
"gists_url": "https://api.github.com/users/nullscc/gists{/gist_id}",
"starred_url": "https://api.github.com/users/nullscc/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/nullscc/subscriptions",
"organizations_url": "https://api.github.com/users/nullscc/orgs",
"repos_url": "https://api.github.com/users/nullscc/repos",
"events_url": "https://api.github.com/users/nullscc/events{/privacy}",
"received_events_url": "https://api.github.com/users/nullscc/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 | 2023-06-22T14:40:58 | 2023-08-22T01:44:04 | 2023-08-22T01:44:04 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | Does Nemo support about streaming inference of Conformer Transducer?
From my view, as long as Conformer Transducer trained using Cache-aware Streaming Conformer, then it can be used on streaming situation.
But from this: https://github.com/NVIDIA/NeMo/issues/3144#issuecomment-962420001, it seems we can train a Cache-aware Streaming based Conformer Transducer model to use it in streaming situation.
So my question is:
1. Can I train a Cache-aware Streaming based Conformer Transducer model using https://github.com/NVIDIA/NeMo/blob/main/examples/asr/conf/conformer/cache_aware_streaming/conformer_transducer_bpe_streaming.yaml?
2. Can I inference using script from https://github.com/NVIDIA/NeMo/blob/main/examples/asr/asr_cache_aware_streaming/speech_to_text_cache_aware_streaming_infer.py with a Confomer Transducer model, which means does speech_to_text_cache_aware_streaming_infer.py support Cache-aware Streaming based Conformer Transducer model?
3. If I want use a Conformer Transducer model in streaming situation, what can I do for now? Because from Librispeech 960 experiments, it seems Conformer Transducer is best with respect to WER?
Thanks for any tips. | {
"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/6905/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/6905/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6904 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6904/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6904/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6904/events | https://github.com/NVIDIA/NeMo/pull/6904 | 1,769,787,509 | PR_kwDOC_bI7s5TqZn- | 6,904 | Make Gradio library optional | {
"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 | 2023-06-22T14:19:18 | 2023-06-22T19:27:46 | 2023-06-22T19:27:45 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6904",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6904",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6904.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6904.patch",
"merged_at": "2023-06-22T19:27:45"
} | Make gradio library optional | {
"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/6904/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/6904/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6903 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6903/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6903/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6903/events | https://github.com/NVIDIA/NeMo/issues/6903 | 1,768,890,773 | I_kwDOC_bI7s5pbyGV | 6,903 | Support BLOOM for finetuning? | {
"login": "dat-browny",
"id": 73813209,
"node_id": "MDQ6VXNlcjczODEzMjA5",
"avatar_url": "https://avatars.githubusercontent.com/u/73813209?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/dat-browny",
"html_url": "https://github.com/dat-browny",
"followers_url": "https://api.github.com/users/dat-browny/followers",
"following_url": "https://api.github.com/users/dat-browny/following{/other_user}",
"gists_url": "https://api.github.com/users/dat-browny/gists{/gist_id}",
"starred_url": "https://api.github.com/users/dat-browny/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/dat-browny/subscriptions",
"organizations_url": "https://api.github.com/users/dat-browny/orgs",
"repos_url": "https://api.github.com/users/dat-browny/repos",
"events_url": "https://api.github.com/users/dat-browny/events{/privacy}",
"received_events_url": "https://api.github.com/users/dat-browny/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 | {
"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 | 2023-06-22T04:05:32 | 2023-07-31T01:51:47 | 2023-07-31T01:51:47 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | I noticed that you referenced the ALiBi in the flash-attn module, but I couldn't find any mention or description of its usage in your project. I'm curious to know if you have incorporated support for models like BLOOM for fine-tuning and converting their checkpoints from Huggingface format to .nemo or MegatronLM. | {
"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/6903/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/6903/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6902 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6902/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6902/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6902/events | https://github.com/NVIDIA/NeMo/pull/6902 | 1,768,413,989 | PR_kwDOC_bI7s5TlsMm | 6,902 | Update Frame-VAD doc | {
"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": 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 | 2023-06-21T20:22:40 | 2023-06-22T23:06:54 | 2023-06-22T23:06:52 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6902",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6902",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6902.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6902.patch",
"merged_at": "2023-06-22T23:06:52"
} | # What does this PR do ?
Update doc for Frame-VAD
**Collection**: [ASR]
| {
"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/6902/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/6902/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6901 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6901/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6901/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6901/events | https://github.com/NVIDIA/NeMo/pull/6901 | 1,768,359,410 | PR_kwDOC_bI7s5TlgGb | 6,901 | Fix flash-attention | {
"login": "hsiehjackson",
"id": 37269846,
"node_id": "MDQ6VXNlcjM3MjY5ODQ2",
"avatar_url": "https://avatars.githubusercontent.com/u/37269846?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/hsiehjackson",
"html_url": "https://github.com/hsiehjackson",
"followers_url": "https://api.github.com/users/hsiehjackson/followers",
"following_url": "https://api.github.com/users/hsiehjackson/following{/other_user}",
"gists_url": "https://api.github.com/users/hsiehjackson/gists{/gist_id}",
"starred_url": "https://api.github.com/users/hsiehjackson/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/hsiehjackson/subscriptions",
"organizations_url": "https://api.github.com/users/hsiehjackson/orgs",
"repos_url": "https://api.github.com/users/hsiehjackson/repos",
"events_url": "https://api.github.com/users/hsiehjackson/events{/privacy}",
"received_events_url": "https://api.github.com/users/hsiehjackson/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 | 2023-06-21T19:45:57 | 2023-06-26T22:06:30 | 2023-06-26T22:06:29 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6901",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6901",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6901.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6901.patch",
"merged_at": "2023-06-26T22:06:29"
} | # What does this PR do ?
1. Set default `apply_query_key_layer_scaling=False`
2. Fix flash_attention_triton attention_bias broadcast issue (reduce memory usage)
3. Add testing for flash cross-attention suggested by @yidong72
**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**:
- [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?
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": "hsiehjackson",
"id": 37269846,
"node_id": "MDQ6VXNlcjM3MjY5ODQ2",
"avatar_url": "https://avatars.githubusercontent.com/u/37269846?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/hsiehjackson",
"html_url": "https://github.com/hsiehjackson",
"followers_url": "https://api.github.com/users/hsiehjackson/followers",
"following_url": "https://api.github.com/users/hsiehjackson/following{/other_user}",
"gists_url": "https://api.github.com/users/hsiehjackson/gists{/gist_id}",
"starred_url": "https://api.github.com/users/hsiehjackson/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/hsiehjackson/subscriptions",
"organizations_url": "https://api.github.com/users/hsiehjackson/orgs",
"repos_url": "https://api.github.com/users/hsiehjackson/repos",
"events_url": "https://api.github.com/users/hsiehjackson/events{/privacy}",
"received_events_url": "https://api.github.com/users/hsiehjackson/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6901/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/6901/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6900 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6900/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6900/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6900/events | https://github.com/NVIDIA/NeMo/pull/6900 | 1,768,202,694 | PR_kwDOC_bI7s5TlFY9 | 6,900 | removed some tests | {
"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": 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 | 2023-06-21T18:45:12 | 2023-06-21T20:55:49 | 2023-06-21T20:55:48 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6900",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6900",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6900.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6900.patch",
"merged_at": "2023-06-21T20:55:48"
} | # 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/6900/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/6900/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6899 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6899/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6899/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6899/events | https://github.com/NVIDIA/NeMo/pull/6899 | 1,768,127,545 | PR_kwDOC_bI7s5Tk1Qb | 6,899 | Fix transcribe_utils.py for hybrid models in partial transcribe mode | {
"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": 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 | 2023-06-21T17:58:30 | 2023-06-21T21:41:33 | 2023-06-21T21:41:32 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6899",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6899",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6899.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6899.patch",
"merged_at": "2023-06-21T21:41:32"
} | The bug is that the hybrid-ctc model has a slightly different transcribe function implementation from the original CTC model. | {
"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/6899/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/6899/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6898 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6898/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6898/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6898/events | https://github.com/NVIDIA/NeMo/pull/6898 | 1,768,123,110 | PR_kwDOC_bI7s5Tk0TS | 6,898 | Enable `rpe` methods in bert-like models | {
"login": "sararb",
"id": 17721108,
"node_id": "MDQ6VXNlcjE3NzIxMTA4",
"avatar_url": "https://avatars.githubusercontent.com/u/17721108?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sararb",
"html_url": "https://github.com/sararb",
"followers_url": "https://api.github.com/users/sararb/followers",
"following_url": "https://api.github.com/users/sararb/following{/other_user}",
"gists_url": "https://api.github.com/users/sararb/gists{/gist_id}",
"starred_url": "https://api.github.com/users/sararb/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/sararb/subscriptions",
"organizations_url": "https://api.github.com/users/sararb/orgs",
"repos_url": "https://api.github.com/users/sararb/repos",
"events_url": "https://api.github.com/users/sararb/events{/privacy}",
"received_events_url": "https://api.github.com/users/sararb/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 | {
"login": "sararb",
"id": 17721108,
"node_id": "MDQ6VXNlcjE3NzIxMTA4",
"avatar_url": "https://avatars.githubusercontent.com/u/17721108?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sararb",
"html_url": "https://github.com/sararb",
"followers_url": "https://api.github.com/users/sararb/followers",
"following_url": "https://api.github.com/users/sararb/following{/other_user}",
"gists_url": "https://api.github.com/users/sararb/gists{/gist_id}",
"starred_url": "https://api.github.com/users/sararb/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/sararb/subscriptions",
"organizations_url": "https://api.github.com/users/sararb/orgs",
"repos_url": "https://api.github.com/users/sararb/repos",
"events_url": "https://api.github.com/users/sararb/events{/privacy}",
"received_events_url": "https://api.github.com/users/sararb/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "sararb",
"id": 17721108,
"node_id": "MDQ6VXNlcjE3NzIxMTA4",
"avatar_url": "https://avatars.githubusercontent.com/u/17721108?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sararb",
"html_url": "https://github.com/sararb",
"followers_url": "https://api.github.com/users/sararb/followers",
"following_url": "https://api.github.com/users/sararb/following{/other_user}",
"gists_url": "https://api.github.com/users/sararb/gists{/gist_id}",
"starred_url": "https://api.github.com/users/sararb/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/sararb/subscriptions",
"organizations_url": "https://api.github.com/users/sararb/orgs",
"repos_url": "https://api.github.com/users/sararb/repos",
"events_url": "https://api.github.com/users/sararb/events{/privacy}",
"received_events_url": "https://api.github.com/users/sararb/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | null | 2023-06-21T17:54:58 | 2023-06-28T21:19:08 | 2023-06-28T21:19:01 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6898",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6898",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6898.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6898.patch",
"merged_at": "2023-06-28T21:19:01"
} | # What does this PR do ?
- Add extrapolatable position embedding to bert-like model, based on the recent changes introduced in #6666
- I am already using this change in the BioNeMo framework and running experiments to test how `rpe` techniques help bert-like model to extrapolate to longer sequences.
**Collection**: [NLP]
# Changelog
- BERT:
- nemo/collections/nlp/models/language_modeling/megatron_bert_model.py
- nemo/collections/nlp/models/language_modeling/megatron/bert_model.py
# Usage
```python
model.position_embedding_type='' # Options ['learned_absolute', 'rope', 'alibi', 'xpos', 'sandwich', 'kerple']
```
# 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": "sararb",
"id": 17721108,
"node_id": "MDQ6VXNlcjE3NzIxMTA4",
"avatar_url": "https://avatars.githubusercontent.com/u/17721108?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sararb",
"html_url": "https://github.com/sararb",
"followers_url": "https://api.github.com/users/sararb/followers",
"following_url": "https://api.github.com/users/sararb/following{/other_user}",
"gists_url": "https://api.github.com/users/sararb/gists{/gist_id}",
"starred_url": "https://api.github.com/users/sararb/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/sararb/subscriptions",
"organizations_url": "https://api.github.com/users/sararb/orgs",
"repos_url": "https://api.github.com/users/sararb/repos",
"events_url": "https://api.github.com/users/sararb/events{/privacy}",
"received_events_url": "https://api.github.com/users/sararb/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6898/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/6898/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6897 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6897/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6897/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6897/events | https://github.com/NVIDIA/NeMo/pull/6897 | 1,767,372,695 | PR_kwDOC_bI7s5TiRf_ | 6,897 | hot fix SDE | {
"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 | null | [] | null | null | 2023-06-21T11:26:25 | 2023-06-22T14:16:50 | 2023-06-22T14:16:49 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6897",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6897",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6897.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6897.patch",
"merged_at": "2023-06-22T14:16:49"
} | # What does this PR do ?
Got report about unexpected behaviour. fixed
**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": "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/6897/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/6897/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6896 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6896/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6896/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6896/events | https://github.com/NVIDIA/NeMo/issues/6896 | 1,767,360,767 | I_kwDOC_bI7s5pV8j_ | 6,896 | Best way to get started for real-time ASR through microphone | {
"login": "Caet-pip",
"id": 121129243,
"node_id": "U_kgDOBzhJGw",
"avatar_url": "https://avatars.githubusercontent.com/u/121129243?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Caet-pip",
"html_url": "https://github.com/Caet-pip",
"followers_url": "https://api.github.com/users/Caet-pip/followers",
"following_url": "https://api.github.com/users/Caet-pip/following{/other_user}",
"gists_url": "https://api.github.com/users/Caet-pip/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Caet-pip/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Caet-pip/subscriptions",
"organizations_url": "https://api.github.com/users/Caet-pip/orgs",
"repos_url": "https://api.github.com/users/Caet-pip/repos",
"events_url": "https://api.github.com/users/Caet-pip/events{/privacy}",
"received_events_url": "https://api.github.com/users/Caet-pip/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 | 2023-06-21T11:18:28 | 2024-06-25T11:31:29 | 2023-08-22T01:44:05 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | Hello,
I want to implement real time inferencing on microphone audio and saw that the Cache-aware Streaming Conformer models are being used for this in NeMo, I wanted to know if there are any available models for real-time transcription though mic and any example scripts to implement this in python.
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/6896/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/6896/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6895 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6895/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6895/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6895/events | https://github.com/NVIDIA/NeMo/issues/6895 | 1,766,891,158 | I_kwDOC_bI7s5pUJ6W | 6,895 | Finetuning ASR on another language (tokenizer) | {
"login": "gabitza-tech",
"id": 53811655,
"node_id": "MDQ6VXNlcjUzODExNjU1",
"avatar_url": "https://avatars.githubusercontent.com/u/53811655?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/gabitza-tech",
"html_url": "https://github.com/gabitza-tech",
"followers_url": "https://api.github.com/users/gabitza-tech/followers",
"following_url": "https://api.github.com/users/gabitza-tech/following{/other_user}",
"gists_url": "https://api.github.com/users/gabitza-tech/gists{/gist_id}",
"starred_url": "https://api.github.com/users/gabitza-tech/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/gabitza-tech/subscriptions",
"organizations_url": "https://api.github.com/users/gabitza-tech/orgs",
"repos_url": "https://api.github.com/users/gabitza-tech/repos",
"events_url": "https://api.github.com/users/gabitza-tech/events{/privacy}",
"received_events_url": "https://api.github.com/users/gabitza-tech/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2023-06-21T07:00:12 | 2023-07-20T09:30:09 | 2023-06-23T07:58:25 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | Hello,
I would like to finetune the stt_en_conformer_transducer_medium on 200 hours of another language (romanian). I am at the tokenization step and I am unsure of what would be the best practice:
1. Prepare the tokens using the text from the training audio transcripts (around 2 million words, not necessarily different). I would use sentencepiece, with unigram and vocab of 1024.
2. Prepare the tokens using a big text corpus of around 50 million words, one that is usually used for training LMs. Same configuration.
I am unsure if it is ok to train the tokenizer on more text than the training audios, I was not able to find if this is an OK practice and can lead to better/worse results. I was thinking that it might generalize better if it has more possible tokens. The tokenizer I obtained on just 200 hours looked rather limited, with lots of big words that are not actually that common outside of the training set. However, I am scared that it might not learn to pronounce some tokens based on the audio dataset, but I consider this rather unlikely.
One last question would be: What are the "##" , "_", "_de", etc. tokens representing?
Thank you in advance! | {
"login": "gabitza-tech",
"id": 53811655,
"node_id": "MDQ6VXNlcjUzODExNjU1",
"avatar_url": "https://avatars.githubusercontent.com/u/53811655?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/gabitza-tech",
"html_url": "https://github.com/gabitza-tech",
"followers_url": "https://api.github.com/users/gabitza-tech/followers",
"following_url": "https://api.github.com/users/gabitza-tech/following{/other_user}",
"gists_url": "https://api.github.com/users/gabitza-tech/gists{/gist_id}",
"starred_url": "https://api.github.com/users/gabitza-tech/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/gabitza-tech/subscriptions",
"organizations_url": "https://api.github.com/users/gabitza-tech/orgs",
"repos_url": "https://api.github.com/users/gabitza-tech/repos",
"events_url": "https://api.github.com/users/gabitza-tech/events{/privacy}",
"received_events_url": "https://api.github.com/users/gabitza-tech/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6895/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/6895/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6894 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6894/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6894/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6894/events | https://github.com/NVIDIA/NeMo/pull/6894 | 1,766,599,422 | PR_kwDOC_bI7s5TfxVJ | 6,894 | Enables manifest sharding for bucketed ASR datasets | {
"login": "bmwshop",
"id": 482416,
"node_id": "MDQ6VXNlcjQ4MjQxNg==",
"avatar_url": "https://avatars.githubusercontent.com/u/482416?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/bmwshop",
"html_url": "https://github.com/bmwshop",
"followers_url": "https://api.github.com/users/bmwshop/followers",
"following_url": "https://api.github.com/users/bmwshop/following{/other_user}",
"gists_url": "https://api.github.com/users/bmwshop/gists{/gist_id}",
"starred_url": "https://api.github.com/users/bmwshop/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/bmwshop/subscriptions",
"organizations_url": "https://api.github.com/users/bmwshop/orgs",
"repos_url": "https://api.github.com/users/bmwshop/repos",
"events_url": "https://api.github.com/users/bmwshop/events{/privacy}",
"received_events_url": "https://api.github.com/users/bmwshop/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 | 2023-06-21T03:14:43 | 2023-07-13T02:07:42 | 2023-07-13T02:07:42 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | true | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6894",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6894",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6894.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6894.patch",
"merged_at": null
} | # What does this PR do ?
Enables manifest sharding for bucketed ASR datasets
Add a one line overview of what this PR aims to accomplish.
**Collection**: 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": "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/6894/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/6894/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6893 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6893/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6893/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6893/events | https://github.com/NVIDIA/NeMo/pull/6893 | 1,766,505,667 | PR_kwDOC_bI7s5Tfdc7 | 6,893 | [TTS][ZH] add Chinese TTS recipes based on IPA symbol sets. | {
"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": 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 | 2023-06-21T01:18:50 | 2023-08-02T06:00:07 | 2023-08-02T06:00:03 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6893",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6893",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6893.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6893.patch",
"merged_at": "2023-08-02T06:00:03"
} | # What does this PR do ?
This PR provides 4 options for Chinese phoneme symbol sets, 24-final pinyin and its IPA conversion, 36-final pinyin and its IPA conversion. From my experiments, the IPA symbol set converted from 36-final pinyin produces the best sounds in terms of naturalness and prosody among these 4 setups. It's up to user's judge to pick either recipe.
* add new pinyin and ipa dictionaries with 36 finals.
* add yaml configs for 24-final pinyin and ipa.
* add copyright header
* add a directory level 24finals to discriminate from 36 finals.
```
$ pytest --nightly tests/collections/tts/models/
========================================= 27 passed, 3 warnings in 330.76s (0:05:30) ==========================================
```
**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**:
- [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": "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/6893/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/6893/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6892 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6892/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6892/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6892/events | https://github.com/NVIDIA/NeMo/pull/6892 | 1,766,368,095 | PR_kwDOC_bI7s5TfIka | 6,892 | add end-of-file-fixer hook | {
"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": 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
},
{
"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": ""
},
{
"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": ""
},
{
"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 | 2023-06-21T00:00:27 | 2024-07-30T17:49:12 | 2023-07-30T01:49:45 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6892",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6892",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6892.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6892.patch",
"merged_at": null
} | # What does this PR do ?
enforce this hook to ensure that a file is either empty, or ends with one newline. This feature can help avoiding review code changes relevant to newlines.
**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**:
- [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/6892/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/6892/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6891 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6891/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6891/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6891/events | https://github.com/NVIDIA/NeMo/pull/6891 | 1,766,364,388 | PR_kwDOC_bI7s5TfHzi | 6,891 | add a pre-commit hook to trim trailing whitespace. | {
"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": 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": 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
},
{
"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": ""
},
{
"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": ""
},
{
"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": ""
},
{
"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 | 2023-06-20T23:55:02 | 2024-07-30T17:49:15 | 2023-07-29T01:45:30 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6891",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6891",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6891.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6891.patch",
"merged_at": null
} | # What does this PR do ?
enforce this hook to avoid reviewing useless code changes relevant to trailing whitespaces.
**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**:
- [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/6891/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/6891/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6889 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6889/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6889/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6889/events | https://github.com/NVIDIA/NeMo/pull/6889 | 1,765,822,728 | PR_kwDOC_bI7s5Tddg_ | 6,889 | [TTS] Add output audio format to preprocessing | {
"login": "rlangman",
"id": 9942053,
"node_id": "MDQ6VXNlcjk5NDIwNTM=",
"avatar_url": "https://avatars.githubusercontent.com/u/9942053?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/rlangman",
"html_url": "https://github.com/rlangman",
"followers_url": "https://api.github.com/users/rlangman/followers",
"following_url": "https://api.github.com/users/rlangman/following{/other_user}",
"gists_url": "https://api.github.com/users/rlangman/gists{/gist_id}",
"starred_url": "https://api.github.com/users/rlangman/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/rlangman/subscriptions",
"organizations_url": "https://api.github.com/users/rlangman/orgs",
"repos_url": "https://api.github.com/users/rlangman/repos",
"events_url": "https://api.github.com/users/rlangman/events{/privacy}",
"received_events_url": "https://api.github.com/users/rlangman/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 | 2023-06-20T17:16:47 | 2023-08-02T06:05:06 | 2023-08-02T06:05:05 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6889",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6889",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6889.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6889.patch",
"merged_at": "2023-08-02T06:05:05"
} | # What does this PR do ?
Add option in TTS audio preprocessing to convert audio to desired output format. Default to flac since it is lossless and uses much less space than wav.
Adding this to save on disk space for large datasets, and because I was hitting cryptic dataloader errors when training with mp3 files.
**Collection**: [TTS]
# Changelog
- Add `output_format` CLI parameter.
# 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
| {
"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/6889/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/6889/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6888 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6888/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6888/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6888/events | https://github.com/NVIDIA/NeMo/pull/6888 | 1,765,395,486 | PR_kwDOC_bI7s5TcDsp | 6,888 | Update setup_os2s_decoders.py | {
"login": "mdvdv",
"id": 83948828,
"node_id": "MDQ6VXNlcjgzOTQ4ODI4",
"avatar_url": "https://avatars.githubusercontent.com/u/83948828?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/mdvdv",
"html_url": "https://github.com/mdvdv",
"followers_url": "https://api.github.com/users/mdvdv/followers",
"following_url": "https://api.github.com/users/mdvdv/following{/other_user}",
"gists_url": "https://api.github.com/users/mdvdv/gists{/gist_id}",
"starred_url": "https://api.github.com/users/mdvdv/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/mdvdv/subscriptions",
"organizations_url": "https://api.github.com/users/mdvdv/orgs",
"repos_url": "https://api.github.com/users/mdvdv/repos",
"events_url": "https://api.github.com/users/mdvdv/events{/privacy}",
"received_events_url": "https://api.github.com/users/mdvdv/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 | 2023-06-20T13:18:12 | 2023-07-07T08:20:48 | 2023-07-07T08:20:48 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6888",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6888",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6888.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6888.patch",
"merged_at": null
} | # What does this PR do ?
Install ctc_decoders without distutils duplication to avoid `AttributeError: module 'distutils' has no attribute 'ccompiler'`.
**Collection**: ASR
# Changelog
- Fix install_beamsearch_decoders.sh to work outside container
# Usage
```
# make NeMo venv
conda create --name nemo python==3.8.10
conda activate nemo
# install NeMo
sudo apt-get update && apt-get install -y libsndfile1 ffmpeg
git clone https://github.com/NVIDIA/NeMo
cd NeMo
./reinstall.sh
# install ctc-decoders
./scripts/asr_language_modeling/ngram_lm/install_beamsearch_decoders.sh $(pwd)
```
**PR Type**:
- [ x ] Bugfix
# Who can review?
@titu1994 | {
"login": "mdvdv",
"id": 83948828,
"node_id": "MDQ6VXNlcjgzOTQ4ODI4",
"avatar_url": "https://avatars.githubusercontent.com/u/83948828?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/mdvdv",
"html_url": "https://github.com/mdvdv",
"followers_url": "https://api.github.com/users/mdvdv/followers",
"following_url": "https://api.github.com/users/mdvdv/following{/other_user}",
"gists_url": "https://api.github.com/users/mdvdv/gists{/gist_id}",
"starred_url": "https://api.github.com/users/mdvdv/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/mdvdv/subscriptions",
"organizations_url": "https://api.github.com/users/mdvdv/orgs",
"repos_url": "https://api.github.com/users/mdvdv/repos",
"events_url": "https://api.github.com/users/mdvdv/events{/privacy}",
"received_events_url": "https://api.github.com/users/mdvdv/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6888/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/6888/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6887 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6887/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6887/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6887/events | https://github.com/NVIDIA/NeMo/pull/6887 | 1,765,337,005 | PR_kwDOC_bI7s5Tb253 | 6,887 | Fix k2 build topo helper | {
"login": "artbataev",
"id": 9380560,
"node_id": "MDQ6VXNlcjkzODA1NjA=",
"avatar_url": "https://avatars.githubusercontent.com/u/9380560?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/artbataev",
"html_url": "https://github.com/artbataev",
"followers_url": "https://api.github.com/users/artbataev/followers",
"following_url": "https://api.github.com/users/artbataev/following{/other_user}",
"gists_url": "https://api.github.com/users/artbataev/gists{/gist_id}",
"starred_url": "https://api.github.com/users/artbataev/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/artbataev/subscriptions",
"organizations_url": "https://api.github.com/users/artbataev/orgs",
"repos_url": "https://api.github.com/users/artbataev/repos",
"events_url": "https://api.github.com/users/artbataev/events{/privacy}",
"received_events_url": "https://api.github.com/users/artbataev/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 | 2023-06-20T12:48:36 | 2023-06-20T16:52:20 | 2023-06-20T16:52:19 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6887",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6887",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6887.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6887.patch",
"merged_at": "2023-06-20T16:52:19"
} | # What does this PR do ?
Fix `build_topo` function for `k2`-based topologies.
This solution fixes failing tests `tests/collections/asr/k2/test_ctc.py::TestCTCLossK2::test_case_small_blank_last` + `tests/collections/asr/k2/test_rnnt.py::TestRNNTLossK2::test_case_small_blank_last`
**Collection**: [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**:
- [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**:
- [ ] 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": "artbataev",
"id": 9380560,
"node_id": "MDQ6VXNlcjkzODA1NjA=",
"avatar_url": "https://avatars.githubusercontent.com/u/9380560?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/artbataev",
"html_url": "https://github.com/artbataev",
"followers_url": "https://api.github.com/users/artbataev/followers",
"following_url": "https://api.github.com/users/artbataev/following{/other_user}",
"gists_url": "https://api.github.com/users/artbataev/gists{/gist_id}",
"starred_url": "https://api.github.com/users/artbataev/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/artbataev/subscriptions",
"organizations_url": "https://api.github.com/users/artbataev/orgs",
"repos_url": "https://api.github.com/users/artbataev/repos",
"events_url": "https://api.github.com/users/artbataev/events{/privacy}",
"received_events_url": "https://api.github.com/users/artbataev/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6887/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/6887/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6886 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6886/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6886/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6886/events | https://github.com/NVIDIA/NeMo/pull/6886 | 1,764,370,843 | PR_kwDOC_bI7s5TYp6c | 6,886 | Update distopt API for coalesced NCCL calls | {
"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": 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": 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": 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 | 2023-06-19T23:47:19 | 2023-07-03T15:01:36 | 2023-07-03T15:01:35 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6886",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6886",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6886.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6886.patch",
"merged_at": "2023-07-03T15:01:35"
} | # What does this PR do ?
The Pytorch API for coalescing NCCL calls has changed in 2.1.0 (see https://github.com/pytorch/pytorch/pull/98793), which required some changes in the Apex wrapper functions (see https://github.com/NVIDIA/apex/pull/1663).
**Collection**: NLP
# Changelog
- Use updated API for coalescing NCCL calls in distributed optimizer
# Usage
Set the optimizer to `distributed_fused_adam` in the config 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)
- [ ] 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**:
- [ ] 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
* PyTorch API change in https://github.com/pytorch/pytorch/pull/98793
* Apex API change in https://github.com/NVIDIA/apex/pull/1663 | {
"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/6886/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/6886/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6885 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6885/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6885/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6885/events | https://github.com/NVIDIA/NeMo/pull/6885 | 1,764,345,232 | PR_kwDOC_bI7s5TYkSa | 6,885 | Start using ModelParallelConfig from Megatron Core | {
"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": 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": 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 | 2023-06-19T23:13:10 | 2023-08-18T12:26:59 | 2023-08-14T04:55:14 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6885",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6885",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6885.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6885.patch",
"merged_at": "2023-08-14T04:55:14"
} | # What does this PR do ?
This PR is adding the ModelParallelConfig arguments to be used with the next release of Megatron Core.
**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/6885/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/6885/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6884 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6884/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6884/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6884/events | https://github.com/NVIDIA/NeMo/pull/6884 | 1,764,252,810 | PR_kwDOC_bI7s5TYQvr | 6,884 | removed unnecessary print | {
"login": "dimapihtar",
"id": 37850217,
"node_id": "MDQ6VXNlcjM3ODUwMjE3",
"avatar_url": "https://avatars.githubusercontent.com/u/37850217?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/dimapihtar",
"html_url": "https://github.com/dimapihtar",
"followers_url": "https://api.github.com/users/dimapihtar/followers",
"following_url": "https://api.github.com/users/dimapihtar/following{/other_user}",
"gists_url": "https://api.github.com/users/dimapihtar/gists{/gist_id}",
"starred_url": "https://api.github.com/users/dimapihtar/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/dimapihtar/subscriptions",
"organizations_url": "https://api.github.com/users/dimapihtar/orgs",
"repos_url": "https://api.github.com/users/dimapihtar/repos",
"events_url": "https://api.github.com/users/dimapihtar/events{/privacy}",
"received_events_url": "https://api.github.com/users/dimapihtar/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 | 2023-06-19T21:33:09 | 2023-06-21T15:30:56 | 2023-06-21T15:30:55 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6884",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6884",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6884.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6884.patch",
"merged_at": "2023-06-21T15:30:55"
} | # What does this PR do ?
removes unnecessary log
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**:
- [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?
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": "dimapihtar",
"id": 37850217,
"node_id": "MDQ6VXNlcjM3ODUwMjE3",
"avatar_url": "https://avatars.githubusercontent.com/u/37850217?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/dimapihtar",
"html_url": "https://github.com/dimapihtar",
"followers_url": "https://api.github.com/users/dimapihtar/followers",
"following_url": "https://api.github.com/users/dimapihtar/following{/other_user}",
"gists_url": "https://api.github.com/users/dimapihtar/gists{/gist_id}",
"starred_url": "https://api.github.com/users/dimapihtar/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/dimapihtar/subscriptions",
"organizations_url": "https://api.github.com/users/dimapihtar/orgs",
"repos_url": "https://api.github.com/users/dimapihtar/repos",
"events_url": "https://api.github.com/users/dimapihtar/events{/privacy}",
"received_events_url": "https://api.github.com/users/dimapihtar/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6884/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/6884/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6883 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6883/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6883/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6883/events | https://github.com/NVIDIA/NeMo/pull/6883 | 1,762,489,039 | PR_kwDOC_bI7s5TSSCz | 6,883 | Update container for import action | {
"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": 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 | 2023-06-18T22:30:31 | 2023-06-20T22:24:25 | 2023-06-20T22:24:23 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6883",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6883",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6883.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6883.patch",
"merged_at": "2023-06-20T22:24:23"
} | # What does this PR do ?
Updates the base container for import action. Testing if this helps the NLP import test which is currently timing out due to solver issues with pip.
Edit: The issue was with installing and uninstalling nemo for each of the domains. This was wreaking havoc on the solver so we instead test each domain in parallel from a clean environment.
**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/6883/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/6883/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6882 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6882/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6882/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6882/events | https://github.com/NVIDIA/NeMo/pull/6882 | 1,761,956,538 | PR_kwDOC_bI7s5TQl9J | 6,882 | typo fix from #6666 | {
"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": ""
}
] | closed | false | null | [] | null | null | 2023-06-17T18:49:47 | 2023-06-19T05:46:04 | 2023-06-19T05:46:03 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6882",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6882",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6882.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6882.patch",
"merged_at": "2023-06-19T05:46:03"
} | # What does this PR do ?
fixes small error introduced in #6666
**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
- [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": "ekmb",
"id": 10428420,
"node_id": "MDQ6VXNlcjEwNDI4NDIw",
"avatar_url": "https://avatars.githubusercontent.com/u/10428420?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ekmb",
"html_url": "https://github.com/ekmb",
"followers_url": "https://api.github.com/users/ekmb/followers",
"following_url": "https://api.github.com/users/ekmb/following{/other_user}",
"gists_url": "https://api.github.com/users/ekmb/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ekmb/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ekmb/subscriptions",
"organizations_url": "https://api.github.com/users/ekmb/orgs",
"repos_url": "https://api.github.com/users/ekmb/repos",
"events_url": "https://api.github.com/users/ekmb/events{/privacy}",
"received_events_url": "https://api.github.com/users/ekmb/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6882/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/6882/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6881 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6881/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6881/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6881/events | https://github.com/NVIDIA/NeMo/issues/6881 | 1,761,953,365 | I_kwDOC_bI7s5pBUZV | 6,881 | inference.compute_logprob=True fails | {
"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": 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 | 2023-06-17T18:37:20 | 2023-06-17T18:52:48 | 2023-06-17T18:52:47 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
setting compute_logprob=Ture in the megatron_gpt_eval.py script is failing.
**Steps/Code to reproduce bug**
1. checkout main
2. Run ```python examples/nlp/language_modeling/megatron_gpt_eval.py gpt_model_file=path_to_nemo_gpt_model/megatron_gpt.nemo 'prompts=[How to fix GPU memory? A:]' tensor_model_parallel_size=1 inference.tokens_to_generate=32 trainer.precision=16 inference.compute_logprob=True```
This is the error:
```
Traceback (most recent call last):
File "examples/nlp/language_modeling/megatron_gpt_eval.py", line 273, in main
response = model.generate(
File "/home/adithyare/NeMo/nemo/collections/nlp/models/language_modeling/megatron_gpt_model.py", line 1100, in generate
return megatron_gpt_generate(self.cuda(), inputs, self.tokenizer, length_params, sampling_params)
File "/home/adithyare/NeMo/nemo/collections/nlp/modules/common/text_generation_utils.py", line 108, in megatron_gpt_generate
compute_attention_mask=sampling_params.get("compute_attention_mask", True) ** strategy_args,
TypeError: unsupported operand type(s) for ** or pow(): 'bool' and 'dict'
```
Please list *minimal* steps or code snippet for us to be able to reproduce the bug.
A helpful guide on on how to craft a minimal bug report http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports.
**Expected behavior**
A clear and concise description of what you expected to happen.
**Environment overview (please complete the following information)**
- Environment location: [Bare-metal, Docker, Cloud(specify cloud provider - AWS, Azure, GCP, Collab)]
- Method of NeMo install: [pip install or from source]. Please specify exact commands you used to install.
- If method of install is [Docker], provide `docker pull` & `docker run` commands used
**Environment details**
If NVIDIA docker image is used you don't need to specify these.
Otherwise, please provide:
- OS version
- PyTorch version
- Python version
**Additional context**
Add any other context about the problem here.
Example: GPU model
| {
"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/6881/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/6881/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6879 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6879/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6879/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6879/events | https://github.com/NVIDIA/NeMo/issues/6879 | 1,760,632,113 | I_kwDOC_bI7s5o8R0x | 6,879 | Fine-tuning ECAPA-TDNN | {
"login": "carlosfranzreb",
"id": 30505353,
"node_id": "MDQ6VXNlcjMwNTA1MzUz",
"avatar_url": "https://avatars.githubusercontent.com/u/30505353?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/carlosfranzreb",
"html_url": "https://github.com/carlosfranzreb",
"followers_url": "https://api.github.com/users/carlosfranzreb/followers",
"following_url": "https://api.github.com/users/carlosfranzreb/following{/other_user}",
"gists_url": "https://api.github.com/users/carlosfranzreb/gists{/gist_id}",
"starred_url": "https://api.github.com/users/carlosfranzreb/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/carlosfranzreb/subscriptions",
"organizations_url": "https://api.github.com/users/carlosfranzreb/orgs",
"repos_url": "https://api.github.com/users/carlosfranzreb/repos",
"events_url": "https://api.github.com/users/carlosfranzreb/events{/privacy}",
"received_events_url": "https://api.github.com/users/carlosfranzreb/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 | 2023-06-16T13:22:24 | 2023-09-09T01:43:45 | 2023-09-09T01:43:45 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | Hi, I am trying to fine-tune the ECAPA-TDNN model, following the recommendations made in the examples for fine-tuning the titanet-large. At some point during training, the loss is infinite and then everything turns into "not a number", including the embeddings. I already clip the gradients, as is done in the example that I follow. Could you check the demo below and tell me what's wrong? Thank you very much.
The packages that I install in the conda environment with python 3.9.16 are:
- torch==2.0.0
- torchvision==0.15.1
- torchaudio==2.0.1
- nemo_toolkit[all]==1.17.0
- resampy==0.4.2
- speechbrain==0.5.14
- openai-whisper==20230314
Here is a minimal dataset with 10 samples that I use to reproduce the error.
[ls-dev-clean-2.txt](https://github.com/NVIDIA/NeMo/files/11770634/ls-dev-clean-2.txt)
And the code:
```python
from omegaconf import OmegaConf
from nemo.collections.asr.models.label_models import EncDecSpeakerLabelModel
import pytorch_lightning as pl
from pytorch_lightning import seed_everything
from torchmetrics import Accuracy
from omegaconf import OmegaConf
seed_everything(42)
def finetune_nemo():
"""
Fine-tune the model on the given datafiles. Store the resulting model
in the given experiment folder. The decoder of the model is modified to
output the number of speakers given in `n_speakers`. Its parameters are
initialized again, meaning that the decoder is trained from scratch. Save
the model parameters in the experiment folder.
"""
datafiles = ["ls-dev-clean-2.txt"]
n_speakers = 3
device = "cpu"
batch_size = 2
num_workers = 0
cfg = OmegaConf.create(
{
"trainer": {
"accelerator": "cpu",
"max_epochs": 5,
"gradient_clip_val": 1.0,
"logger": False,
"deterministic": True,
"max_steps": -1,
},
"optim": {
"name": "adamw",
"lr": 0.0001,
"weight_decay": 0.0002,
"sched": {
"name": "CosineAnnealing",
"warmup_ratio": 0.1,
"min_lr": 0.0,
},
},
}
)
trainer = pl.Trainer(**cfg.trainer)
model = EncDecSpeakerLabelModel.from_pretrained(model_name="ecapa_tdnn")
# update training data config
OmegaConf.set_struct(model.cfg, False)
model.cfg.train_ds.update(
{
"manifest_filepath": datafiles,
"batch_size": batch_size,
"num_workers": num_workers,
"drop_last": True,
}
)
if "noise" in model.cfg.train_ds.augmentor:
del model.cfg.train_ds.augmentor.noise
model.setup_training_data(model.cfg.train_ds)
# initialize decoder with new number of classes
decoder_cfg = model.cfg.decoder
decoder_cfg.num_classes = n_speakers
model.decoder = EncDecSpeakerLabelModel.from_config_dict(decoder_cfg)
model.decoder.final.bias = None
# initialize optimizer and update accuracy metric
model._macro_accuracy = Accuracy(
task="multiclass", num_classes=n_speakers, average="macro"
)
model.setup_optimization(cfg.optim)
# train model
model.train()
trainer.fit(model)
model.eval()
model.to(device)
# save model
model.save_to("spkid.nemo")
if __name__ == "__main__":
finetune_nemo()
``` | {
"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/6879/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/6879/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6878 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6878/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6878/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6878/events | https://github.com/NVIDIA/NeMo/issues/6878 | 1,760,007,558 | I_kwDOC_bI7s5o55WG | 6,878 | Do Contextnet and Squeezeformer support streaming training | {
"login": "nullscc",
"id": 17700282,
"node_id": "MDQ6VXNlcjE3NzAwMjgy",
"avatar_url": "https://avatars.githubusercontent.com/u/17700282?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/nullscc",
"html_url": "https://github.com/nullscc",
"followers_url": "https://api.github.com/users/nullscc/followers",
"following_url": "https://api.github.com/users/nullscc/following{/other_user}",
"gists_url": "https://api.github.com/users/nullscc/gists{/gist_id}",
"starred_url": "https://api.github.com/users/nullscc/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/nullscc/subscriptions",
"organizations_url": "https://api.github.com/users/nullscc/orgs",
"repos_url": "https://api.github.com/users/nullscc/repos",
"events_url": "https://api.github.com/users/nullscc/events{/privacy}",
"received_events_url": "https://api.github.com/users/nullscc/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2023-06-16T06:30:50 | 2023-06-19T08:13:18 | 2023-06-19T08:13:18 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | So far, I know all ASR model but Contextnet and Squeezeformer in Nemo support streaming training.
So I want to ask : Do Contextnet and Squeezeformer support streaming training?
BTW, by streaming training, I mean chunk based like Cache-aware Streaming Conformer or model support streaming input itself during both training and inference stage, not the strategy just in inference stage like Buffered streaming.
Thanks in advance for anyone who give tips on this. | {
"login": "nullscc",
"id": 17700282,
"node_id": "MDQ6VXNlcjE3NzAwMjgy",
"avatar_url": "https://avatars.githubusercontent.com/u/17700282?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/nullscc",
"html_url": "https://github.com/nullscc",
"followers_url": "https://api.github.com/users/nullscc/followers",
"following_url": "https://api.github.com/users/nullscc/following{/other_user}",
"gists_url": "https://api.github.com/users/nullscc/gists{/gist_id}",
"starred_url": "https://api.github.com/users/nullscc/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/nullscc/subscriptions",
"organizations_url": "https://api.github.com/users/nullscc/orgs",
"repos_url": "https://api.github.com/users/nullscc/repos",
"events_url": "https://api.github.com/users/nullscc/events{/privacy}",
"received_events_url": "https://api.github.com/users/nullscc/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6878/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/6878/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6877 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6877/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6877/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6877/events | https://github.com/NVIDIA/NeMo/pull/6877 | 1,759,687,942 | PR_kwDOC_bI7s5TJEmx | 6,877 | Import Enum for chatbot component | {
"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 | 2023-06-15T23:57:54 | 2023-06-17T15:05:53 | 2023-06-17T15:05:52 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6877",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6877",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6877.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6877.patch",
"merged_at": "2023-06-17T15:05:52"
} | # 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/6877/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/6877/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6876 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6876/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6876/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6876/events | https://github.com/NVIDIA/NeMo/pull/6876 | 1,759,580,770 | PR_kwDOC_bI7s5TIt8L | 6,876 | Update mcore requirement to 0.2.0 | {
"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 | 2023-06-15T21:35:05 | 2023-06-18T21:49:04 | 2023-06-18T21:49:03 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6876",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6876",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6876.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6876.patch",
"merged_at": "2023-06-18T21:49:03"
} | # 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/6876/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/6876/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6875 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6875/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6875/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6875/events | https://github.com/NVIDIA/NeMo/pull/6875 | 1,759,510,604 | PR_kwDOC_bI7s5TIemq | 6,875 | Update mcore requirement to 0.2.0 | {
"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 | 2023-06-15T20:32:53 | 2023-06-15T21:34:32 | 2023-06-15T21:34:31 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6875",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6875",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6875.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6875.patch",
"merged_at": "2023-06-15T21:34:31"
} | # 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/6875/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/6875/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6873 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6873/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6873/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6873/events | https://github.com/NVIDIA/NeMo/issues/6873 | 1,758,499,445 | I_kwDOC_bI7s5o0JJ1 | 6,873 | transcribing with stt_en_fastconformer_ctc_large fails with oom on RAM (not gpu memory) on large manifest | {
"login": "bene-ges",
"id": 61418381,
"node_id": "MDQ6VXNlcjYxNDE4Mzgx",
"avatar_url": "https://avatars.githubusercontent.com/u/61418381?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/bene-ges",
"html_url": "https://github.com/bene-ges",
"followers_url": "https://api.github.com/users/bene-ges/followers",
"following_url": "https://api.github.com/users/bene-ges/following{/other_user}",
"gists_url": "https://api.github.com/users/bene-ges/gists{/gist_id}",
"starred_url": "https://api.github.com/users/bene-ges/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/bene-ges/subscriptions",
"organizations_url": "https://api.github.com/users/bene-ges/orgs",
"repos_url": "https://api.github.com/users/bene-ges/repos",
"events_url": "https://api.github.com/users/bene-ges/events{/privacy}",
"received_events_url": "https://api.github.com/users/bene-ges/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 | 2023-06-15T10:13:46 | 2023-07-28T01:44:34 | 2023-07-28T01:44:34 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
Transcribing with `stt_en_fastconformer_ctc_large` fails with oom on RAM (not gpu memory) on large manifest.
I run
```
python ${NEMO_PATH}/examples/asr/transcribe_speech.py \
pretrained_name="stt_en_fastconformer_ctc_large" \
dataset_manifest=${part}.json \
output_filename=pred_fast_ctc.$part.json \
batch_size=128
```
where input manifest contains 168'000+ records (small phrases).
This is run on Slurm where I have to specify the required amount of RAM for the process (this is common RAM, not gpu-memory).
If I specify 15G, the process fails on 46% of transcription, for 20G - same on 64%. So this is not connected with specific record in manifest - seems to be a cumulative memory consumption.
I tested other models on the same manifest (`stt_en_conformer_transducer_large`, `stt_en_conformer_ctc_large`, `stt_en_quartznet15x5`), all work fine.
Error message:
```
Detected 1 oom-kill event(s) in StepId=753012.0 cgroup. Some of your processes may have been killed by the cgroup out-of-memory handler.
srun: error: nodeXXX: task 0: Out Of Memory
```
**Steps/Code to reproduce bug**
Please list *minimal* steps or code snippet for us to be able to reproduce the bug.
A helpful guide on on how to craft a minimal bug report http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports.
**Expected behavior**
A clear and concise description of what you expected to happen.
**Environment overview (please complete the following information)**
- Environment location: [Bare-metal, Docker, Cloud(specify cloud provider - AWS, Azure, GCP, Collab)]
- Method of NeMo install: [pip install or from source]. Please specify exact commands you used to install.
- If method of install is [Docker], provide `docker pull` & `docker run` commands used
**Environment details**
OS version: CentOS Stream 8
PyTorch version: 1.13.0+cu117
Python version: 3.9
GPU model: Titan_X
| {
"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/6873/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/6873/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6872 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6872/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6872/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6872/events | https://github.com/NVIDIA/NeMo/pull/6872 | 1,757,767,033 | PR_kwDOC_bI7s5TCmQ8 | 6,872 | Apply garbage collection interval to validation steps | {
"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 | 2023-06-14T22:51:08 | 2023-06-28T21:07:32 | 2023-06-28T21:07:31 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6872",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6872",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6872.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6872.patch",
"merged_at": "2023-06-28T21:07:31"
} | # What does this PR do ?
Apply garbage collection interval to validation steps
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
Auto-applied when using `gc_interval > 0`
```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/6872/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/6872/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6871 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6871/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6871/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6871/events | https://github.com/NVIDIA/NeMo/issues/6871 | 1,757,679,792 | I_kwDOC_bI7s5oxBCw | 6,871 | Fine-tuning ASR (stt_en_conformer) on another language. | {
"login": "gabitza-tech",
"id": 53811655,
"node_id": "MDQ6VXNlcjUzODExNjU1",
"avatar_url": "https://avatars.githubusercontent.com/u/53811655?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/gabitza-tech",
"html_url": "https://github.com/gabitza-tech",
"followers_url": "https://api.github.com/users/gabitza-tech/followers",
"following_url": "https://api.github.com/users/gabitza-tech/following{/other_user}",
"gists_url": "https://api.github.com/users/gabitza-tech/gists{/gist_id}",
"starred_url": "https://api.github.com/users/gabitza-tech/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/gabitza-tech/subscriptions",
"organizations_url": "https://api.github.com/users/gabitza-tech/orgs",
"repos_url": "https://api.github.com/users/gabitza-tech/repos",
"events_url": "https://api.github.com/users/gabitza-tech/events{/privacy}",
"received_events_url": "https://api.github.com/users/gabitza-tech/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | null | 2023-06-14T21:31:35 | 2023-06-16T13:42:56 | 2023-06-16T13:42:55 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | Hello,
I would like to use the pre-trained stt_en_conformer_ctc or stt_en_conformer_transducer models for transfer learning on another language ( I only have around 1000 hours of audio, and I would like to jump-start the training/try to see if it may bring better results). I am new to ASR, so I am not very familiar if this is possible. From my understanding, I could use the pre-trained weights from the encoder, and only change the tokens of the decoder ( but the tokenizer should have the same size as the pretrained model, even if the tokens are different).
I would greatly appreciate some clarifications/corrections to what I mentioned above and if someone could guide me towards a tutorial/example that tackles something similar?
Thank you in advance! | {
"login": "gabitza-tech",
"id": 53811655,
"node_id": "MDQ6VXNlcjUzODExNjU1",
"avatar_url": "https://avatars.githubusercontent.com/u/53811655?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/gabitza-tech",
"html_url": "https://github.com/gabitza-tech",
"followers_url": "https://api.github.com/users/gabitza-tech/followers",
"following_url": "https://api.github.com/users/gabitza-tech/following{/other_user}",
"gists_url": "https://api.github.com/users/gabitza-tech/gists{/gist_id}",
"starred_url": "https://api.github.com/users/gabitza-tech/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/gabitza-tech/subscriptions",
"organizations_url": "https://api.github.com/users/gabitza-tech/orgs",
"repos_url": "https://api.github.com/users/gabitza-tech/repos",
"events_url": "https://api.github.com/users/gabitza-tech/events{/privacy}",
"received_events_url": "https://api.github.com/users/gabitza-tech/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6871/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/6871/timeline | null | completed | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6870 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6870/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6870/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6870/events | https://github.com/NVIDIA/NeMo/pull/6870 | 1,757,285,468 | PR_kwDOC_bI7s5TA8Mv | 6,870 | Apply garbage collection interval to validation steps | {
"login": "erhoo82",
"id": 15861259,
"node_id": "MDQ6VXNlcjE1ODYxMjU5",
"avatar_url": "https://avatars.githubusercontent.com/u/15861259?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/erhoo82",
"html_url": "https://github.com/erhoo82",
"followers_url": "https://api.github.com/users/erhoo82/followers",
"following_url": "https://api.github.com/users/erhoo82/following{/other_user}",
"gists_url": "https://api.github.com/users/erhoo82/gists{/gist_id}",
"starred_url": "https://api.github.com/users/erhoo82/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/erhoo82/subscriptions",
"organizations_url": "https://api.github.com/users/erhoo82/orgs",
"repos_url": "https://api.github.com/users/erhoo82/repos",
"events_url": "https://api.github.com/users/erhoo82/events{/privacy}",
"received_events_url": "https://api.github.com/users/erhoo82/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 | 2023-06-14T16:40:34 | 2023-06-14T22:50:38 | 2023-06-14T22:50:37 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6870",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6870",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6870.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6870.patch",
"merged_at": "2023-06-14T22:50:37"
} | # What does this PR do ?
Apply garbage collection interval to validation steps
# Changelog
- Add specific line by line info of high level changes in this PR.
# Usage
Auto-applied when using `gc_interval > 0`
```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/6870/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/6870/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6869 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6869/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6869/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6869/events | https://github.com/NVIDIA/NeMo/issues/6869 | 1,757,016,167 | I_kwDOC_bI7s5oufBn | 6,869 | Order of scheduler/optimizer calls | {
"login": "msis",
"id": 577139,
"node_id": "MDQ6VXNlcjU3NzEzOQ==",
"avatar_url": "https://avatars.githubusercontent.com/u/577139?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/msis",
"html_url": "https://github.com/msis",
"followers_url": "https://api.github.com/users/msis/followers",
"following_url": "https://api.github.com/users/msis/following{/other_user}",
"gists_url": "https://api.github.com/users/msis/gists{/gist_id}",
"starred_url": "https://api.github.com/users/msis/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/msis/subscriptions",
"organizations_url": "https://api.github.com/users/msis/orgs",
"repos_url": "https://api.github.com/users/msis/repos",
"events_url": "https://api.github.com/users/msis/events{/privacy}",
"received_events_url": "https://api.github.com/users/msis/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": "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
} | [
{
"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
}
] | null | null | 2023-06-14T14:16:42 | 2023-10-08T08:12:12 | 2023-07-23T01:57:16 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
PyTorch is warning about the order at which is the scheduler is being called.
**Steps/Code to reproduce bug**
Any training script will generate this warning:
```
[...]/lib/python3.10/site-packages/torch/optim/lr_scheduler.py:139: UserWarning: Detected call of `lr_scheduler.step()` before `optimizer.step()`. In PyTorch 1.1.0 and later, you should call them in the opposite order: `optimizer.step()` before `lr_scheduler.step()`. Failure to do this will result in PyTorch skipping the first value of the learning rate schedule. See more details at https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate
warnings.warn("Detected call of `lr_scheduler.step()` before `optimizer.step()`. "
```
**Expected behavior**
Reorder the calls.
**Environment overview (please complete the following information)**
- Environment location: Bare metal
- Method of NeMo install: `pip isntall nemo_toolkit[asr]`
**Environment details**
- NeMo: 1.18.1
- OS version: Ubuntu 22.04
- PyTorch version: '2.0.1+cu117'
- Python version: Python 3.10.6
**Additional context**
GPUs: A100 40G (8x)
| {
"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/6869/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/6869/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6866 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6866/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6866/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6866/events | https://github.com/NVIDIA/NeMo/pull/6866 | 1,756,075,858 | PR_kwDOC_bI7s5S82Pi | 6,866 | fix ptuning residuals bug | {
"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": ""
}
] | closed | false | null | [] | null | null | 2023-06-14T05:18:04 | 2023-06-22T16:22:29 | 2023-06-22T16:22:28 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6866",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6866",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6866.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6866.patch",
"merged_at": "2023-06-22T16:22:28"
} | # What does this PR do ?
- fixes peft ptuning bug which was causing ptuning to use a residual connection
- simplifies mixin use for all NLP PEFT models
- ptuning includes a inference only table which is needed for FT
-
**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
- [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/6866/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/6866/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6865 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6865/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6865/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6865/events | https://github.com/NVIDIA/NeMo/pull/6865 | 1,755,505,322 | PR_kwDOC_bI7s5S66CO | 6,865 | Update transcribe_utils.py | {
"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": 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 | 2023-06-13T19:15:42 | 2023-06-15T05:57:57 | 2023-06-15T05:57:56 | COLLABORATOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6865",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6865",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6865.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6865.patch",
"merged_at": "2023-06-15T05:57:56"
} | fix ctc decoding for hybrid model in partial transcribe
| {
"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
} | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/issues/6865/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/6865/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6864 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6864/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6864/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6864/events | https://github.com/NVIDIA/NeMo/pull/6864 | 1,755,353,754 | PR_kwDOC_bI7s5S6ZAq | 6,864 | Add API docs for NeMo Megatron | {
"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 | 2023-06-13T17:25:46 | 2023-06-14T22:18:02 | 2023-06-14T22:18:01 | CONTRIBUTOR | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6864",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6864",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6864.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6864.patch",
"merged_at": "2023-06-14T22:18:01"
} | # What does this PR do ?
Adds API documentation.
**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/6864/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/6864/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6863 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6863/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6863/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6863/events | https://github.com/NVIDIA/NeMo/pull/6863 | 1,755,053,015 | PR_kwDOC_bI7s5S5WiR | 6,863 | [Bugfix] Move nemo file config logic of GPT eval script | {
"login": "ZiyueXu77",
"id": 71786575,
"node_id": "MDQ6VXNlcjcxNzg2NTc1",
"avatar_url": "https://avatars.githubusercontent.com/u/71786575?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ZiyueXu77",
"html_url": "https://github.com/ZiyueXu77",
"followers_url": "https://api.github.com/users/ZiyueXu77/followers",
"following_url": "https://api.github.com/users/ZiyueXu77/following{/other_user}",
"gists_url": "https://api.github.com/users/ZiyueXu77/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ZiyueXu77/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ZiyueXu77/subscriptions",
"organizations_url": "https://api.github.com/users/ZiyueXu77/orgs",
"repos_url": "https://api.github.com/users/ZiyueXu77/repos",
"events_url": "https://api.github.com/users/ZiyueXu77/events{/privacy}",
"received_events_url": "https://api.github.com/users/ZiyueXu77/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 | 2023-06-13T14:40:32 | 2023-07-06T02:07:24 | 2023-07-06T02:07:24 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | false | {
"url": "https://api.github.com/repos/NVIDIA/NeMo/pulls/6863",
"html_url": "https://github.com/NVIDIA/NeMo/pull/6863",
"diff_url": "https://github.com/NVIDIA/NeMo/pull/6863.diff",
"patch_url": "https://github.com/NVIDIA/NeMo/pull/6863.patch",
"merged_at": null
} | # What does this PR do ?
Move the cfg.gpt_model_file logic into the "if" block to avoid error during ckpt eval
**Collection**: [Note which collection this PR will affect]
NLP
examples/nlp/language_modelling
# Changelog
- Add specific line by line info of high level changes in this PR.
Move https://github.com/NVIDIA/NeMo/blob/main/examples/nlp/language_modeling/megatron_gpt_eval.py#L172-L189
to after https://github.com/NVIDIA/NeMo/blob/main/examples/nlp/language_modeling/megatron_gpt_eval.py#L191
# Usage
* You can potentially add a usage example below
Otherwise, if following this
```python
python megatron_gpt_eval.py \
checkpoint_dir=PATH_TO_CHECKPOINT_FILE \
checkpoint_name=CHECKPOINT_FILE_NAME \
hparams_file=HPARAMS_FILE \
inference.greedy=True \
inference.add_BOS=True \
trainer.devices=1 \
trainer.num_nodes=1 \
tensor_model_parallel_size=-1 \
pipeline_model_parallel_size=-1 \
prompts=[prompt1,prompt2]
```
Since in this case gpt_model_file=null, the current script will error out due to restore path unknown
```
Traceback (most recent call last):
File "megatron_gpt_eval.py", line 176, in main
model_config = MegatronGPTModel.restore_from(
File "/workspace/NeMo-1.18.0/nemo/core/classes/modelPT.py", line 425, in restore_from
restore_path = os.path.abspath(os.path.expanduser(restore_path))
File "/usr/lib/python3.8/posixpath.py", line 231, in expanduser
path = os.fspath(path)
TypeError: expected str, bytes or os.PathLike object, not NoneType
```
# 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? No
- [ ] Did you add or update any necessary documentation? No
- [ ] Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc) No
- [ ] 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": "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/6863/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/6863/timeline | null | null | true |
https://api.github.com/repos/NVIDIA/NeMo/issues/6862 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6862/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6862/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6862/events | https://github.com/NVIDIA/NeMo/issues/6862 | 1,755,036,885 | I_kwDOC_bI7s5om7zV | 6,862 | Conformer-Transducer XXLarge Finetuning Error: size mismatch for encoder.layers.N.conv.depthwise_conv.weight | {
"login": "abarcovschi",
"id": 26311007,
"node_id": "MDQ6VXNlcjI2MzExMDA3",
"avatar_url": "https://avatars.githubusercontent.com/u/26311007?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/abarcovschi",
"html_url": "https://github.com/abarcovschi",
"followers_url": "https://api.github.com/users/abarcovschi/followers",
"following_url": "https://api.github.com/users/abarcovschi/following{/other_user}",
"gists_url": "https://api.github.com/users/abarcovschi/gists{/gist_id}",
"starred_url": "https://api.github.com/users/abarcovschi/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/abarcovschi/subscriptions",
"organizations_url": "https://api.github.com/users/abarcovschi/orgs",
"repos_url": "https://api.github.com/users/abarcovschi/repos",
"events_url": "https://api.github.com/users/abarcovschi/events{/privacy}",
"received_events_url": "https://api.github.com/users/abarcovschi/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": "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
}
] | null | null | 2023-06-13T14:32:03 | 2023-12-18T10:23:04 | 2023-07-23T01:57:18 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | **Describe the bug**
I am trying to finetune the XXL Conformer-Transducer model from a base .nemo checkpoint that I downloaded from https://catalog.ngc.nvidia.com/orgs/nvidia/teams/nemo/models/stt_en_conformer_transducer_xxlarge
**Steps/Code to reproduce bug**
I run the following command:
```
HYDRA_FULL_ERROR=1 python speech_to_text_rnnt_bpe.py --config-path=configs --config-name=conformer_transducer_bpe model.train_ds.batch_size=4 model.validation_ds.batch_size=4 model.test_ds.batch_size=4 trainer.devices=-1 trainer.accelerator=gpu +init_from_nemo_model=non-finetuned/models/stt_en_conformer_transducer_xxlarge.nemo ~model.encoder.stochastic_depth_drop_prob ~model.encoder.stochastic_depth_mode ~model.encoder.stochastic_depth_start_layer
```
I am using the same hyperparameters in my YAML config file as those recommended for the XLarge model, since there are no explicit recommendations for the XXLarge model.
My config YAML file (conformer_transducer_bpe.yaml) has the following content:
```
# It contains the default values for training a Conformer-Transducer ASR model, large size (~120M) with Transducer loss and sub-word encoding.
# Architecture and training config:
# Default learning parameters in this config are set for effective batch size of 2K. To train it with smaller effective
# batch sizes, you may need to re-tune the learning parameters or use higher accumulate_grad_batches.
# Here are the recommended configs for different variants of Conformer-Transducer, other parameters are the same as in this config file.
#
# +--------------+---------+---------+----------+------------------+--------------+--------------------------+-----------------+
# | Model | d_model | n_heads | n_layers | conv_kernel_size | weight_decay | pred_hidden/joint_hidden | pred_rnn_layers |
# +==============+=========+========+===========+==================+==============+==========================+=================+
# | Small (14M)| 176 | 4 | 16 | 31 | 0.0 | 320 | 1 |
# +--------------+---------+--------+-----------+------------------+--------------+--------------------------+-----------------+
# | Medium (32M)| 256 | 4 | 16 | 31 | 1e-3 | 640 | 1 |
# +--------------+---------+--------+-----------+------------------+--------------+--------------------------+-----------------+
# | Large (120M)| 512 | 8 | 17 | 31 | 1e-3 | 640 | 1 |
# +--------------+---------+--------+-----------+------------------+--------------+--------------------------+-----------------+
# | XLarge (644M)| 1024 | 8 | 24 | 5 | 1e-3 | 640 | 2 |
# +--------------+---------+--------+-----------+------------------+--------------+--------------------------+-----------------+
# You may find more info about Conformer-Transducer here: https://docs.nvidia.com/deeplearning/nemo/user-guide/docs/en/stable/asr/models.html#conformer-transducer
# Pre-trained models of Conformer-Transducer can be found here: https://docs.nvidia.com/deeplearning/nemo/user-guide/docs/en/stable/asr/results.html
# The checkpoint of the large model trained on NeMo ASRSET with this recipe can be found here: https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_conformer_transducer_large
name: "Conformer-Transducer-BPE-Large"
model:
sample_rate: 16000
compute_eval_loss: false # eval samples can be very long and exhaust memory. Disable computation of transducer loss during validation/testing with this flag.
log_prediction: true # enables logging sample predictions in the output during training
skip_nan_grad: false
model_defaults:
enc_hidden: ${model.encoder.d_model}
pred_hidden: 640
joint_hidden: 640
train_ds:
manifest_filepath: "/workspace/datasets/myst_w2v2_asr/train_manifest.json"
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
use_start_end_token: false
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: "/workspace/datasets/myst_w2v2_asr/test_manifest.json"
sample_rate: ${model.sample_rate}
batch_size: 16
shuffle: false
num_workers: 8
pin_memory: true
use_start_end_token: false
test_ds:
manifest_filepath: "/workspace/datasets/myst_w2v2_asr/test_manifest.json"
sample_rate: ${model.sample_rate}
batch_size: 16
shuffle: false
num_workers: 8
pin_memory: true
use_start_end_token: false
# You may find more detail on how to train a tokenizer at: /scripts/tokenizers/process_asr_text_tokenizer.py
tokenizer:
dir: "/workspace/projects/nemo_asr/tokenizers/myst_w2v2_asr/tokenizer_spe_unigram_v1024" # path to directory which contains either tokenizer.model (bpe) or vocab.txt (for wpe)
type: bpe # Can be either bpe (SentencePiece tokenizer) or wpe (WordPiece tokenizer)
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
frame_splicing: 1
dither: 0.00001
pad_to: 0
spec_augment:
_target_: nemo.collections.asr.modules.SpectrogramAugmentation
freq_masks: 2 # set to zero to disable it
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: 24
d_model: 1024
# Sub-sampling parameters
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
# Reduction parameters: Can be used to add another subsampling layer at a given position.
# Having a 2x reduction will speedup the training and inference speech while keeping similar WER.
# Adding it at the end will give the best WER while adding it at the beginning will give the best speedup.
reduction: null # pooling, striding, or null
reduction_position: null # Encoder block index or -1 for subsampling at the end of encoder
reduction_factor: 1
# 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: 5
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
# set to non-zero to enable stochastic depth
stochastic_depth_drop_prob: 0.0
stochastic_depth_mode: linear # linear or uniform
stochastic_depth_start_layer: 1
decoder:
_target_: nemo.collections.asr.modules.RNNTDecoder
normalization_mode: null # Currently only null is supported for export.
random_state_sampling: false # Random state sampling: https://arxiv.org/pdf/1910.11455.pdf
blank_as_pad: true # This flag must be set in order to support exporting of RNNT models + efficient inference.
prednet:
pred_hidden: ${model.model_defaults.pred_hidden}
pred_rnn_layers: 2
t_max: null
dropout: 0.2
joint:
_target_: nemo.collections.asr.modules.RNNTJoint
log_softmax: null # 'null' would set it automatically according to CPU/GPU device
preserve_memory: false # dramatically slows down training, but might preserve some memory
# Fuses the computation of prediction net + joint net + loss + WER calculation
# to be run on sub-batches of size `fused_batch_size`.
# When this flag is set to true, consider the `batch_size` of *_ds to be just `encoder` batch size.
# `fused_batch_size` is the actual batch size of the prediction net, joint net and transducer loss.
# Using small values here will preserve a lot of memory during training, but will make training slower as well.
# An optimal ratio of fused_batch_size : *_ds.batch_size is 1:1.
# However, to preserve memory, this ratio can be 1:8 or even 1:16.
# Extreme case of 1:B (i.e. fused_batch_size=1) should be avoided as training speed would be very slow.
fuse_loss_wer: true
fused_batch_size: 16
jointnet:
joint_hidden: ${model.model_defaults.joint_hidden}
activation: "relu"
dropout: 0.2
decoding:
strategy: "greedy_batch" # can be greedy, greedy_batch, beam, tsd, alsd.
# greedy strategy config
greedy:
max_symbols: 10
# beam strategy config
beam:
beam_size: 2
return_best_hypothesis: False
score_norm: true
tsd_max_sym_exp: 50 # for Time Synchronous Decoding
alsd_max_target_len: 2.0 # for Alignment-Length Synchronous Decoding
loss:
loss_name: "default"
warprnnt_numba_kwargs:
# FastEmit regularization: https://arxiv.org/abs/2010.11148
# You may enable FastEmit to reduce the latency of the model for streaming
fastemit_lambda: 0.0 # Recommended values to be in range [1e-4, 1e-2], 0.001 is a good start.
clamp: -1.0 # if > 0, applies gradient clamping in range [-clamp, clamp] for the joint tensor only.
# Adds Gaussian noise to the gradients of the decoder to avoid overfitting
variational_noise:
start_step: 0
std: 0.0
optim:
name: adamw
lr: 3.0
# optimizer arguments
betas: [0.9, 0.98]
weight_decay: 1e-3
# scheduler setup
sched:
name: NoamAnnealing
d_model: ${model.encoder.d_model}
# scheduler config override
warmup_steps: 40000
warmup_ratio: null
min_lr: 1e-6
trainer:
devices: -1 # number of GPUs, -1 would use all available GPUs
num_nodes: 1
max_epochs: 1000
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
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: null
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" # the objective function that determines model performance
mode: "min"
save_top_k: 2 # number of best performing .ckpt files to save (last epoch is always saved as a .ckpt file separately)
save_best_model: true # saves best model as nemo file instead of the default last checkpoint
always_save_nemo: True # saves the checkpoints as nemo files instead of PTL checkpoints (last checkpoint is saved by default)
resume_if_exists: false
resume_ignore_no_checkpoint: false
create_wandb_logger: true
wandb_logger_kwargs:
name: ${name}
project: "Conformer-Transducer-ASR"
```
**Output**
I get the following error:
```
Traceback (most recent call last):
File "speech_to_text_rnnt_bpe.py", line 133, in <module>
main() # noqa pylint: disable=no-value-for-parameter
File "/usr/local/lib/python3.8/dist-packages/nemo/core/config/hydra_runner.py", line 105, in wrapper
_run_hydra(
File "/usr/local/lib/python3.8/dist-packages/hydra/_internal/utils.py", line 389, in _run_hydra
_run_app(
File "/usr/local/lib/python3.8/dist-packages/hydra/_internal/utils.py", line 452, in _run_app
run_and_report(
File "/usr/local/lib/python3.8/dist-packages/hydra/_internal/utils.py", line 216, in run_and_report
raise ex
File "/usr/local/lib/python3.8/dist-packages/hydra/_internal/utils.py", line 213, in run_and_report
return func()
File "/usr/local/lib/python3.8/dist-packages/hydra/_internal/utils.py", line 453, in <lambda>
lambda: hydra.run(
File "/usr/local/lib/python3.8/dist-packages/hydra/_internal/hydra.py", line 132, in run
_ = ret.return_value
File "/usr/local/lib/python3.8/dist-packages/hydra/core/utils.py", line 260, in return_value
raise self._return_value
File "/usr/local/lib/python3.8/dist-packages/hydra/core/utils.py", line 186, in run_job
ret.return_value = task_function(task_cfg)
File "speech_to_text_rnnt_bpe_custom.py", line 114, in main
asr_model.maybe_init_from_pretrained_checkpoint(cfg)
File "/usr/local/lib/python3.8/dist-packages/lightning_utilities/core/rank_zero.py", line 27, in wrapped_fn
return fn(*args, **kwargs)
File "/usr/local/lib/python3.8/dist-packages/nemo/core/classes/modelPT.py", line 1220, in maybe_init_from_pretrained_checkpoint
self.load_state_dict(restored_model.state_dict(), strict=False)
File "/usr/local/lib/python3.8/dist-packages/torch/nn/modules/module.py", line 2001, in load_state_dict
raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
RuntimeError: Error(s) in loading state_dict for EncDecRNNTBPEModel:
size mismatch for encoder.layers.0.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.1.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.2.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.3.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.4.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.5.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.6.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.7.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.8.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.9.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.10.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.11.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.12.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.13.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.14.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.15.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.16.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.17.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.18.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.19.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.20.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.21.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.22.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
size mismatch for encoder.layers.23.conv.depthwise_conv.weight: copying a param with shape torch.Size([1024, 1, 31]) from checkpoint, the shape in current model is torch.Size([1024, 1, 5]).
```
**Expected behavior**
Should run finetuning successfully, which is the case when I tested with the Conformer-Transducer Small, Medium, Large, and XLarge .nemo models.
**Environment details**
Docker container created from image: nvcr.io/nvidia/nemo:23.02
**Additional context**
GPU model: quad A6000
| {
"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/6862/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/6862/timeline | null | not_planned | false |
https://api.github.com/repos/NVIDIA/NeMo/issues/6860 | https://api.github.com/repos/NVIDIA/NeMo | https://api.github.com/repos/NVIDIA/NeMo/issues/6860/labels{/name} | https://api.github.com/repos/NVIDIA/NeMo/issues/6860/comments | https://api.github.com/repos/NVIDIA/NeMo/issues/6860/events | https://github.com/NVIDIA/NeMo/issues/6860 | 1,754,012,828 | I_kwDOC_bI7s5ojByc | 6,860 | Fade on long TTS inputs | {
"login": "nharodig",
"id": 110415743,
"node_id": "U_kgDOBpTPfw",
"avatar_url": "https://avatars.githubusercontent.com/u/110415743?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/nharodig",
"html_url": "https://github.com/nharodig",
"followers_url": "https://api.github.com/users/nharodig/followers",
"following_url": "https://api.github.com/users/nharodig/following{/other_user}",
"gists_url": "https://api.github.com/users/nharodig/gists{/gist_id}",
"starred_url": "https://api.github.com/users/nharodig/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/nharodig/subscriptions",
"organizations_url": "https://api.github.com/users/nharodig/orgs",
"repos_url": "https://api.github.com/users/nharodig/repos",
"events_url": "https://api.github.com/users/nharodig/events{/privacy}",
"received_events_url": "https://api.github.com/users/nharodig/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 | 2023-06-13T04:29:10 | 2023-07-22T01:52:17 | 2023-07-22T01:52:17 | NONE | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | null | null | null | GPU T4
Riva 2.10
TTS es multispeaker fastpitch-hifigan
I increased the default input when using TTS SSML tags because I need to use 1000 characters, when using it in offline mode there is a big audio quality degradation or fade at the end of the generated audio, to avoid this I want to use online/streaming mode but the size of the chunks is very small (less than 1s), is it possible to increase the size of the chunks to 10s? Is there another way to solve this problem? I couldn't find Riva documentation on how to fix it.
This is the configuration that I'm using but in online mode it stays inferring and does not synthesize any audio:
riva-build speech_synthesis -f \
/servicemaker-dev/tts_rmir:tlt_encode \
/servicemaker-dev/tts_es_fastpitch.riva:None \
/servicemaker-dev/tts_es_hifigan_fastpitch.riva:None\
--sample_rate=44100 \
--preprocessor.g2p_ignore_ambiguous=true \
--phone_set="ipa" \
--voice_name=tts_spanish \
--wfst_tokenizer_model=/servicemaker-dev/tokenize_and_classify.far \
--wfst_verbalizer_model=/servicemaker-dev/verbalize.far \
--num_speakers=174 \
--preprocessor.max_sequence_length=1000 \
--chunk_length=350 \
--num_mels=350\
--num_samples_per_frame=1120 \
--chunkerFastPitch.max_sequence_idle_microseconds=30000 \
--chunkerFastPitch.max_queue_delay_microseconds=30000 \
--denoiser.fade_length=30000
| {
"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/6860/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/6860/timeline | null | not_planned | false |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.