url stringlengths 62 66 | repository_url stringclasses 1
value | labels_url stringlengths 76 80 | comments_url stringlengths 71 75 | events_url stringlengths 69 73 | html_url stringlengths 50 56 | id int64 377M 2.15B | node_id stringlengths 18 32 | number int64 1 29.2k | title stringlengths 1 487 | user dict | labels list | state stringclasses 2
values | locked bool 2
classes | assignee dict | assignees list | comments list | created_at int64 1.54k 1.71k | updated_at int64 1.54k 1.71k | closed_at int64 1.54k 1.71k ⌀ | author_association stringclasses 4
values | active_lock_reason stringclasses 2
values | body stringlengths 0 234k ⌀ | reactions dict | timeline_url stringlengths 71 75 | state_reason stringclasses 3
values | draft bool 2
classes | pull_request dict |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
https://api.github.com/repos/huggingface/transformers/issues/10230 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10230/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10230/comments | https://api.github.com/repos/huggingface/transformers/issues/10230/events | https://github.com/huggingface/transformers/pull/10230 | 810,036,646 | MDExOlB1bGxSZXF1ZXN0NTc0NzkyNjcx | 10,230 | Making TF GPT2 compliant with XLA and AMP | {
"login": "jplu",
"id": 959590,
"node_id": "MDQ6VXNlcjk1OTU5MA==",
"avatar_url": "https://avatars.githubusercontent.com/u/959590?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jplu",
"html_url": "https://github.com/jplu",
"followers_url": "https://api.github.com/users/jplu/followers",
... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # What does this PR do?
This PR makes the TF GPT2 model compliant with XLA and AMP. All the slow tests are passing as well. | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10230/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/huggingface/transformers/issues/10230/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10230",
"html_url": "https://github.com/huggingface/transformers/pull/10230",
"diff_url": "https://github.com/huggingface/transformers/pull/10230.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10230.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10229 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10229/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10229/comments | https://api.github.com/repos/huggingface/transformers/issues/10229/events | https://github.com/huggingface/transformers/pull/10229 | 810,020,839 | MDExOlB1bGxSZXF1ZXN0NTc0Nzc5NTUy | 10,229 | Introduce warmup_ratio training argument | {
"login": "tanmay17061",
"id": 32801726,
"node_id": "MDQ6VXNlcjMyODAxNzI2",
"avatar_url": "https://avatars.githubusercontent.com/u/32801726?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tanmay17061",
"html_url": "https://github.com/tanmay17061",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [
"As per current implementation, any non-zero value given for `warmup_steps` will override any effects of `warmup_ratio`. It made sense for me to give higher precedence to `warmup_steps` as it seems to be the more inconvenient argument of the 2 to provide from user perspective. Please let me know if this default beh... | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | Introduce warmup_ratio training argument in both
TrainingArguments and TFTrainingArguments classes (#6673)
# What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure i... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10229/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/huggingface/transformers/issues/10229/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10229",
"html_url": "https://github.com/huggingface/transformers/pull/10229",
"diff_url": "https://github.com/huggingface/transformers/pull/10229.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10229.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10228 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10228/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10228/comments | https://api.github.com/repos/huggingface/transformers/issues/10228/events | https://github.com/huggingface/transformers/issues/10228 | 810,013,350 | MDU6SXNzdWU4MTAwMTMzNTA= | 10,228 | Converting original T5 to be used in Transformers | {
"login": "marton-avrios",
"id": 59836119,
"node_id": "MDQ6VXNlcjU5ODM2MTE5",
"avatar_url": "https://avatars.githubusercontent.com/u/59836119?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/marton-avrios",
"html_url": "https://github.com/marton-avrios",
"followers_url": "https://api.githu... | [] | closed | false | null | [] | [
"Hi,\r\n\r\nthis file exists, it can be found here: https://github.com/huggingface/transformers/blob/master/src/transformers/models/t5/convert_t5_original_tf_checkpoint_to_pytorch.py\r\n\r\n",
"Thank you! I tried the script and it misses a `config.json` file. Where can I find this?",
"The config.json should be ... | 1,613 | 1,613 | 1,613 | NONE | null | I want to use original T5 checkpoint in Transformers library. I found multiple answers referring to `convert_t5_original_tf_checkpoint_to_pytorch.py` which does not seem to exist. Any other way? Or where can I find a (currently working) version of that file? | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10228/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/huggingface/transformers/issues/10228/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10227 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10227/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10227/comments | https://api.github.com/repos/huggingface/transformers/issues/10227/events | https://github.com/huggingface/transformers/issues/10227 | 809,961,223 | MDU6SXNzdWU4MDk5NjEyMjM= | 10,227 | Showing individual token and corresponding score during beam search | {
"login": "monmanuela",
"id": 35655790,
"node_id": "MDQ6VXNlcjM1NjU1Nzkw",
"avatar_url": "https://avatars.githubusercontent.com/u/35655790?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/monmanuela",
"html_url": "https://github.com/monmanuela",
"followers_url": "https://api.github.com/use... | [] | closed | false | null | [] | [
"Hey @monmanuela,\r\n\r\nThanks for checking out the post! We try to keep the repository for github issues and kindly ask you to post these kinds of questions on the [forum](https://discuss.huggingface.co/). Feel free to tag me there (@patrickvonplaten) :-)",
"@patrickvonplaten thanks for your quick reply! Posted... | 1,613 | 1,613 | 1,613 | NONE | null | ## Who can help
@patrickvonplaten
## Information
Hello,
I am using beam search with a pre-trained T5 model for summarization. I would like to visualize the beam search process by showing the tokens with the highest scores, and eventually the chosen beam like this diagram:
![image](https://user-images.github... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10227/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/huggingface/transformers/issues/10227/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10226 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10226/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10226/comments | https://api.github.com/repos/huggingface/transformers/issues/10226/events | https://github.com/huggingface/transformers/issues/10226 | 809,955,336 | MDU6SXNzdWU4MDk5NTUzMzY= | 10,226 | Trainer.train() is stuck | {
"login": "saichandrapandraju",
"id": 41769919,
"node_id": "MDQ6VXNlcjQxNzY5OTE5",
"avatar_url": "https://avatars.githubusercontent.com/u/41769919?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/saichandrapandraju",
"html_url": "https://github.com/saichandrapandraju",
"followers_url": "ht... | [] | closed | false | null | [] | [
"Hi,\r\n\r\nFor training-related issues, it might be better to ask your question on the [forum](https://discuss.huggingface.co/).\r\n\r\nThe authors of HuggingFace (and community members) are happy to help you there!\r\n"
] | 1,613 | 1,613 | 1,613 | NONE | null | Hi,
I'm training roberta-base using HF Trainer, but it's stuck at the starting itself. Here's my code -
```
train_dataset[0]
{'input_ids': tensor([ 0, 100, 657, ..., 1, 1, 1]),
'attention_mask': tensor([1, 1, 1, ..., 0, 0, 0]),
'labels': tensor(0)}
val_dataset[0]
{'input_ids': tensor([ 0, 110... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10226/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/huggingface/transformers/issues/10226/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10225 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10225/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10225/comments | https://api.github.com/repos/huggingface/transformers/issues/10225/events | https://github.com/huggingface/transformers/pull/10225 | 809,777,966 | MDExOlB1bGxSZXF1ZXN0NTc0NTgyMTE0 | 10,225 | [Trainer] memory tracker metrics | {
"login": "stas00",
"id": 10676103,
"node_id": "MDQ6VXNlcjEwNjc2MTAz",
"avatar_url": "https://avatars.githubusercontent.com/u/10676103?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stas00",
"html_url": "https://github.com/stas00",
"followers_url": "https://api.github.com/users/stas00/fo... | [] | closed | false | null | [] | [
"> Thanks for adding this functionality! One general comment I have is on the type of the `stage` argument. Since it has only four possible values from what I can see, it would be better to create an enum for those (to avoid typos and have auto-complete in an IDE).\r\n\r\nOh, let me make it absolutely automatic wit... | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | This PR introduced memory usage metrics in Trainer:
* [x] adds `TrainerMemoryTracker` (pytorch only, no-op for tf), which records deltas of the first gpu and cpu of the main process - and records them for `init|train|eval|test` stages - if there is no gpu it reports cpu only.
* [x] adds `--skip_memory_metrics` to d... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10225/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/huggingface/transformers/issues/10225/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10225",
"html_url": "https://github.com/huggingface/transformers/pull/10225",
"diff_url": "https://github.com/huggingface/transformers/pull/10225.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10225.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10224 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10224/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10224/comments | https://api.github.com/repos/huggingface/transformers/issues/10224/events | https://github.com/huggingface/transformers/issues/10224 | 809,726,134 | MDU6SXNzdWU4MDk3MjYxMzQ= | 10,224 | No module named 'tasks' | {
"login": "daniel-z-kaplan",
"id": 48258016,
"node_id": "MDQ6VXNlcjQ4MjU4MDE2",
"avatar_url": "https://avatars.githubusercontent.com/u/48258016?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/daniel-z-kaplan",
"html_url": "https://github.com/daniel-z-kaplan",
"followers_url": "https://api... | [] | closed | false | null | [] | [
"I think you did not clone the repository properly or are not running the command from the folder `examples/legacy/token-classification`, since that folder does have a task.py file.",
"Ah, I was searching for \"tasks.py\" instead. Just user error, thanks for the fast reply."
] | 1,613 | 1,613 | 1,613 | NONE | null | ## Environment info
- `transformers` version: 4.3.2
- Platform: 5.10.8-200.fc33.x86_64
- Python version: 3.8.3
- PyTorch version (GPU?): 1.6.0 (GPU)
- Tensorflow version (GPU?): 2.4.1 (GPU)
- Using GPU in script?: No.
- Using distributed or parallel set-up in script?: No.
@sgugger, @patil-suraj
## Inform... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10224/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/huggingface/transformers/issues/10224/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10223 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10223/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10223/comments | https://api.github.com/repos/huggingface/transformers/issues/10223/events | https://github.com/huggingface/transformers/issues/10223 | 809,714,942 | MDU6SXNzdWU4MDk3MTQ5NDI= | 10,223 | Slow Multi-GPU DDP training with run_clm.py and GPT2 | {
"login": "ktrapeznikov",
"id": 4052002,
"node_id": "MDQ6VXNlcjQwNTIwMDI=",
"avatar_url": "https://avatars.githubusercontent.com/u/4052002?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ktrapeznikov",
"html_url": "https://github.com/ktrapeznikov",
"followers_url": "https://api.github.com... | [] | closed | false | null | [] | [
"The answers to those two points is not necessarily yes.\r\n- DDP training is only faster if you have NVLinks between your GPUs, otherwise the slow communication between them can slow down training.\r\n- DDP training takes more space on GPU then a single-process training since there is some gradients caching.\r\n\r... | 1,613 | 1,619 | 1,619 | CONTRIBUTOR | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 4.3.0
- Platform: Linux-3.10.0-1160.11.1.el7.x86_64-x86_64-with-glibc2.10
- Python version: 3.8.6
- PyTorch ... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10223/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/huggingface/transformers/issues/10223/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10222 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10222/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10222/comments | https://api.github.com/repos/huggingface/transformers/issues/10222/events | https://github.com/huggingface/transformers/pull/10222 | 809,682,524 | MDExOlB1bGxSZXF1ZXN0NTc0NTAzODA3 | 10,222 | the change from single mask to multi mask support for pytorch | {
"login": "naveenjafer",
"id": 7025448,
"node_id": "MDQ6VXNlcjcwMjU0NDg=",
"avatar_url": "https://avatars.githubusercontent.com/u/7025448?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/naveenjafer",
"html_url": "https://github.com/naveenjafer",
"followers_url": "https://api.github.com/us... | [
{
"id": 2796628563,
"node_id": "MDU6TGFiZWwyNzk2NjI4NTYz",
"url": "https://api.github.com/repos/huggingface/transformers/labels/WIP",
"name": "WIP",
"color": "234C99",
"default": false,
"description": "Label your PR/Issue with WIP for some long outstanding Issues/PRs that are work in pro... | open | false | null | [] | [
"@Narsil @LysandreJik How do you suggest we go about with the targets param? At the moment, targets can either be a list of strings or a string. In case of multiple masks, there are 2 ways to go about with it.\r\n\r\n1. Provide a way for the user to define targets for each mask. \r\n2. One single target list th... | 1,613 | 1,649 | null | NONE | null | # What does this PR do?
A draft PR for the Feature request to change from single mask to multi mask support for the fill mask pipeline.
As discussed this is one a draft PR to discuss the changes that need to be made to the output format to jointly support multiple and single mask in one pipeline call. The PR impl... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10222/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/huggingface/transformers/issues/10222/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10222",
"html_url": "https://github.com/huggingface/transformers/pull/10222",
"diff_url": "https://github.com/huggingface/transformers/pull/10222.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10222.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10221 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10221/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10221/comments | https://api.github.com/repos/huggingface/transformers/issues/10221/events | https://github.com/huggingface/transformers/issues/10221 | 809,640,318 | MDU6SXNzdWU4MDk2NDAzMTg= | 10,221 | T5 relative attention bias: Discrepancy to original implementation | {
"login": "maurice-g",
"id": 2892585,
"node_id": "MDQ6VXNlcjI4OTI1ODU=",
"avatar_url": "https://avatars.githubusercontent.com/u/2892585?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/maurice-g",
"html_url": "https://github.com/maurice-g",
"followers_url": "https://api.github.com/users/ma... | [] | closed | false | null | [] | [
"The implementation of T5 is based on the original implementation which can be found [here](https://github.com/google-research/text-to-text-transfer-transformer).\r\n\r\nThe implementation you are referring to is a general Transformer implementation from Tensorflow Mesh. It seems like this repo does not implement T... | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | ### Who can help
@patrickvonplaten
## Information
Model I am using: T5
In the huggingface TF T5 implementation, the relative attention bias only seems to be applied to the first layer of the stack. If I understand the original implementation correctly, though, it is applied to all layers there.
HF: https:... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10221/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/huggingface/transformers/issues/10221/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10220 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10220/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10220/comments | https://api.github.com/repos/huggingface/transformers/issues/10220/events | https://github.com/huggingface/transformers/pull/10220 | 809,615,232 | MDExOlB1bGxSZXF1ZXN0NTc0NDQ3NjU1 | 10,220 | fix deprecated reference `tokenizer.max_len` in glue.py | {
"login": "poedator",
"id": 24738311,
"node_id": "MDQ6VXNlcjI0NzM4MzEx",
"avatar_url": "https://avatars.githubusercontent.com/u/24738311?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/poedator",
"html_url": "https://github.com/poedator",
"followers_url": "https://api.github.com/users/poe... | [] | closed | false | null | [] | [] | 1,613 | 1,614 | 1,614 | CONTRIBUTOR | null | This is to fix deprecated reference to `tokenizer.max_len` with `tokenizer.model_max_length` - similar to [issue 8739](https://github.com/huggingface/transformers/issues/8739) and [PR 8604](https://github.com/huggingface/transformers/pull/8604).
See error example [in Colab here](https://colab.research.google.com/gist... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10220/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/huggingface/transformers/issues/10220/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10220",
"html_url": "https://github.com/huggingface/transformers/pull/10220",
"diff_url": "https://github.com/huggingface/transformers/pull/10220.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10220.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10219 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10219/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10219/comments | https://api.github.com/repos/huggingface/transformers/issues/10219/events | https://github.com/huggingface/transformers/pull/10219 | 809,571,923 | MDExOlB1bGxSZXF1ZXN0NTc0NDExOTY5 | 10,219 | [trainer] fix ignored columns logger | {
"login": "stas00",
"id": 10676103,
"node_id": "MDQ6VXNlcjEwNjc2MTAz",
"avatar_url": "https://avatars.githubusercontent.com/u/10676103?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stas00",
"html_url": "https://github.com/stas00",
"followers_url": "https://api.github.com/users/stas00/fo... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | This PR fixes a confusing log entry that says:
```
The following columns in the evaluation set don't have a corresponding argument in `T5ForConditionalGeneration.forward` and have been ignored: .
```
when everything is in order.
@sgugger | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10219/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/huggingface/transformers/issues/10219/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10219",
"html_url": "https://github.com/huggingface/transformers/pull/10219",
"diff_url": "https://github.com/huggingface/transformers/pull/10219.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10219.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10218 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10218/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10218/comments | https://api.github.com/repos/huggingface/transformers/issues/10218/events | https://github.com/huggingface/transformers/issues/10218 | 809,555,024 | MDU6SXNzdWU4MDk1NTUwMjQ= | 10,218 | discrepancy between the Huggingface T5Tokenizer and the original T5tokenizer | {
"login": "dorost1234",
"id": 79165106,
"node_id": "MDQ6VXNlcjc5MTY1MTA2",
"avatar_url": "https://avatars.githubusercontent.com/u/79165106?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/dorost1234",
"html_url": "https://github.com/dorost1234",
"followers_url": "https://api.github.com/use... | [] | closed | false | null | [] | [
"Hi, \r\n\r\nT5 is an encoder-decoder Transformer. The `run_mlm.py` script can only be used for encoder-only models, such as BERT, RoBERTa, DeBERTa, etc. \r\n\r\nBesides this, T5 does not use the regular [MASK] token as BERT. Rather than masked language modeling, T5 is pre-trained on \"unsupervised denoising traini... | 1,613 | 1,624 | 1,619 | NONE | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `Transformers` version: 4.3.2
- Platform: Linux
- Python version: 3.7
- PyTorch version (GPU?): yes
- Tensorflow version (GPU?): -... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10218/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/huggingface/transformers/issues/10218/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10217 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10217/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10217/comments | https://api.github.com/repos/huggingface/transformers/issues/10217/events | https://github.com/huggingface/transformers/pull/10217 | 809,495,804 | MDExOlB1bGxSZXF1ZXN0NTc0MzQ5MTI1 | 10,217 | Fix add_token_positions in custom datasets tutorial | {
"login": "joeddav",
"id": 9353833,
"node_id": "MDQ6VXNlcjkzNTM4MzM=",
"avatar_url": "https://avatars.githubusercontent.com/u/9353833?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/joeddav",
"html_url": "https://github.com/joeddav",
"followers_url": "https://api.github.com/users/joeddav/... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | Discussed in #10210. The example `add_token_positions` function incorrectly converts `answers[i]['answer_end']` to its corresponding tokenized index rather than `answers[i]['answer_end'] - 1`. | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10217/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/huggingface/transformers/issues/10217/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10217",
"html_url": "https://github.com/huggingface/transformers/pull/10217",
"diff_url": "https://github.com/huggingface/transformers/pull/10217.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10217.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10216 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10216/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10216/comments | https://api.github.com/repos/huggingface/transformers/issues/10216/events | https://github.com/huggingface/transformers/pull/10216 | 809,483,522 | MDExOlB1bGxSZXF1ZXN0NTc0MzM4NTQ3 | 10,216 | Making TF Funnel compliant with AMP | {
"login": "jplu",
"id": 959590,
"node_id": "MDQ6VXNlcjk1OTU5MA==",
"avatar_url": "https://avatars.githubusercontent.com/u/959590?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jplu",
"html_url": "https://github.com/jplu",
"followers_url": "https://api.github.com/users/jplu/followers",
... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # What does this PR do?
This PR makes the TF Funnel model compliant with AMP. All the slow tests are passing as well.
| {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10216/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/huggingface/transformers/issues/10216/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10216",
"html_url": "https://github.com/huggingface/transformers/pull/10216",
"diff_url": "https://github.com/huggingface/transformers/pull/10216.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10216.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10215 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10215/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10215/comments | https://api.github.com/repos/huggingface/transformers/issues/10215/events | https://github.com/huggingface/transformers/pull/10215 | 809,463,604 | MDExOlB1bGxSZXF1ZXN0NTc0MzIxNjMw | 10,215 | Factor out methods | {
"login": "LysandreJik",
"id": 30755778,
"node_id": "MDQ6VXNlcjMwNzU1Nzc4",
"avatar_url": "https://avatars.githubusercontent.com/u/30755778?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/LysandreJik",
"html_url": "https://github.com/LysandreJik",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [
"@sgugger @patrickvonplaten verified this fixed the issue in #10214"
] | 1,613 | 1,613 | 1,613 | MEMBER | null | With PyTorch's DataParallel, it is not possible to simply iterate over parameters in order to find the `nn.Module`'s dtype or device.
Some efforts were made to catch the error (`StopIteration`) in most cases, but the some were forgotten. This PR factors the try/except in a method, which is applied everywhere instead... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10215/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/huggingface/transformers/issues/10215/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10215",
"html_url": "https://github.com/huggingface/transformers/pull/10215",
"diff_url": "https://github.com/huggingface/transformers/pull/10215.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10215.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10214 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10214/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10214/comments | https://api.github.com/repos/huggingface/transformers/issues/10214/events | https://github.com/huggingface/transformers/issues/10214 | 809,443,349 | MDU6SXNzdWU4MDk0NDMzNDk= | 10,214 | StopIteration Error when running beam search for squad 2.0 | {
"login": "neufang",
"id": 689203,
"node_id": "MDQ6VXNlcjY4OTIwMw==",
"avatar_url": "https://avatars.githubusercontent.com/u/689203?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/neufang",
"html_url": "https://github.com/neufang",
"followers_url": "https://api.github.com/users/neufang/fo... | [] | closed | false | null | [] | [
"Indeed! I can reproduce, will fix.",
"Can you tell me if https://github.com/huggingface/transformers/pull/10215 fixes it? You can try by installing the following:\r\n\r\n```\r\npip install git+https://github.com/huggingface/transformers@parameter-device-dtype\r\n```",
"Thanks a lot for the quick fix. I'm runni... | 1,613 | 1,613 | 1,613 | NONE | null | I'm using `huggingface/transformers-pytorch-gpu:4.3.0` on Ubuntu DGX1 server with 8 V100 GPUs.
`NVIDIA-SMI 418.126.02 Driver Version: 418.126.02 CUDA Version: 10.1`
When running the step in `examples/question_answering/README.md` for beam search for squad 2.0
```
python run_qa_beam_search.py \
--model... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10214/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/huggingface/transformers/issues/10214/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10213 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10213/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10213/comments | https://api.github.com/repos/huggingface/transformers/issues/10213/events | https://github.com/huggingface/transformers/pull/10213 | 809,431,838 | MDExOlB1bGxSZXF1ZXN0NTc0Mjk1MTIz | 10,213 | Store FLOS as floats to avoid overflow. | {
"login": "sgugger",
"id": 35901082,
"node_id": "MDQ6VXNlcjM1OTAxMDgy",
"avatar_url": "https://avatars.githubusercontent.com/u/35901082?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sgugger",
"html_url": "https://github.com/sgugger",
"followers_url": "https://api.github.com/users/sgugge... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | COLLABORATOR | null | # What does this PR do?
As pointed out in #10212, the `total_flos` stored as ints can result in overflowing errors: when in Python ints there is no risk, but when in distributed training, we use torch.int64 to gather all FLOS on all processes which can trigger that error.
Fixes #10212 | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10213/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/huggingface/transformers/issues/10213/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10213",
"html_url": "https://github.com/huggingface/transformers/pull/10213",
"diff_url": "https://github.com/huggingface/transformers/pull/10213.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10213.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10212 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10212/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10212/comments | https://api.github.com/repos/huggingface/transformers/issues/10212/events | https://github.com/huggingface/transformers/issues/10212 | 809,350,471 | MDU6SXNzdWU4MDkzNTA0NzE= | 10,212 | RuntimeError: Overflow when unpacking long | {
"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.gi... | [] | closed | false | null | [] | [
"This comes from the `_total_flos` being stored as long and overflowing in a big training. Will fix this by storing them as floats (hoping for a PR by the end of today)."
] | 1,613 | 1,613 | 1,613 | NONE | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 4.4.0.dev0
- Platform: linux
- Python version: 3.6.10
- PyTorch version (GPU?): 1.7.0a0. (gpu)
- Using ... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10212/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/huggingface/transformers/issues/10212/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10211 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10211/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10211/comments | https://api.github.com/repos/huggingface/transformers/issues/10211/events | https://github.com/huggingface/transformers/pull/10211 | 809,330,723 | MDExOlB1bGxSZXF1ZXN0NTc0MjExMzIx | 10,211 | Making TF XLM-like models XLA and AMP compliant | {
"login": "jplu",
"id": 959590,
"node_id": "MDQ6VXNlcjk1OTU5MA==",
"avatar_url": "https://avatars.githubusercontent.com/u/959590?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jplu",
"html_url": "https://github.com/jplu",
"followers_url": "https://api.github.com/users/jplu/followers",
... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # What does this PR do?
This PR makes the TF XLM-like models compliant with XLA and AMP. All the slow tests are passing as well for these models. | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10211/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/huggingface/transformers/issues/10211/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10211",
"html_url": "https://github.com/huggingface/transformers/pull/10211",
"diff_url": "https://github.com/huggingface/transformers/pull/10211.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10211.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10210 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10210/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10210/comments | https://api.github.com/repos/huggingface/transformers/issues/10210/events | https://github.com/huggingface/transformers/issues/10210 | 809,300,209 | MDU6SXNzdWU4MDkzMDAyMDk= | 10,210 | QA Documentation: I got error just copy and pasting documentation | {
"login": "andreabac3",
"id": 36055796,
"node_id": "MDQ6VXNlcjM2MDU1Nzk2",
"avatar_url": "https://avatars.githubusercontent.com/u/36055796?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/andreabac3",
"html_url": "https://github.com/andreabac3",
"followers_url": "https://api.github.com/use... | [] | closed | false | null | [] | [
"Pinging @joeddav on this one, since he wrote this tutorial :-)",
"Thank you @sgugger for the reply.\r\nOk I can wait for the answer from @joeddav.\r\n\r\nHave a nice day. ",
"Figured it out. `answer_end` is the character position immediately _after_ the answer, so end_position should be derived from `answer_en... | 1,613 | 1,614 | 1,613 | CONTRIBUTOR | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 4.3.1
- Platform:Manjaro Linux
- Python version: 1.5.1
- PyTorch version (GPU?): Yes
- Using GPU in script?... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10210/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/huggingface/transformers/issues/10210/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10209 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10209/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10209/comments | https://api.github.com/repos/huggingface/transformers/issues/10209/events | https://github.com/huggingface/transformers/pull/10209 | 809,247,683 | MDExOlB1bGxSZXF1ZXN0NTc0MTQxNjA5 | 10,209 | Make TF CTRL compliant with XLA and AMP | {
"login": "jplu",
"id": 959590,
"node_id": "MDQ6VXNlcjk1OTU5MA==",
"avatar_url": "https://avatars.githubusercontent.com/u/959590?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jplu",
"html_url": "https://github.com/jplu",
"followers_url": "https://api.github.com/users/jplu/followers",
... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # What does this PR do?
This PR makes the TF CTRL model compliant with XLA and AMP. All the slow tests are passing as well.
| {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10209/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/huggingface/transformers/issues/10209/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10209",
"html_url": "https://github.com/huggingface/transformers/pull/10209",
"diff_url": "https://github.com/huggingface/transformers/pull/10209.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10209.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10208 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10208/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10208/comments | https://api.github.com/repos/huggingface/transformers/issues/10208/events | https://github.com/huggingface/transformers/issues/10208 | 809,236,920 | MDU6SXNzdWU4MDkyMzY5MjA= | 10,208 | different behavior for get_input_embeddings() between 4.2.x and 4.3.x in Tensorflow | {
"login": "AndyTheFactory",
"id": 863810,
"node_id": "MDQ6VXNlcjg2MzgxMA==",
"avatar_url": "https://avatars.githubusercontent.com/u/863810?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/AndyTheFactory",
"html_url": "https://github.com/AndyTheFactory",
"followers_url": "https://api.github... | [] | closed | false | null | [] | [
"Hello!\r\n\r\nThis is because in the 4.3 version the implementation of the embedding have changed and `get_input_embeddings()` returns only the word embeddings layer, hence only `input_ids` can be passed.\r\n\r\nThis was an unexpected behavior and will be fixed for the next release (the fix is already in master if... | 1,613 | 1,613 | 1,613 | NONE | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 4.2.x vs 4.3.x
- Platform: Colab
- Python version: 3.6
- Tensorflow version (GPU?): 2.4.1
@jplu
## In... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10208/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/huggingface/transformers/issues/10208/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10207 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10207/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10207/comments | https://api.github.com/repos/huggingface/transformers/issues/10207/events | https://github.com/huggingface/transformers/pull/10207 | 809,214,852 | MDExOlB1bGxSZXF1ZXN0NTc0MTEzOTA2 | 10,207 | Unlock XLA test for TF ConvBert | {
"login": "jplu",
"id": 959590,
"node_id": "MDQ6VXNlcjk1OTU5MA==",
"avatar_url": "https://avatars.githubusercontent.com/u/959590?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jplu",
"html_url": "https://github.com/jplu",
"followers_url": "https://api.github.com/users/jplu/followers",
... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # What does this PR do?
This PR allows the XLA test for TF ConvBert.
| {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10207/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/huggingface/transformers/issues/10207/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10207",
"html_url": "https://github.com/huggingface/transformers/pull/10207",
"diff_url": "https://github.com/huggingface/transformers/pull/10207.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10207.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10206 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10206/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10206/comments | https://api.github.com/repos/huggingface/transformers/issues/10206/events | https://github.com/huggingface/transformers/issues/10206 | 809,028,481 | MDU6SXNzdWU4MDkwMjg0ODE= | 10,206 | Tokenizer is working different from expected functionality. | {
"login": "pyturn",
"id": 25935364,
"node_id": "MDQ6VXNlcjI1OTM1MzY0",
"avatar_url": "https://avatars.githubusercontent.com/u/25935364?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/pyturn",
"html_url": "https://github.com/pyturn",
"followers_url": "https://api.github.com/users/pyturn/fo... | [] | closed | false | {
"login": "LysandreJik",
"id": 30755778,
"node_id": "MDQ6VXNlcjMwNzU1Nzc4",
"avatar_url": "https://avatars.githubusercontent.com/u/30755778?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/LysandreJik",
"html_url": "https://github.com/LysandreJik",
"followers_url": "https://api.github.com/... | [
{
"login": "LysandreJik",
"id": 30755778,
"node_id": "MDQ6VXNlcjMwNzU1Nzc4",
"avatar_url": "https://avatars.githubusercontent.com/u/30755778?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/LysandreJik",
"html_url": "https://github.com/LysandreJik",
"followers_url": "htt... | [
"Hi! Could you provide a reproducible example? I don't understand what's your `TOKENIZER` here. Thanks",
"Hi @LysandreJik , \r\n\r\nThanks for your response , I have done deep-dive and modified code a bit to reproduce the same issue. I am extending the vocabulary of tokenizer and using some automated logic to add... | 1,613 | 1,619 | 1,619 | NONE | null | Hi,
I updated the vocabulary for pre-trained tokenizer. Pretrained tokenizer was taken from this model - https://huggingface.co/microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract
When I am using the updated tokenizer , it is creating the sub tokens of the word which are present in vocabulary dictionary. And th... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10206/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/huggingface/transformers/issues/10206/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10205 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10205/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10205/comments | https://api.github.com/repos/huggingface/transformers/issues/10205/events | https://github.com/huggingface/transformers/pull/10205 | 809,019,053 | MDExOlB1bGxSZXF1ZXN0NTczOTU0MTUw | 10,205 | set tgt_lang of MBart Tokenizer for summarization | {
"login": "HeroadZ",
"id": 17962682,
"node_id": "MDQ6VXNlcjE3OTYyNjgy",
"avatar_url": "https://avatars.githubusercontent.com/u/17962682?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/HeroadZ",
"html_url": "https://github.com/HeroadZ",
"followers_url": "https://api.github.com/users/Heroad... | [] | closed | false | null | [] | [
"Thanks a lot for fixing this!",
"Other models/tasks have this issue as well https://github.com/huggingface/transformers/issues/10292\r\n\r\nThese features require tests. Without tests this is an endless work-work-work - otherwise we keep on breaking what was working before.\r\n"
] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # What does this PR do?
To set tgt_lang of MBart Tokenizer for summarization.
Otherwise, the error `AttributeError: 'MBartTokenizerFast' object has no attribute 'tgt_lang'` occurred.
I have read your discussion and know that you will modify the part of MBart later. So this PR will be meaningless at that time.
B... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10205/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/huggingface/transformers/issues/10205/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10205",
"html_url": "https://github.com/huggingface/transformers/pull/10205",
"diff_url": "https://github.com/huggingface/transformers/pull/10205.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10205.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10204 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10204/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10204/comments | https://api.github.com/repos/huggingface/transformers/issues/10204/events | https://github.com/huggingface/transformers/issues/10204 | 809,007,443 | MDU6SXNzdWU4MDkwMDc0NDM= | 10,204 | 1.3GB dataset creates over 107GB of cache file! | {
"login": "DarshanDeshpande",
"id": 39432636,
"node_id": "MDQ6VXNlcjM5NDMyNjM2",
"avatar_url": "https://avatars.githubusercontent.com/u/39432636?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/DarshanDeshpande",
"html_url": "https://github.com/DarshanDeshpande",
"followers_url": "https://... | [] | closed | false | null | [] | [
"cc @lhoestq ",
"Related to https://github.com/huggingface/datasets/issues/861\r\nMaybe on-the-fly tokenization can help.\r\nOr if we stick to having the tokenization in the preprocessing, at least reduce the precision of the integers stored on disk and maybe do the padding on the fly.",
"@lhoestq Are there any... | 1,613 | 1,615 | 1,613 | NONE | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 4.4.0 dev0
- Platform: Google Colab
- Python version: 3.6
- PyTorch version (GPU?): 1.7
- Tensorflow versio... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10204/reactions",
"total_count": 3,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 3
} | https://api.github.com/repos/huggingface/transformers/issues/10204/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10203 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10203/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10203/comments | https://api.github.com/repos/huggingface/transformers/issues/10203/events | https://github.com/huggingface/transformers/pull/10203 | 809,002,194 | MDExOlB1bGxSZXF1ZXN0NTczOTQwMTUy | 10,203 | [run_glue] Add MNLI compatible mode | {
"login": "JetRunner",
"id": 22514219,
"node_id": "MDQ6VXNlcjIyNTE0MjE5",
"avatar_url": "https://avatars.githubusercontent.com/u/22514219?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/JetRunner",
"html_url": "https://github.com/JetRunner",
"followers_url": "https://api.github.com/users/... | [] | closed | false | null | [] | [
"The CI failed but it seems irrelevant. Could you please give it a check? @LysandreJik ",
"The CI issue is with `test_hf_api`, which is fixed on master now, rebasing should make the CI green!",
"@sgugger Sorry but I can't really agree with you here. It is a problem introduced by mistake and we shouldn't just tr... | 1,613 | 1,651 | 1,618 | CONTRIBUTOR | null | In this PR:
- Upgrade `datasets` to `1.3.0`
- Rename `datasets` variable to `task_datasets` in `run_glue.py` to avoid confusion with the library `datasets`
- Add a `--mnli_compat_mode` option to use the old label assignment for MNLI | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10203/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/huggingface/transformers/issues/10203/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10203",
"html_url": "https://github.com/huggingface/transformers/pull/10203",
"diff_url": "https://github.com/huggingface/transformers/pull/10203.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10203.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10202 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10202/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10202/comments | https://api.github.com/repos/huggingface/transformers/issues/10202/events | https://github.com/huggingface/transformers/issues/10202 | 809,000,058 | MDU6SXNzdWU4MDkwMDAwNTg= | 10,202 | Fast Tokenizers instantiated via vocab/merge files do not respect skip_special_tokens=True | {
"login": "minimaxir",
"id": 2179708,
"node_id": "MDQ6VXNlcjIxNzk3MDg=",
"avatar_url": "https://avatars.githubusercontent.com/u/2179708?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/minimaxir",
"html_url": "https://github.com/minimaxir",
"followers_url": "https://api.github.com/users/mi... | [] | closed | false | null | [] | [
"Indeed, I can reproduce! Do you know what might be causing this @n1t0?",
"This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelin... | 1,613 | 1,618 | 1,618 | NONE | null | ## Environment info
- `transformers` version: 4.3.2
- Platform: macOS-11.2.1-x86_64-i386-64bit
- Python version: 3.9.1
- PyTorch version (GPU?): 1.7.1 (False)
- Tensorflow version (GPU?): not installed (NA)
- Using GPU in script?: No
- Using distributed or parallel set-up in script?: No
## Information
See ... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10202/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/huggingface/transformers/issues/10202/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10201 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10201/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10201/comments | https://api.github.com/repos/huggingface/transformers/issues/10201/events | https://github.com/huggingface/transformers/issues/10201 | 808,961,103 | MDU6SXNzdWU4MDg5NjExMDM= | 10,201 | Better Fine-Tuning by Reducing Representational Collapse | {
"login": "mingruimingrui",
"id": 18568364,
"node_id": "MDQ6VXNlcjE4NTY4MzY0",
"avatar_url": "https://avatars.githubusercontent.com/u/18568364?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/mingruimingrui",
"html_url": "https://github.com/mingruimingrui",
"followers_url": "https://api.gi... | [
{
"id": 2648621985,
"node_id": "MDU6TGFiZWwyNjQ4NjIxOTg1",
"url": "https://api.github.com/repos/huggingface/transformers/labels/Feature%20request",
"name": "Feature request",
"color": "FBCA04",
"default": false,
"description": "Request for a new feature"
}
] | open | false | null | [] | [
"Would be lovely to see it, seems promising!",
"Any progress? @LysandreJik "
] | 1,613 | 1,656 | null | CONTRIBUTOR | null | # 🚀 Feature request
Add r3f/r4f to some popular objective functions as suggested by [Armen et. al](https://arxiv.org/abs/2008.03156).
## Motivation
Finetuning is a primary use case of many users of the transformers library.
We can use r3f/r4f to reduce representation collapse by vocabulary inefficiencies.
T... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10201/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/huggingface/transformers/issues/10201/timeline | null | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10200 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10200/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10200/comments | https://api.github.com/repos/huggingface/transformers/issues/10200/events | https://github.com/huggingface/transformers/pull/10200 | 808,955,258 | MDExOlB1bGxSZXF1ZXN0NTczOTAxMTAw | 10,200 | Bugfix: Removal of padding_idx in BartLearnedPositionalEmbedding | {
"login": "mingruimingrui",
"id": 18568364,
"node_id": "MDQ6VXNlcjE4NTY4MzY0",
"avatar_url": "https://avatars.githubusercontent.com/u/18568364?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/mingruimingrui",
"html_url": "https://github.com/mingruimingrui",
"followers_url": "https://api.gi... | [] | closed | false | null | [] | [
"Thanks for the PR @mingruimingrui !\r\nBut I'm not sure how filling the embeddings with 0 will avoid them being learnable, and `nn.Embedding` actually handles this itself, if pad index is specified then the output of the embedding layer at that index will be all zeros.\r\n\r\nPlus BART's offest is very specific to... | 1,613 | 1,614 | 1,614 | CONTRIBUTOR | null | # What does this PR do?
This PR removes the unnecessary padding_idx argument from positional embedding and instead uses the pre-determined offset.
In the event that padding_idx > 2, positional embedding at some position can be fixed to 0 instead of being learnable.
## Before submitting
- [x] This PR fixes a t... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10200/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/huggingface/transformers/issues/10200/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10200",
"html_url": "https://github.com/huggingface/transformers/pull/10200",
"diff_url": "https://github.com/huggingface/transformers/pull/10200.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10200.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10199 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10199/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10199/comments | https://api.github.com/repos/huggingface/transformers/issues/10199/events | https://github.com/huggingface/transformers/issues/10199 | 808,750,705 | MDU6SXNzdWU4MDg3NTA3MDU= | 10,199 | StopIteration error happened | {
"login": "brightbsit",
"id": 75205503,
"node_id": "MDQ6VXNlcjc1MjA1NTAz",
"avatar_url": "https://avatars.githubusercontent.com/u/75205503?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/brightbsit",
"html_url": "https://github.com/brightbsit",
"followers_url": "https://api.github.com/use... | [] | closed | false | null | [] | [
"Hi! It seems that you're on a very old transformers library version. I would recommend you upgrade to a more recent versions, as this particular has been patched several months ago.",
"@LysandreJik Hi! thank you for answering. But I'm using version 4.3.2",
"Are you sure? The error happens on the following line... | 1,613 | 1,619 | 1,619 | NONE | null | I'm using cuda11.0 with RTX3090.(using ubuntu 18.04)
I don't know how can I solve this problem.
I saw some people solve this StopIteration error with downgrade torch.
But rtx3090 is only compatible with cuda 11.
Please help!
[pip list]
anytree 2.8.0
apex 0.1
boto3 1.... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10199/reactions",
"total_count": 3,
"+1": 3,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/huggingface/transformers/issues/10199/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10198 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10198/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10198/comments | https://api.github.com/repos/huggingface/transformers/issues/10198/events | https://github.com/huggingface/transformers/issues/10198 | 808,747,886 | MDU6SXNzdWU4MDg3NDc4ODY= | 10,198 | ONNX Export - cannot resolve operator 'Shape' with opsets: ai.onnx v11 | {
"login": "biro-mark",
"id": 58680214,
"node_id": "MDQ6VXNlcjU4NjgwMjE0",
"avatar_url": "https://avatars.githubusercontent.com/u/58680214?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/biro-mark",
"html_url": "https://github.com/biro-mark",
"followers_url": "https://api.github.com/users/... | [] | closed | false | null | [] | [
"Sorry I think I misinterpreted the operators table https://github.com/onnx/onnx/blob/master/docs/Operators.md and `Shape` should have been available since opset 1, getting an update in opset 13. \r\n\r\nThis seems to be an issue with `onnxjs` not implementing the full set of operators in opset 11."
] | 1,613 | 1,613 | 1,613 | NONE | null | ## Environment info
- `transformers` version: 4.3.2
- Platform: Windows-10-10.0.18362-SP0
- Python version: 3.8.7
- PyTorch version (GPU?): 1.7.1+cu110 (True)
- Tensorflow version (GPU?): not installed (NA)
- Using GPU in script?: yes
- Using distributed or parallel set-up in script?: no
### Who can help
@Ly... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10198/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/huggingface/transformers/issues/10198/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10197 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10197/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10197/comments | https://api.github.com/repos/huggingface/transformers/issues/10197/events | https://github.com/huggingface/transformers/issues/10197 | 808,733,108 | MDU6SXNzdWU4MDg3MzMxMDg= | 10,197 | Fine-tuning Seq2Seq models for Machine translation | {
"login": "MorenoLaQuatra",
"id": 10062811,
"node_id": "MDQ6VXNlcjEwMDYyODEx",
"avatar_url": "https://avatars.githubusercontent.com/u/10062811?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/MorenoLaQuatra",
"html_url": "https://github.com/MorenoLaQuatra",
"followers_url": "https://api.gi... | [] | closed | false | null | [] | [
"Maybe @patrickvonplaten or @patil-suraj can chime in here!",
"Hey @MorenoLaQuatra \r\n\r\nThe `run_seq2seq.py` supports translation with custom dataset. And you can use T5, mT5, MarianMT, mBART, mBART-50 for fine-tuning.\r\nhttps://github.com/huggingface/transformers/tree/master/examples/seq2seq\r\n\r\nAnd the b... | 1,613 | 1,619 | 1,619 | CONTRIBUTOR | null | Good morning,
@micheledaddetta1
We were experimenting with Seq2Seq models such as MarianMT or T5.
I was wondering if there is a common way to fine-tune those models with custom datasets for the machine translation task.
Specifically we did the following:
```python
embeddings = self.batch_encode_plus(senten... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10197/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/huggingface/transformers/issues/10197/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10196 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10196/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10196/comments | https://api.github.com/repos/huggingface/transformers/issues/10196/events | https://github.com/huggingface/transformers/pull/10196 | 808,729,550 | MDExOlB1bGxSZXF1ZXN0NTczNzE5NTIx | 10,196 | [CI] make the examples sub-group of tests run always | {
"login": "stas00",
"id": 10676103,
"node_id": "MDQ6VXNlcjEwNjc2MTAz",
"avatar_url": "https://avatars.githubusercontent.com/u/10676103?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stas00",
"html_url": "https://github.com/stas00",
"followers_url": "https://api.github.com/users/stas00/fo... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | the examples tests weren't running if some previous job failed. this fixes it to always run.
@LysandreJik | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10196/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/huggingface/transformers/issues/10196/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10196",
"html_url": "https://github.com/huggingface/transformers/pull/10196",
"diff_url": "https://github.com/huggingface/transformers/pull/10196.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10196.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10195 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10195/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10195/comments | https://api.github.com/repos/huggingface/transformers/issues/10195/events | https://github.com/huggingface/transformers/pull/10195 | 808,723,338 | MDExOlB1bGxSZXF1ZXN0NTczNzE0NDc5 | 10,195 | Specify dataset dtype | {
"login": "LysandreJik",
"id": 30755778,
"node_id": "MDQ6VXNlcjMwNzU1Nzc4",
"avatar_url": "https://avatars.githubusercontent.com/u/30755778?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/LysandreJik",
"html_url": "https://github.com/LysandreJik",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | MEMBER | null | There was an issue in `datasets` <1.3.0 where the datasets type wouldn't be kept. Since this bug was patched, we have to specify the correct type for the dataset.
Co-authored-by: Quentin Lhoest <lhoest.q@gmail.com>
| {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10195/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/huggingface/transformers/issues/10195/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10195",
"html_url": "https://github.com/huggingface/transformers/pull/10195",
"diff_url": "https://github.com/huggingface/transformers/pull/10195.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10195.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10194 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10194/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10194/comments | https://api.github.com/repos/huggingface/transformers/issues/10194/events | https://github.com/huggingface/transformers/issues/10194 | 808,691,866 | MDU6SXNzdWU4MDg2OTE4NjY= | 10,194 | Uploaded a new model but is not found on the hub. | {
"login": "zolekode",
"id": 25635679,
"node_id": "MDQ6VXNlcjI1NjM1Njc5",
"avatar_url": "https://avatars.githubusercontent.com/u/25635679?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/zolekode",
"html_url": "https://github.com/zolekode",
"followers_url": "https://api.github.com/users/zol... | [
{
"id": 1843244711,
"node_id": "MDU6TGFiZWwxODQzMjQ0NzEx",
"url": "https://api.github.com/repos/huggingface/transformers/labels/New%20model",
"name": "New model",
"color": "fbca04",
"default": false,
"description": ""
}
] | closed | false | null | [] | [
"Hi @zolekode ,\r\n\r\nthe folder structure is not quite correct:\r\n\r\nhttps://huggingface.co/flexudy/t5-small-wav2vec2-grammar-fixer/tree/main\r\n\r\nYou just need to move everything from the `t5-small-wav2vec2-grammar-fixer` folder to the root folder. Then it should work :hugs: ",
"Ah I see. awesome. Thanks a... | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # 🌟 New model addition
I recently added this model: https://huggingface.co/flexudy/t5-small-wav2vec2-grammar-fixer
However, I get this error whilst trying to download it.
```
Can't load tokenizer for 'flexudy/t5-small-wav2vec2-grammar-fixer'
```
How can I fix it please?
| {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10194/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/huggingface/transformers/issues/10194/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10193 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10193/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10193/comments | https://api.github.com/repos/huggingface/transformers/issues/10193/events | https://github.com/huggingface/transformers/issues/10193 | 808,632,319 | MDU6SXNzdWU4MDg2MzIzMTk= | 10,193 | Make use of our copy-consistency script for task-specific models | {
"login": "sgugger",
"id": 35901082,
"node_id": "MDQ6VXNlcjM1OTAxMDgy",
"avatar_url": "https://avatars.githubusercontent.com/u/35901082?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sgugger",
"html_url": "https://github.com/sgugger",
"followers_url": "https://api.github.com/users/sgugge... | [
{
"id": 1990918270,
"node_id": "MDU6TGFiZWwxOTkwOTE4Mjcw",
"url": "https://api.github.com/repos/huggingface/transformers/labels/Good%20First%20Issue",
"name": "Good First Issue",
"color": "bbf794",
"default": false,
"description": ""
},
{
"id": 2392046359,
"node_id": "MDU6TGF... | closed | false | null | [] | [
"Hi @sgugger I'm up for the mission. After looking though the code I think I have a basic understanding of what you mean. However, would you be able to provide an example on one model just for clarification? I'm a bit confused on the second step\r\n\r\n\r\n>modeling_bert.py\r\n\r\n```\r\n_CONFIG_FOR_DOC = \"BertCo... | 1,613 | 1,614 | 1,614 | COLLABORATOR | null | This is an intermediate issue, which is why it gets both the good first issue and good second issue tags.
We have an automated script to check when copies of the same code are consistent inside the library, which allows us to avoid subclassing and keep all code for one model's forward pass inside one file (see our [... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10193/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/huggingface/transformers/issues/10193/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10192 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10192/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10192/comments | https://api.github.com/repos/huggingface/transformers/issues/10192/events | https://github.com/huggingface/transformers/issues/10192 | 808,610,222 | MDU6SXNzdWU4MDg2MTAyMjI= | 10,192 | run_mlm.py not utilizing TPU | {
"login": "DarshanDeshpande",
"id": 39432636,
"node_id": "MDQ6VXNlcjM5NDMyNjM2",
"avatar_url": "https://avatars.githubusercontent.com/u/39432636?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/DarshanDeshpande",
"html_url": "https://github.com/DarshanDeshpande",
"followers_url": "https://... | [] | closed | false | null | [] | [
"`--pad_to_max_length False` is the reason you have a very slow training: this creates batches of different sequence lengths but TPUs need fixed shapes to be efficient.\r\n\r\nThere was a bug in our argument parser before that ignored bool setting like this, so it may be the reason you are seeing that slow down now... | 1,613 | 1,613 | 1,613 | NONE | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 4.3.2 and Latest version forked from github
- Platform: Linux (Colab env)
- Python version: 3.6
- PyTorch ve... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10192/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/huggingface/transformers/issues/10192/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10191 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10191/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10191/comments | https://api.github.com/repos/huggingface/transformers/issues/10191/events | https://github.com/huggingface/transformers/pull/10191 | 808,595,707 | MDExOlB1bGxSZXF1ZXN0NTczNjEwMzE0 | 10,191 | Making TF BART-like models XLA and AMP compliant | {
"login": "jplu",
"id": 959590,
"node_id": "MDQ6VXNlcjk1OTU5MA==",
"avatar_url": "https://avatars.githubusercontent.com/u/959590?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jplu",
"html_url": "https://github.com/jplu",
"followers_url": "https://api.github.com/users/jplu/followers",
... | [] | closed | false | null | [] | [
"I succeed to fix Marian and Pegasus, and my first guess was the good one. I basically reworked a bit how the embedding was created, and now it works in XLA_GPU. Of course, all the corresponding slow tests are passing, and the weights are properly loaded."
] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # What does this PR do?
This PR makes the TF BART-like models compliant with AMP and XLA. The main issue for XLA was all the asserts, XLA is not compliant with them (see the [TF doc](https://www.tensorflow.org/xla/known_issues)), so I had to disable them if the model is run with another mode than eager.
TF Marian... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10191/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/huggingface/transformers/issues/10191/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10191",
"html_url": "https://github.com/huggingface/transformers/pull/10191",
"diff_url": "https://github.com/huggingface/transformers/pull/10191.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10191.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10190 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10190/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10190/comments | https://api.github.com/repos/huggingface/transformers/issues/10190/events | https://github.com/huggingface/transformers/issues/10190 | 808,555,015 | MDU6SXNzdWU4MDg1NTUwMTU= | 10,190 | 0% GPU usage when using `hyperparameter_search` | {
"login": "neel04",
"id": 11617870,
"node_id": "MDQ6VXNlcjExNjE3ODcw",
"avatar_url": "https://avatars.githubusercontent.com/u/11617870?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/neel04",
"html_url": "https://github.com/neel04",
"followers_url": "https://api.github.com/users/neel04/fo... | [] | closed | false | null | [] | [
"I tried decorating a function that contains the tranier command like this:-\r\n```\r\n\r\n@ray.remote(num_cpus=3, num_gpus=1, accelerator_type=ray.accelerators.NVIDIA_TESLA_V100)\r\ndef search():\r\n trainer.hyperparameter_search(n_trials=100, compute_objective='accuracy', direction=\"maximize\", backend='ray',\r... | 1,613 | 1,613 | 1,613 | NONE | null | _##_ Environment info
- `transformers` version: 4.4.0.dev0
- Platform: Linux-4.19.112+-x86_64-with-Ubuntu-18.04-bionic
- Python version: 3.6.9
- PyTorch version (GPU?): 1.7.0+cu101 (True)
- Tensorflow version (GPU?): 2.4.1 (True)
- Using GPU in script?: Yes
- Using distributed or parallel set-up in script?: No... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10190/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/huggingface/transformers/issues/10190/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10189 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10189/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10189/comments | https://api.github.com/repos/huggingface/transformers/issues/10189/events | https://github.com/huggingface/transformers/pull/10189 | 808,539,545 | MDExOlB1bGxSZXF1ZXN0NTczNTYzNzc1 | 10,189 | Fix TF template | {
"login": "jplu",
"id": 959590,
"node_id": "MDQ6VXNlcjk1OTU5MA==",
"avatar_url": "https://avatars.githubusercontent.com/u/959590?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jplu",
"html_url": "https://github.com/jplu",
"followers_url": "https://api.github.com/users/jplu/followers",
... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # What does this PR do?
This PR fixes the TF template for the tests by adding the missing onnx boolean. | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10189/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/huggingface/transformers/issues/10189/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10189",
"html_url": "https://github.com/huggingface/transformers/pull/10189",
"diff_url": "https://github.com/huggingface/transformers/pull/10189.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10189.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10188 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10188/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10188/comments | https://api.github.com/repos/huggingface/transformers/issues/10188/events | https://github.com/huggingface/transformers/issues/10188 | 808,496,626 | MDU6SXNzdWU4MDg0OTY2MjY= | 10,188 | Failing Multi-GPU trainer test | {
"login": "LysandreJik",
"id": 30755778,
"node_id": "MDQ6VXNlcjMwNzU1Nzc4",
"avatar_url": "https://avatars.githubusercontent.com/u/30755778?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/LysandreJik",
"html_url": "https://github.com/LysandreJik",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [
"I have tried on my machine and the test passes, so the bug is linked to the setup of the machine executing the multi-GPU tests. I have never seen that error before but I would guess there is something wrong with nccl/cuda?",
"Retrieved the backtrace:\r\n\r\nThe error is: `RuntimeError: Address already in use`\r\... | 1,613 | 1,615 | 1,615 | MEMBER | null | This test is currently [failing in a multi-GPU setup](https://github.com/huggingface/transformers/runs/1902689616?check_suite_focus=true):
```
FAILED tests/test_trainer_distributed.py::TestTrainerDistributed::test_trainer
```
The error is the following:
```
RuntimeError: NCCL error in: /pytorch/torch/lib/c1... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10188/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/huggingface/transformers/issues/10188/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10187 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10187/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10187/comments | https://api.github.com/repos/huggingface/transformers/issues/10187/events | https://github.com/huggingface/transformers/pull/10187 | 808,455,983 | MDExOlB1bGxSZXF1ZXN0NTczNDk0ODE2 | 10,187 | Add new model to labels that should not stale | {
"login": "LysandreJik",
"id": 30755778,
"node_id": "MDQ6VXNlcjMwNzU1Nzc4",
"avatar_url": "https://avatars.githubusercontent.com/u/30755778?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/LysandreJik",
"html_url": "https://github.com/LysandreJik",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | MEMBER | null | The `New model` label gets added to the labels that should not become stale. | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10187/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/huggingface/transformers/issues/10187/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10187",
"html_url": "https://github.com/huggingface/transformers/pull/10187",
"diff_url": "https://github.com/huggingface/transformers/pull/10187.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10187.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10186 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10186/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10186/comments | https://api.github.com/repos/huggingface/transformers/issues/10186/events | https://github.com/huggingface/transformers/issues/10186 | 808,450,576 | MDU6SXNzdWU4MDg0NTA1NzY= | 10,186 | Support for DeBERTa V2 models | {
"login": "saichandrapandraju",
"id": 41769919,
"node_id": "MDQ6VXNlcjQxNzY5OTE5",
"avatar_url": "https://avatars.githubusercontent.com/u/41769919?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/saichandrapandraju",
"html_url": "https://github.com/saichandrapandraju",
"followers_url": "ht... | [] | closed | false | null | [] | [
"Hi, no workaround, we're working on the implementation now (https://github.com/huggingface/transformers/pull/10018). It should be available in a few days.",
"Thanks @LysandreJik ",
"@saichandrapandraju the PR was merged, so I think this issue can be closed now?",
"ok @yaysummeriscoming ,\r\n\r\nMay I know wh... | 1,613 | 1,619 | 1,614 | NONE | null | Hi,
I downloaded [DeBERTa V2-XLarge](https://github.com/microsoft/DeBERTa) from [here](https://huggingface.co/microsoft/deberta-v2-xlarge) and trying to implement V2-XLarge model but I'm getting this error -
**RuntimeError: Error(s) in loading state_dict for DebertaForSequenceClassification:
size mismatch for d... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10186/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/huggingface/transformers/issues/10186/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10185 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10185/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10185/comments | https://api.github.com/repos/huggingface/transformers/issues/10185/events | https://github.com/huggingface/transformers/issues/10185 | 808,423,710 | MDU6SXNzdWU4MDg0MjM3MTA= | 10,185 | Saving HF wrapped in Keras | {
"login": "saboof",
"id": 59536094,
"node_id": "MDQ6VXNlcjU5NTM2MDk0",
"avatar_url": "https://avatars.githubusercontent.com/u/59536094?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/saboof",
"html_url": "https://github.com/saboof",
"followers_url": "https://api.github.com/users/saboof/fo... | [] | closed | false | {
"login": "jplu",
"id": 959590,
"node_id": "MDQ6VXNlcjk1OTU5MA==",
"avatar_url": "https://avatars.githubusercontent.com/u/959590?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jplu",
"html_url": "https://github.com/jplu",
"followers_url": "https://api.github.com/users/jplu/followers",
... | [
{
"login": "jplu",
"id": 959590,
"node_id": "MDQ6VXNlcjk1OTU5MA==",
"avatar_url": "https://avatars.githubusercontent.com/u/959590?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jplu",
"html_url": "https://github.com/jplu",
"followers_url": "https://api.github.com/users... | [
"Hello @saboof!\r\n\r\nFirst of all, which version of Transformers and TF are you using? And can you share with us a Colab from which we can reproduce the issue. Thanks!!",
"version of transformers: transformers==4.2.2\r\nversion of TF2: tensorflow==2.3.1\r\n\r\nafter editing, the code is attached to the main iss... | 1,613 | 1,619 | 1,619 | NONE | null | Hi,
Trying to save a Keras model which has a HF model and a liner layer (Dense layer) on top of it.
To save a model, Keras requires that every layer would have a serialize_layer_fn implemented.
However it seems that HF models don't include this function.
Spending some time understanding and googling this issu... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10185/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/huggingface/transformers/issues/10185/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10184 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10184/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10184/comments | https://api.github.com/repos/huggingface/transformers/issues/10184/events | https://github.com/huggingface/transformers/pull/10184 | 808,403,145 | MDExOlB1bGxSZXF1ZXN0NTczNDUyMjQ0 | 10,184 | Fixing NER pipeline for list inputs. | {
"login": "Narsil",
"id": 204321,
"node_id": "MDQ6VXNlcjIwNDMyMQ==",
"avatar_url": "https://avatars.githubusercontent.com/u/204321?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Narsil",
"html_url": "https://github.com/Narsil",
"followers_url": "https://api.github.com/users/Narsil/follow... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # What does this PR do?
- Changes TokenArgumentHandler(*args) into (inputs) signature to follow `__call__` signature.
- Fixes the bug.
- Backward compatible for single sentences
- Not backward compatible for multiple sentences, but it "worked" only for same length sentences in tokens (the result was bogus as it c... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10184/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/huggingface/transformers/issues/10184/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10184",
"html_url": "https://github.com/huggingface/transformers/pull/10184",
"diff_url": "https://github.com/huggingface/transformers/pull/10184.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10184.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10183 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10183/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10183/comments | https://api.github.com/repos/huggingface/transformers/issues/10183/events | https://github.com/huggingface/transformers/pull/10183 | 808,355,105 | MDExOlB1bGxSZXF1ZXN0NTczNDEzMDY1 | 10,183 | BigBird | {
"login": "thevasudevgupta",
"id": 53136577,
"node_id": "MDQ6VXNlcjUzMTM2NTc3",
"avatar_url": "https://avatars.githubusercontent.com/u/53136577?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/thevasudevgupta",
"html_url": "https://github.com/thevasudevgupta",
"followers_url": "https://api... | [] | closed | false | null | [] | [
"Will BigBird-Pegasus be added, and then `BigBirdForConditionalGeneration` so that summarization will be possible?",
"Yes, we will be adding that soon.\r\n\r\n> Will BigBird-Pegasus be added, and then `BigBirdForConditionalGeneration` so that summarization will be possible?\r\n\r\n",
"Once pre-trained checkpoin... | 1,613 | 1,619 | 1,617 | CONTRIBUTOR | null | # What does this PR do?
This PR will add Google's BigBird "Roberta".
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awes... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10183/reactions",
"total_count": 22,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 10,
"rocket": 12,
"eyes": 0
} | https://api.github.com/repos/huggingface/transformers/issues/10183/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10183",
"html_url": "https://github.com/huggingface/transformers/pull/10183",
"diff_url": "https://github.com/huggingface/transformers/pull/10183.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10183.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10182 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10182/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10182/comments | https://api.github.com/repos/huggingface/transformers/issues/10182/events | https://github.com/huggingface/transformers/issues/10182 | 808,234,454 | MDU6SXNzdWU4MDgyMzQ0NTQ= | 10,182 | `super()` does not have `prepare_seq2seq_batch()` in `transformers/models/rag/tokenization_rag.py` | {
"login": "moyapchen",
"id": 72097364,
"node_id": "MDQ6VXNlcjcyMDk3MzY0",
"avatar_url": "https://avatars.githubusercontent.com/u/72097364?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/moyapchen",
"html_url": "https://github.com/moyapchen",
"followers_url": "https://api.github.com/users/... | [] | closed | false | null | [] | [
"Hi ! Thanks for reporting\r\n#10167 should fix this issue",
"Convenient to see that the fix was already in the pipeline. Thanks!"
] | 1,613 | 1,613 | 1,613 | NONE | null | ## Environment info
- `transformers` version: 4.3.2
- Platform: Linux-5.4.0-52-generic-x86_64-with-debian-bullseye-sid
- Python version: 3.6.12
- PyTorch version (GPU?): 1.7.1+cu110 (True)
- Tensorflow version (GPU?): not installed (NA)
- Using GPU in script?: N/A
- Using distributed or parallel set-up in scri... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10182/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/huggingface/transformers/issues/10182/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10181 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10181/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10181/comments | https://api.github.com/repos/huggingface/transformers/issues/10181/events | https://github.com/huggingface/transformers/issues/10181 | 808,100,525 | MDU6SXNzdWU4MDgxMDA1MjU= | 10,181 | Inconsistent loss computation? | {
"login": "LuCeHe",
"id": 9610770,
"node_id": "MDQ6VXNlcjk2MTA3NzA=",
"avatar_url": "https://avatars.githubusercontent.com/u/9610770?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/LuCeHe",
"html_url": "https://github.com/LuCeHe",
"followers_url": "https://api.github.com/users/LuCeHe/foll... | [] | closed | false | null | [] | [
"Hello! If you look inside the `TFGPT2LMHeadModel`, you'll see it automatically shifts the labels for you.\r\n\r\nThe model generates tokens given a past. It then compares the generated token to the \"true\" token contained in the labels you passed to it. If you shift the tokens as it is done in the model, you shou... | 1,613 | 1,619 | 1,619 | NONE | null | `transformers` version: 4.3.2, Python version: 3.7, Tensorflow version (GPU?): 2.3.1, Using GPU in script?: No, Using distributed or parallel set-up in script?: No
## To reproduce
Steps to reproduce the behavior:
``` python
import tensorflow as tf
from transformers import TFGPT2LMHeadModel
from transformers... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10181/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/huggingface/transformers/issues/10181/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10180 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10180/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10180/comments | https://api.github.com/repos/huggingface/transformers/issues/10180/events | https://github.com/huggingface/transformers/issues/10180 | 808,013,284 | MDU6SXNzdWU4MDgwMTMyODQ= | 10,180 | ONNX Export for Fine-Tuned DistilBertForTokenClassification | {
"login": "biro-mark",
"id": 58680214,
"node_id": "MDQ6VXNlcjU4NjgwMjE0",
"avatar_url": "https://avatars.githubusercontent.com/u/58680214?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/biro-mark",
"html_url": "https://github.com/biro-mark",
"followers_url": "https://api.github.com/users/... | [] | closed | false | null | [] | [
"I figured out how to do it. Within python, you have to use `model.save_pretrained(\"path/to/output_dir\")` where `model` is the fine-tuned `model = DistilBertForTokenClassification.from_pretrained('distilbert-base-uncased', num_labels=...)`\r\n\r\nThen inside an empty directory, run\r\n`python -m transformers.conv... | 1,613 | 1,613 | 1,613 | NONE | null | # 🚀 Feature request
I'd like to export a fine-tuned DistilBertForTokenClassification model to ONNX. Right now the conversion script convert_graph_to_onnx.py looks like it takes a string like "--model bert-base-cased" but I'd like to pass in the model object that I've fine tuned on my data set.
## Motivation
... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10180/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/huggingface/transformers/issues/10180/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10179 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10179/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10179/comments | https://api.github.com/repos/huggingface/transformers/issues/10179/events | https://github.com/huggingface/transformers/issues/10179 | 807,993,885 | MDU6SXNzdWU4MDc5OTM4ODU= | 10,179 | Why is the attention_mask added to the attn_weights instead of multiplying/masking? | {
"login": "fomalhautb",
"id": 14837467,
"node_id": "MDQ6VXNlcjE0ODM3NDY3",
"avatar_url": "https://avatars.githubusercontent.com/u/14837467?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/fomalhautb",
"html_url": "https://github.com/fomalhautb",
"followers_url": "https://api.github.com/use... | [] | closed | false | null | [] | [
"Hello, Maybe this comment can help you out https://github.com/huggingface/transformers/issues/1935#issuecomment-561305086!",
"@LysandreJik Oh, I got it now, thank you!"
] | 1,613 | 1,613 | 1,613 | NONE | null | https://github.com/huggingface/transformers/blob/8fae93ca1972c39d19c8cf3d3c6a3dd2530cc59a/src/transformers/models/bart/modeling_bart.py#L219-L227
As far as I understand, attention_mask is to prevent the model peek into the future or padded positions, shouldn't the weights in these positions be masked out? What does ... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10179/reactions",
"total_count": 2,
"+1": 2,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/huggingface/transformers/issues/10179/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10178 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10178/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10178/comments | https://api.github.com/repos/huggingface/transformers/issues/10178/events | https://github.com/huggingface/transformers/pull/10178 | 807,983,671 | MDExOlB1bGxSZXF1ZXN0NTczMTA3NjYw | 10,178 | Fix datasets set_format | {
"login": "sgugger",
"id": 35901082,
"node_id": "MDQ6VXNlcjM1OTAxMDgy",
"avatar_url": "https://avatars.githubusercontent.com/u/35901082?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sgugger",
"html_url": "https://github.com/sgugger",
"followers_url": "https://api.github.com/users/sgugge... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | COLLABORATOR | null | # What does this PR do?
This PR fixes a problem in `Trainer` when user provide a dataset using the new functionality in the upcoming v2 of `datasets` `set_transform` (see [here](https://github.com/huggingface/datasets/issues/1867) for more details). This is a hotfix that is not perfect and we will need to take some ... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10178/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/huggingface/transformers/issues/10178/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10178",
"html_url": "https://github.com/huggingface/transformers/pull/10178",
"diff_url": "https://github.com/huggingface/transformers/pull/10178.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10178.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10177 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10177/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10177/comments | https://api.github.com/repos/huggingface/transformers/issues/10177/events | https://github.com/huggingface/transformers/issues/10177 | 807,970,091 | MDU6SXNzdWU4MDc5NzAwOTE= | 10,177 | Loading a model from local files achieves way too lower accuracy in comparison to model downloading | {
"login": "sstojanoska",
"id": 17052700,
"node_id": "MDQ6VXNlcjE3MDUyNzAw",
"avatar_url": "https://avatars.githubusercontent.com/u/17052700?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sstojanoska",
"html_url": "https://github.com/sstojanoska",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [
"> The problem arises when using the model from local files. I have noticed that the accuracy of the local model for the exact same configuration (same data, number of epochs, lr, etc.) is around 20% lower than downloading the model for each experiment.\r\n\r\nThis is extremely vague and we can't help you solve you... | 1,613 | 1,613 | 1,613 | NONE | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 4.3.2
- Platform: Linux-4.19.112+-x86_64-with-Ubuntu-18.04-bionic
- Python version: 3.6.9
- PyTorch version ... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10177/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/huggingface/transformers/issues/10177/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10176 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10176/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10176/comments | https://api.github.com/repos/huggingface/transformers/issues/10176/events | https://github.com/huggingface/transformers/issues/10176 | 807,926,442 | MDU6SXNzdWU4MDc5MjY0NDI= | 10,176 | Conditional generation with T5 | {
"login": "ShivanshuPurohit",
"id": 42869065,
"node_id": "MDQ6VXNlcjQyODY5MDY1",
"avatar_url": "https://avatars.githubusercontent.com/u/42869065?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ShivanshuPurohit",
"html_url": "https://github.com/ShivanshuPurohit",
"followers_url": "https://... | [] | closed | false | null | [] | [
"You can decode them back to a string using `T5Tokenizer`, like so:\r\n\r\n`tokenizer.decode(outputs.squeeze().tolist(), skip_special_tokens=True)`\r\n\r\nBtw, for a really good guide on the different generation strategies of models like T5, see this blog post: https://huggingface.co/blog/how-to-generate",
"This ... | 1,613 | 1,613 | 1,613 | NONE | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 4.3.2
- Platform: Linux-4.19.112+-x86_64-with-Ubuntu-18.04-bionic
- Python version: 3.6.9
- PyTorch version ... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10176/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/huggingface/transformers/issues/10176/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10175 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10175/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10175/comments | https://api.github.com/repos/huggingface/transformers/issues/10175/events | https://github.com/huggingface/transformers/pull/10175 | 807,915,511 | MDExOlB1bGxSZXF1ZXN0NTczMDU3ODQy | 10,175 | Speech2TextTransformer | {
"login": "patil-suraj",
"id": 27137566,
"node_id": "MDQ6VXNlcjI3MTM3NTY2",
"avatar_url": "https://avatars.githubusercontent.com/u/27137566?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/patil-suraj",
"html_url": "https://github.com/patil-suraj",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [
"@patrickvonplaten, @sgugger , @LysandreJik The PR is now finalized and ready for your review :) ",
"I've added proper instructions to install the extra dependencies and addressed Patrick and Sylvain's comments regarding the docs and imports. All slow/non-slow tests are passing!\r\n\r\nMerging!",
"Edit: please... | 1,613 | 1,615 | 1,615 | MEMBER | null | # What does this PR do?
This PR adds the S2T model from [fairseq](https://github.com/pytorch/fairseq/tree/master/examples/speech_to_text) for end-to-end ASR and Speech-Translation (ST).
The model architecture is somewhat similar to the mBART model, except
- the encoder contains the convolutional subsampling modu... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10175/reactions",
"total_count": 5,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 3,
"rocket": 2,
"eyes": 0
} | https://api.github.com/repos/huggingface/transformers/issues/10175/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10175",
"html_url": "https://github.com/huggingface/transformers/pull/10175",
"diff_url": "https://github.com/huggingface/transformers/pull/10175.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10175.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10174 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10174/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10174/comments | https://api.github.com/repos/huggingface/transformers/issues/10174/events | https://github.com/huggingface/transformers/issues/10174 | 807,882,608 | MDU6SXNzdWU4MDc4ODI2MDg= | 10,174 | How to train an MBart model from scratch for a new language pair? | {
"login": "vineetha-thomas",
"id": 60310418,
"node_id": "MDQ6VXNlcjYwMzEwNDE4",
"avatar_url": "https://avatars.githubusercontent.com/u/60310418?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/vineetha-thomas",
"html_url": "https://github.com/vineetha-thomas",
"followers_url": "https://api... | [] | closed | false | null | [] | [
"Hello, thanks for opening an issue! We try to keep the github issues for bugs/feature requests.\r\nCould you ask your question on the [forum](https://discusss.huggingface.co) instead?\r\n\r\nThanks!"
] | 1,613 | 1,613 | 1,613 | NONE | null | I want to train an MBART model from scratch, for a new language pair, unsupervised translation. I have monolingual data from both languages. Specifically, how do I prepare the data for the same?
Currently I start with a code as follows
_tokenizer = MBartTokenizer.from_pretrained('./tokenizer_de_hsb.model') //My ... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10174/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/huggingface/transformers/issues/10174/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10173 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10173/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10173/comments | https://api.github.com/repos/huggingface/transformers/issues/10173/events | https://github.com/huggingface/transformers/issues/10173 | 807,853,292 | MDU6SXNzdWU4MDc4NTMyOTI= | 10,173 | What does the "<s> token" mean in Longformer's global_attention_mask? | {
"login": "Anthonyive",
"id": 8257285,
"node_id": "MDQ6VXNlcjgyNTcyODU=",
"avatar_url": "https://avatars.githubusercontent.com/u/8257285?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Anthonyive",
"html_url": "https://github.com/Anthonyive",
"followers_url": "https://api.github.com/users... | [] | closed | false | null | [] | [
"Ok, I got it. That means [CLS] token."
] | 1,613 | 1,613 | 1,613 | NONE | null | This might be a stupid question, but I couldn't find an answer. The documentation says "For example, for classification, the \<s\> token should be given global attention.". I've also checked the original [longformer paper](https://arxiv.org/pdf/2004.05150.pdf), but "\<s\> token" was only mentioned once. Can someone tel... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10173/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/huggingface/transformers/issues/10173/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10172 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10172/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10172/comments | https://api.github.com/repos/huggingface/transformers/issues/10172/events | https://github.com/huggingface/transformers/issues/10172 | 807,791,750 | MDU6SXNzdWU4MDc3OTE3NTA= | 10,172 | Saving PruneBERT notebook fails to run on torch > 1.5 | {
"login": "lewtun",
"id": 26859204,
"node_id": "MDQ6VXNlcjI2ODU5MjA0",
"avatar_url": "https://avatars.githubusercontent.com/u/26859204?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/lewtun",
"html_url": "https://github.com/lewtun",
"followers_url": "https://api.github.com/users/lewtun/fo... | [] | closed | false | null | [] | [
"Thanks for reporting that @lewtun!\r\nFeel free to open a PR when you have a working solution for higher versions of PyTorch!",
"This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note th... | 1,613 | 1,619 | 1,619 | MEMBER | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 4.4.0.dev0
- Platform: macOS-10.16-x86_64-i386-64bit
- Python version: 3.8.5
- PyTorch version (GPU?): 1.7.1... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10172/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/huggingface/transformers/issues/10172/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10171 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10171/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10171/comments | https://api.github.com/repos/huggingface/transformers/issues/10171/events | https://github.com/huggingface/transformers/pull/10171 | 807,748,323 | MDExOlB1bGxSZXF1ZXN0NTcyOTM2NDYw | 10,171 | Revert propagation | {
"login": "LysandreJik",
"id": 30755778,
"node_id": "MDQ6VXNlcjMwNzU1Nzc4",
"avatar_url": "https://avatars.githubusercontent.com/u/30755778?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/LysandreJik",
"html_url": "https://github.com/LysandreJik",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | MEMBER | null | The proposition offered in https://github.com/huggingface/transformers/pull/10092 unfortunately can't be applied as having a default handler and propagation across handlers results in several logged items.
Reverting that PR here as seen offline with @lhoestq and leaving the docs regarding the default handler introdu... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10171/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/huggingface/transformers/issues/10171/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10171",
"html_url": "https://github.com/huggingface/transformers/pull/10171",
"diff_url": "https://github.com/huggingface/transformers/pull/10171.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10171.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10170 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10170/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10170/comments | https://api.github.com/repos/huggingface/transformers/issues/10170/events | https://github.com/huggingface/transformers/issues/10170 | 807,744,975 | MDU6SXNzdWU4MDc3NDQ5NzU= | 10,170 | T5 training with Keras: InvalidArgumentError: logits and labels must have the same first dimension | {
"login": "marton-avrios",
"id": 59836119,
"node_id": "MDQ6VXNlcjU5ODM2MTE5",
"avatar_url": "https://avatars.githubusercontent.com/u/59836119?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/marton-avrios",
"html_url": "https://github.com/marton-avrios",
"followers_url": "https://api.githu... | [] | closed | false | null | [] | [
"Would like to help you here, I've created a [Colab notebook](https://colab.research.google.com/drive/1PtRxbK4oNUsm4lrsOvWoNYzA-BhOWwf2?usp=sharing) that illustrates how to fine-tune `TFT5ForConditionalGeneration` using Keras. However, I'm having the same issue as posted in #6817, namely:\r\n\r\n`ValueError: No gra... | 1,613 | 1,619 | 1,619 | NONE | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 4.3.0.dev0
- Platform: Linux version 4.19.0-14-cloud-amd64 (debian-kernel@lists.debian.org) (gcc version 8.3.0... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10170/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/huggingface/transformers/issues/10170/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10169 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10169/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10169/comments | https://api.github.com/repos/huggingface/transformers/issues/10169/events | https://github.com/huggingface/transformers/issues/10169 | 807,735,522 | MDU6SXNzdWU4MDc3MzU1MjI= | 10,169 | run_langauge_modeling for T5 | {
"login": "ghost",
"id": 10137,
"node_id": "MDQ6VXNlcjEwMTM3",
"avatar_url": "https://avatars.githubusercontent.com/u/10137?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ghost",
"html_url": "https://github.com/ghost",
"followers_url": "https://api.github.com/users/ghost/followers",
"f... | [] | closed | false | null | [] | [
"Hi\r\n\r\nSeems to me this script is the repetition of this other script: transformers/examples/language-modeling/run_mlm.py \r\nDo you mind adding T5 also to this script? thanks ",
"Actually, we can not simply add T5 to this script, because `run_mlm.py` is for encoder-only models (such as BERT, RoBERTa, DeBERT... | 1,613 | 1,619 | 1,619 | NONE | null | Hi
Based on readme on [1], run_langauge_modeling.py does not support T5 model so far, it would be really nice to include this model as well.
There is also this line "data_args.block_size = tokenizer.max_len", max_len does not exist anymore, I searched in pretrainedTokernizer class and did not find an equivalent varia... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10169/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/huggingface/transformers/issues/10169/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10168 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10168/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10168/comments | https://api.github.com/repos/huggingface/transformers/issues/10168/events | https://github.com/huggingface/transformers/issues/10168 | 807,734,670 | MDU6SXNzdWU4MDc3MzQ2NzA= | 10,168 | NER pipeline doesn't work for a list of sequences | {
"login": "elk-cloner",
"id": 5828101,
"node_id": "MDQ6VXNlcjU4MjgxMDE=",
"avatar_url": "https://avatars.githubusercontent.com/u/5828101?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/elk-cloner",
"html_url": "https://github.com/elk-cloner",
"followers_url": "https://api.github.com/users... | [] | closed | false | null | [] | [
"@Narsil, do you want to take a look at this?",
"Took a look, it seems the issue was not padding, but argument handling.\r\n\r\n"
] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: transformers==4.3.2
- Platform: Linux Ubuntu 20.04
- Python version: 3.6
- PyTorch version (GPU?): torch==1... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10168/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/huggingface/transformers/issues/10168/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10167 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10167/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10167/comments | https://api.github.com/repos/huggingface/transformers/issues/10167/events | https://github.com/huggingface/transformers/pull/10167 | 807,693,819 | MDExOlB1bGxSZXF1ZXN0NTcyODk2OTA2 | 10,167 | [RAG] fix tokenizer | {
"login": "patil-suraj",
"id": 27137566,
"node_id": "MDQ6VXNlcjI3MTM3NTY2",
"avatar_url": "https://avatars.githubusercontent.com/u/27137566?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/patil-suraj",
"html_url": "https://github.com/patil-suraj",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | MEMBER | null | # What does this PR do?
- Introduce `as_target_tokenizer` context manager in `RagTokenizer` to later update the docs when `prepare_seq2seq_batch` is depricated.
- `RagTokenizer.prepare_seq2seq_batch` calls `super().prepare_seq2seq_batch`, but it does not inherit from `PreTrainedTokenizer`. Fix the method temporaril... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10167/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/huggingface/transformers/issues/10167/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10167",
"html_url": "https://github.com/huggingface/transformers/pull/10167",
"diff_url": "https://github.com/huggingface/transformers/pull/10167.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10167.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10166 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10166/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10166/comments | https://api.github.com/repos/huggingface/transformers/issues/10166/events | https://github.com/huggingface/transformers/issues/10166 | 807,683,843 | MDU6SXNzdWU4MDc2ODM4NDM= | 10,166 | [tests] failing test only when run in a group | {
"login": "stas00",
"id": 10676103,
"node_id": "MDQ6VXNlcjEwNjc2MTAz",
"avatar_url": "https://avatars.githubusercontent.com/u/10676103?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stas00",
"html_url": "https://github.com/stas00",
"followers_url": "https://api.github.com/users/stas00/fo... | [
{
"id": 1990918270,
"node_id": "MDU6TGFiZWwxOTkwOTE4Mjcw",
"url": "https://api.github.com/repos/huggingface/transformers/labels/Good%20First%20Issue",
"name": "Good First Issue",
"color": "bbf794",
"default": false,
"description": ""
},
{
"id": 2392046359,
"node_id": "MDU6TGF... | closed | false | null | [] | [
"Hi @stas00,\r\n\r\nI could not understand how to use `RUN_SLOW` in the windows command line, When I run it I was getting\r\n```\r\n'RUN_SLOW' is not recognized as an internal or external command,\r\noperable program or batch file.\r\n```\r\n\r\nIt was mentioned in the contribution guidelines but I don't know how t... | 1,613 | 1,615 | 1,615 | CONTRIBUTOR | null | If someone wants to solve a puzzle, this test:
```
RUN_SLOW=1 pytest examples/seq2seq/test_finetune_trainer.py::TestFinetuneTrainer::test_finetune_trainer_slow
```
works on its own, but fails if it's run in the group with other tests:
```
RUN_SLOW=1 pytest examples/seq2seq/test_finetune_trainer.py
```
it does... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10166/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/huggingface/transformers/issues/10166/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10165 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10165/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10165/comments | https://api.github.com/repos/huggingface/transformers/issues/10165/events | https://github.com/huggingface/transformers/issues/10165 | 807,669,826 | MDU6SXNzdWU4MDc2Njk4MjY= | 10,165 | [example scripts] inconsistency around eval vs val | {
"login": "stas00",
"id": 10676103,
"node_id": "MDQ6VXNlcjEwNjc2MTAz",
"avatar_url": "https://avatars.githubusercontent.com/u/10676103?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stas00",
"html_url": "https://github.com/stas00",
"followers_url": "https://api.github.com/users/stas00/fo... | [
{
"id": 1936351150,
"node_id": "MDU6TGFiZWwxOTM2MzUxMTUw",
"url": "https://api.github.com/repos/huggingface/transformers/labels/Examples",
"name": "Examples",
"color": "d4c5f9",
"default": false,
"description": "Which is related to examples in general"
},
{
"id": 2796628563,
... | closed | false | {
"login": "stas00",
"id": 10676103,
"node_id": "MDQ6VXNlcjEwNjc2MTAz",
"avatar_url": "https://avatars.githubusercontent.com/u/10676103?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stas00",
"html_url": "https://github.com/stas00",
"followers_url": "https://api.github.com/users/stas00/fo... | [
{
"login": "stas00",
"id": 10676103,
"node_id": "MDQ6VXNlcjEwNjc2MTAz",
"avatar_url": "https://avatars.githubusercontent.com/u/10676103?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stas00",
"html_url": "https://github.com/stas00",
"followers_url": "https://api.github... | [
"While what you say make sense, I'm unsure it warrants a new change of argument names on all example scripts as it seems more cosmetic to me.\r\n\r\nThe `TrainingArguments` have the proper mode already (`--do_train`, `--do_eval`, `--do_predict`) so it's only the examples. We're less attached to no breaking changes ... | 1,613 | 1,619 | 1,619 | CONTRIBUTOR | null | * `val` == validation set (split)
* `eval` == evaluation (mode)
those two are orthogonal to each other - one is a split, another is a model's run mode.
the trainer args and the scripts are inconsistent around when it's `val` and when it's `eval` in variable names and metrics.
examples:
* `eval_dataset` but ... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10165/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/huggingface/transformers/issues/10165/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10164 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10164/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10164/comments | https://api.github.com/repos/huggingface/transformers/issues/10164/events | https://github.com/huggingface/transformers/issues/10164 | 807,658,996 | MDU6SXNzdWU4MDc2NTg5OTY= | 10,164 | [example scripts] disambiguate language specification API | {
"login": "stas00",
"id": 10676103,
"node_id": "MDQ6VXNlcjEwNjc2MTAz",
"avatar_url": "https://avatars.githubusercontent.com/u/10676103?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stas00",
"html_url": "https://github.com/stas00",
"followers_url": "https://api.github.com/users/stas00/fo... | [] | closed | false | null | [] | [
"Regarding \"case 1\", only the \"old\" T5 models: `t5-small`, `t5-base`, `t5-large`, `t5-3b` and `t5-11b` were trained with the `source_prefix` and not the new T5 models. Also IMO, there is a very legitimate case that people might want to fine-tune `t5-small`, `t5-base`, ... on translation, but don't want to condi... | 1,613 | 1,616 | 1,616 | CONTRIBUTOR | null | Currently in example scripts like `run_seq2seq.py` we have:
1. for t5
```
--task translation_en_to_ro
--source_prefix "translate English to Romanian: "
```
2. Also these 2:
```
--target_lang ro_RO
--source_lang en_XX
```
are used only for MBart and are ignored for other models. Which means that people wi... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10164/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/huggingface/transformers/issues/10164/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10163 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10163/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10163/comments | https://api.github.com/repos/huggingface/transformers/issues/10163/events | https://github.com/huggingface/transformers/issues/10163 | 807,648,782 | MDU6SXNzdWU4MDc2NDg3ODI= | 10,163 | Increasing gradient accummulation steps significantly slows down training | {
"login": "keleog",
"id": 11840053,
"node_id": "MDQ6VXNlcjExODQwMDUz",
"avatar_url": "https://avatars.githubusercontent.com/u/11840053?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/keleog",
"html_url": "https://github.com/keleog",
"followers_url": "https://api.github.com/users/keleog/fo... | [] | closed | false | null | [] | [
"@sgugger \r\n@LysandreJik \r\n\r\npls help",
"A reported step is a training step (with an optimizer pass). When you increase gradient accumulation, you take more input batches to do one step, so it's normal to have less training steps per second.\r\n\r\nPlease note that the issues are for bugs and feature reques... | 1,613 | 1,613 | 1,613 | NONE | null | When training with a batch size of 32 (grad accummulation step = 1), training speed is approximately 6 it/s, however I increase gradient accummulation step to 4 or 8 (equivalent to batch size of 128 and 256), speed reduces to 1.03 it/s.
Is this expected behaviour?
## Environment info
<!-- You can run the comman... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10163/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/huggingface/transformers/issues/10163/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10162 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10162/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10162/comments | https://api.github.com/repos/huggingface/transformers/issues/10162/events | https://github.com/huggingface/transformers/pull/10162 | 807,646,281 | MDExOlB1bGxSZXF1ZXN0NTcyODYwMzk1 | 10,162 | fix run_seq2seq.py; porting trainer tests to it | {
"login": "stas00",
"id": 10676103,
"node_id": "MDQ6VXNlcjEwNjc2MTAz",
"avatar_url": "https://avatars.githubusercontent.com/u/10676103?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stas00",
"html_url": "https://github.com/stas00",
"followers_url": "https://api.github.com/users/stas00/fo... | [] | closed | false | null | [] | [
"OK, I decided to go ahead and port the other scripts instead of waiting for merging of the first set. Had to make some more fixes in the script while at it.\r\n\r\n\r\n"
] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | This PR:
- restores some of the essential dropped functionality from `finetune_trainer.py` - I'm almost sure this is far far from complete since so much was just dropped
- ports wmt_en_ro test data to `jsonlines` - I move the tests dataset into the root of examples so that it can be accessed by a variety of sub-pr... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10162/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/huggingface/transformers/issues/10162/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10162",
"html_url": "https://github.com/huggingface/transformers/pull/10162",
"diff_url": "https://github.com/huggingface/transformers/pull/10162.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10162.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10161 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10161/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10161/comments | https://api.github.com/repos/huggingface/transformers/issues/10161/events | https://github.com/huggingface/transformers/issues/10161 | 807,568,069 | MDU6SXNzdWU4MDc1NjgwNjk= | 10,161 | Seq2seq now has larger memory requirements, OOM w/Deepspeed on previously runnable models | {
"login": "PeterAJansen",
"id": 3813268,
"node_id": "MDQ6VXNlcjM4MTMyNjg=",
"avatar_url": "https://avatars.githubusercontent.com/u/3813268?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/PeterAJansen",
"html_url": "https://github.com/PeterAJansen",
"followers_url": "https://api.github.com... | [
{
"id": 2659267025,
"node_id": "MDU6TGFiZWwyNjU5MjY3MDI1",
"url": "https://api.github.com/repos/huggingface/transformers/labels/DeepSpeed",
"name": "DeepSpeed",
"color": "4D34F7",
"default": false,
"description": ""
}
] | closed | false | {
"login": "stas00",
"id": 10676103,
"node_id": "MDQ6VXNlcjEwNjc2MTAz",
"avatar_url": "https://avatars.githubusercontent.com/u/10676103?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stas00",
"html_url": "https://github.com/stas00",
"followers_url": "https://api.github.com/users/stas00/fo... | [
{
"login": "stas00",
"id": 10676103,
"node_id": "MDQ6VXNlcjEwNjc2MTAz",
"avatar_url": "https://avatars.githubusercontent.com/u/10676103?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stas00",
"html_url": "https://github.com/stas00",
"followers_url": "https://api.github... | [
"it's there:\r\n```\r\n./run_seq2seq.py -h | grep deepspeed\r\n [--sharded_ddp [SHARDED_DDP]] [--deepspeed DEEPSPEED]\r\n --deepspeed DEEPSPEED\r\n Enable deepspeed and pass the path to deepspeed json\r\n```\r\n\r\nof course, it would OOM w/o `--deepspeed` in your situat... | 1,613 | 1,619 | 1,619 | NONE | null | (A continuation of #10149 , since it looks like it's a broader issue:)
It looks like seq2seq has changed in the past week, and now gives out-of-memory errors for @stas00 's impressive recent DeepSpeed work that allowed training/predicting e.g. T5-11B on a single 40GB card.
Here's a simple repeatable example usin... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10161/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/huggingface/transformers/issues/10161/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10160 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10160/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10160/comments | https://api.github.com/repos/huggingface/transformers/issues/10160/events | https://github.com/huggingface/transformers/issues/10160 | 807,557,119 | MDU6SXNzdWU4MDc1NTcxMTk= | 10,160 | past_key_values tuple index out of range error when using text2text-generation pipeline with encoder-decoder model | {
"login": "thominj",
"id": 3819908,
"node_id": "MDQ6VXNlcjM4MTk5MDg=",
"avatar_url": "https://avatars.githubusercontent.com/u/3819908?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/thominj",
"html_url": "https://github.com/thominj",
"followers_url": "https://api.github.com/users/thominj/... | [] | closed | false | {
"login": "patil-suraj",
"id": 27137566,
"node_id": "MDQ6VXNlcjI3MTM3NTY2",
"avatar_url": "https://avatars.githubusercontent.com/u/27137566?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/patil-suraj",
"html_url": "https://github.com/patil-suraj",
"followers_url": "https://api.github.com/... | [
{
"login": "patil-suraj",
"id": 27137566,
"node_id": "MDQ6VXNlcjI3MTM3NTY2",
"avatar_url": "https://avatars.githubusercontent.com/u/27137566?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/patil-suraj",
"html_url": "https://github.com/patil-suraj",
"followers_url": "htt... | [
"I have been digging into this a little bit more and found some information that might be helpful. It looks like the underlying problem is in the EncoderDecoderModel or one of its dependencies, not the pipeline. \r\n\r\n- When I replaced the pipeline call with a manual tokenization and call to the model's generate ... | 1,613 | 1,645 | 1,621 | CONTRIBUTOR | null | ## Environment info
- `transformers` version: 4.3.0
- Platform: Linux-5.4.0-65-generic-x86_64-with-Ubuntu-20.04-focal
- Python version: 3.7.9
- PyTorch version (GPU?): 1.7.0 (False)
- Tensorflow version (GPU?): not installed (NA)
- Using GPU in script?: No
- Using distributed or parallel set-up in script?: No
... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10160/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/huggingface/transformers/issues/10160/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10159 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10159/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10159/comments | https://api.github.com/repos/huggingface/transformers/issues/10159/events | https://github.com/huggingface/transformers/pull/10159 | 807,528,062 | MDExOlB1bGxSZXF1ZXN0NTcyNzYyNjY4 | 10,159 | [hf_api] delete deprecated methods and tests | {
"login": "julien-c",
"id": 326577,
"node_id": "MDQ6VXNlcjMyNjU3Nw==",
"avatar_url": "https://avatars.githubusercontent.com/u/326577?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/julien-c",
"html_url": "https://github.com/julien-c",
"followers_url": "https://api.github.com/users/julien-... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | MEMBER | null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10159/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/huggingface/transformers/issues/10159/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10159",
"html_url": "https://github.com/huggingface/transformers/pull/10159",
"diff_url": "https://github.com/huggingface/transformers/pull/10159.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10159.patch",
"merged_at... | |
https://api.github.com/repos/huggingface/transformers/issues/10158 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10158/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10158/comments | https://api.github.com/repos/huggingface/transformers/issues/10158/events | https://github.com/huggingface/transformers/issues/10158 | 807,489,698 | MDU6SXNzdWU4MDc0ODk2OTg= | 10,158 | Multiple Mask support in Pipeline | {
"login": "naveenjafer",
"id": 7025448,
"node_id": "MDQ6VXNlcjcwMjU0NDg=",
"avatar_url": "https://avatars.githubusercontent.com/u/7025448?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/naveenjafer",
"html_url": "https://github.com/naveenjafer",
"followers_url": "https://api.github.com/us... | [
{
"id": 2648621985,
"node_id": "MDU6TGFiZWwyNjQ4NjIxOTg1",
"url": "https://api.github.com/repos/huggingface/transformers/labels/Feature%20request",
"name": "Feature request",
"color": "FBCA04",
"default": false,
"description": "Request for a new feature"
}
] | open | false | null | [] | [
"@LysandreJik \r\nThe current implementation for a single mask returns the data as a list of \r\n\r\n```\r\n{ \r\n \"sequence\" : \"the final sequence with the mask added\", \r\n \"score\" : \"the softmax score\", \r\n \"token\" : \"the token ID used in filling the MASK\", \r\n \"token_str\" : \"the to... | 1,613 | 1,613 | null | NONE | null | # 🚀 Feature request
The [fill mask](https://huggingface.co/bert-base-uncased?text=Paris+is+the+capital+of+%5BMASK%5D+%3F) feature as a part of the pipeline currently only supports a single mask for the inputs. It could be expanded to predict and return the results for multiple masks in the same sentence too.
## ... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10158/reactions",
"total_count": 9,
"+1": 9,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/huggingface/transformers/issues/10158/timeline | null | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10157 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10157/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10157/comments | https://api.github.com/repos/huggingface/transformers/issues/10157/events | https://github.com/huggingface/transformers/pull/10157 | 807,429,947 | MDExOlB1bGxSZXF1ZXN0NTcyNjgyMTIx | 10,157 | Fix typo in comments | {
"login": "mrm8488",
"id": 3653789,
"node_id": "MDQ6VXNlcjM2NTM3ODk=",
"avatar_url": "https://avatars.githubusercontent.com/u/3653789?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/mrm8488",
"html_url": "https://github.com/mrm8488",
"followers_url": "https://api.github.com/users/mrm8488/... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this w... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10157/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/huggingface/transformers/issues/10157/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10157",
"html_url": "https://github.com/huggingface/transformers/pull/10157",
"diff_url": "https://github.com/huggingface/transformers/pull/10157.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10157.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10156 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10156/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10156/comments | https://api.github.com/repos/huggingface/transformers/issues/10156/events | https://github.com/huggingface/transformers/pull/10156 | 807,428,818 | MDExOlB1bGxSZXF1ZXN0NTcyNjgxMjA0 | 10,156 | Fix typo in comment | {
"login": "mrm8488",
"id": 3653789,
"node_id": "MDQ6VXNlcjM2NTM3ODk=",
"avatar_url": "https://avatars.githubusercontent.com/u/3653789?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/mrm8488",
"html_url": "https://github.com/mrm8488",
"followers_url": "https://api.github.com/users/mrm8488/... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this w... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10156/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/huggingface/transformers/issues/10156/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10156",
"html_url": "https://github.com/huggingface/transformers/pull/10156",
"diff_url": "https://github.com/huggingface/transformers/pull/10156.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10156.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10155 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10155/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10155/comments | https://api.github.com/repos/huggingface/transformers/issues/10155/events | https://github.com/huggingface/transformers/issues/10155 | 807,427,367 | MDU6SXNzdWU4MDc0MjczNjc= | 10,155 | rfc: integration tests need non-example application for testing | {
"login": "stas00",
"id": 10676103,
"node_id": "MDQ6VXNlcjEwNjc2MTAz",
"avatar_url": "https://avatars.githubusercontent.com/u/10676103?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/stas00",
"html_url": "https://github.com/stas00",
"followers_url": "https://api.github.com/users/stas00/fo... | [
{
"id": 1834088753,
"node_id": "MDU6TGFiZWwxODM0MDg4NzUz",
"url": "https://api.github.com/repos/huggingface/transformers/labels/Tests",
"name": "Tests",
"color": "a6fcca",
"default": false,
"description": "Related to tests"
},
{
"id": 2604155188,
"node_id": "MDU6TGFiZWwyNjA0M... | closed | false | null | [] | [
"I'm all for having core integration tests to do regression testing. As you have said, these tests should not be under `examples/` as that is a dedicated `examples/` folder, but should be under `tests/`.\r\n\r\nI'm not 100% sure whether we would want that in existing testing files (for example a BART regression tes... | 1,613 | 1,616 | 1,616 | CONTRIBUTOR | null | # 🚀 Feature request
We have an ongoing conflict with some of the core integration tests needing a serious program to be tested with. The only place these can be found is under `examples/` - and so the tests - e.g. deepspeed/apex/fairscale reside under `examples/` because of that.
The problem is that because they a... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10155/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/huggingface/transformers/issues/10155/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10154 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10154/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10154/comments | https://api.github.com/repos/huggingface/transformers/issues/10154/events | https://github.com/huggingface/transformers/pull/10154 | 807,327,286 | MDExOlB1bGxSZXF1ZXN0NTcyNTk1NzA2 | 10,154 | Add mBART-50 | {
"login": "patil-suraj",
"id": 27137566,
"node_id": "MDQ6VXNlcjI3MTM3NTY2",
"avatar_url": "https://avatars.githubusercontent.com/u/27137566?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/patil-suraj",
"html_url": "https://github.com/patil-suraj",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [
"> As a follow-up, is mBART aligned with mBART-50? We should have the same setter there. It would make a nice first issue I believe, once this PR is merged and provides a good model.\r\n\r\nYes. Here the setter was necessary because of the many to many models. But yes mBART can also be used for multilingual fine-tu... | 1,613 | 1,613 | 1,613 | MEMBER | null | # What does this PR do?
This is the second part of splitting #9811
This PR adds the mBART-50 models.
- Add `MBart50Tokenizer` and `MBart50TokenizerFast`. A new tokenizer is needed because it adds extra languages and the encoding format is different than `MBartTokenizer`. The difference is that for `mbart-50` bot... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10154/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/huggingface/transformers/issues/10154/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10154",
"html_url": "https://github.com/huggingface/transformers/pull/10154",
"diff_url": "https://github.com/huggingface/transformers/pull/10154.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10154.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10153 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10153/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10153/comments | https://api.github.com/repos/huggingface/transformers/issues/10153/events | https://github.com/huggingface/transformers/pull/10153 | 807,323,341 | MDExOlB1bGxSZXF1ZXN0NTcyNTkyNDkw | 10,153 | I-BERT model support | {
"login": "kssteven418",
"id": 50283958,
"node_id": "MDQ6VXNlcjUwMjgzOTU4",
"avatar_url": "https://avatars.githubusercontent.com/u/50283958?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/kssteven418",
"html_url": "https://github.com/kssteven418",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [
"Actually as @patrickvonplaten correctly mentioned, we really need some test files before we can merge this.",
"@kssteven418, \r\n\r\nThanks a mille for your PR - that's an amazing contribution!\r\n\r\nI think before merging we still do need to do a couple of things:\r\n\r\n1) **Tests** - it seems that currently ... | 1,613 | 1,614 | 1,614 | CONTRIBUTOR | null | # What does this PR do?
This PR implements [I-BERT](https://arxiv.org/abs/2101.01321), an integer-only quantization scheme for Transformer architectures. I-BERT is based on the model architecture and the pre-trained parameters of RoBERTa (this can be extended to other architectures as a future task), except that it ... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10153/reactions",
"total_count": 2,
"+1": 2,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/huggingface/transformers/issues/10153/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10153",
"html_url": "https://github.com/huggingface/transformers/pull/10153",
"diff_url": "https://github.com/huggingface/transformers/pull/10153.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10153.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10152 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10152/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10152/comments | https://api.github.com/repos/huggingface/transformers/issues/10152/events | https://github.com/huggingface/transformers/pull/10152 | 807,273,980 | MDExOlB1bGxSZXF1ZXN0NTcyNTUxMDg4 | 10,152 | Reduce the time spent for the TF slow tests | {
"login": "jplu",
"id": 959590,
"node_id": "MDQ6VXNlcjk1OTU5MA==",
"avatar_url": "https://avatars.githubusercontent.com/u/959590?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jplu",
"html_url": "https://github.com/jplu",
"followers_url": "https://api.github.com/users/jplu/followers",
... | [] | closed | false | null | [] | [
"@sgugger Yes, this is exactly that. There was an important overlap across these three tests (all based on creating a saved model and two on testing the output), so merging them was IMO the best way to keep the coverage and reduce the time.\r\n\r\n@patrickvonplaten feel free to merge if the PR looks ok for you!"
] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # What does this PR do?
This PR reduces by half the time spent on running the all the tests (including the slow tests). Here are the time comparison (time recorded on my machine with the models already downloaded):
- albert: from 13mins to 6mins
- bart: from 19mins to 9mins
- bert: from 17mins to 9mins
- blend... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10152/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/huggingface/transformers/issues/10152/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10152",
"html_url": "https://github.com/huggingface/transformers/pull/10152",
"diff_url": "https://github.com/huggingface/transformers/pull/10152.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10152.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10151 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10151/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10151/comments | https://api.github.com/repos/huggingface/transformers/issues/10151/events | https://github.com/huggingface/transformers/issues/10151 | 807,193,248 | MDU6SXNzdWU4MDcxOTMyNDg= | 10,151 | Model Parallelism for Bert Models | {
"login": "saichandrapandraju",
"id": 41769919,
"node_id": "MDQ6VXNlcjQxNzY5OTE5",
"avatar_url": "https://avatars.githubusercontent.com/u/41769919?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/saichandrapandraju",
"html_url": "https://github.com/saichandrapandraju",
"followers_url": "ht... | [
{
"id": 2627272588,
"node_id": "MDU6TGFiZWwyNjI3MjcyNTg4",
"url": "https://api.github.com/repos/huggingface/transformers/labels/Model%20Parallel",
"name": "Model Parallel",
"color": "8B66A5",
"default": false,
"description": "Model Parallelilsm Implementations"
}
] | closed | false | null | [] | [
"We already have naive vertical MP implemented in t5 and gpt, and there is a much easier version of Bart MP - but it's not merged (https://github.com/huggingface/transformers/pull/9384).\r\n\r\nThe problem with naive MP is that it's very inefficient. That's why at the moment the rest of transformers isn't being por... | 1,613 | 1,694 | 1,613 | NONE | null | Hi,
I'm trying to implement Model parallelism for BERT models by splitting and assigning layers across GPUs. I took DeBERTa as an example for this.
For DeBERTa, I'm able to split entire model into 'embedding', 'encoder', 'pooler', 'classifier' and 'dropout' layers as shown in below pic.
![Capture](https://user-... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10151/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/huggingface/transformers/issues/10151/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10150 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10150/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10150/comments | https://api.github.com/repos/huggingface/transformers/issues/10150/events | https://github.com/huggingface/transformers/issues/10150 | 807,130,110 | MDU6SXNzdWU4MDcxMzAxMTA= | 10,150 | Problem with evaluation_strategy | {
"login": "mtortoli",
"id": 29463872,
"node_id": "MDQ6VXNlcjI5NDYzODcy",
"avatar_url": "https://avatars.githubusercontent.com/u/29463872?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/mtortoli",
"html_url": "https://github.com/mtortoli",
"followers_url": "https://api.github.com/users/mto... | [] | closed | false | {
"login": "sgugger",
"id": 35901082,
"node_id": "MDQ6VXNlcjM1OTAxMDgy",
"avatar_url": "https://avatars.githubusercontent.com/u/35901082?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sgugger",
"html_url": "https://github.com/sgugger",
"followers_url": "https://api.github.com/users/sgugge... | [
{
"login": "sgugger",
"id": 35901082,
"node_id": "MDQ6VXNlcjM1OTAxMDgy",
"avatar_url": "https://avatars.githubusercontent.com/u/35901082?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sgugger",
"html_url": "https://github.com/sgugger",
"followers_url": "https://api.git... | [
"`evaluation_strategy` is not an argument fully implemented in `TFTrainer`, it only supports \"steps\". (The PyTorch counterpart supports all the possibilities.)\r\nTo evaluate every epoch, the best is to use the native Keras fit method.",
"Thanks a lot!!"
] | 1,613 | 1,613 | 1,613 | NONE | null | Hi everyone!
I have a problem (i think is a bug but i'm not sure) with the parameter "evaluation_strategy" in TFTrainingArguments.
I created a script for finetuning a transfomers model, based on the example "run_tf_text_classification.py" file.
In "TFTrainingArguments" i put the parameter "evaluation_strategy="epo... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10150/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/huggingface/transformers/issues/10150/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10149 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10149/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10149/comments | https://api.github.com/repos/huggingface/transformers/issues/10149/events | https://github.com/huggingface/transformers/issues/10149 | 807,039,226 | MDU6SXNzdWU4MDcwMzkyMjY= | 10,149 | Issue using num_beams parameter for T5 / DeepSpeed | {
"login": "PeterAJansen",
"id": 3813268,
"node_id": "MDQ6VXNlcjM4MTMyNjg=",
"avatar_url": "https://avatars.githubusercontent.com/u/3813268?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/PeterAJansen",
"html_url": "https://github.com/PeterAJansen",
"followers_url": "https://api.github.com... | [] | closed | false | null | [] | [
"It's `--eval_beams` in that particular script:\r\n```\r\n./finetune_trainer.py -h | grep beams\r\n [--tgt_lang TGT_LANG] [--eval_beams EVAL_BEAMS]\r\n --eval_beams EVAL_BEAMS\r\n # num_beams to use for evaluation.\r\n```\r\n\r\nThis script is going to be retired so... | 1,613 | 1,613 | 1,613 | NONE | null | Using a fine-turned seq2seq model, I'd like to generate some number of possible different generations for a given input. One way of typically doing this is using beam search.
Using @stas00 's amazing DeepSpeed additions so that T5-11B will fit in my GPUs, I'm calling the trainer ( finetune_trainer.py
) with only... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10149/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/huggingface/transformers/issues/10149/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10148 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10148/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10148/comments | https://api.github.com/repos/huggingface/transformers/issues/10148/events | https://github.com/huggingface/transformers/pull/10148 | 806,957,026 | MDExOlB1bGxSZXF1ZXN0NTcyMjg1MDU1 | 10,148 | Fix typo in GPT2DoubleHeadsModel docs | {
"login": "M-Salti",
"id": 9285264,
"node_id": "MDQ6VXNlcjkyODUyNjQ=",
"avatar_url": "https://avatars.githubusercontent.com/u/9285264?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/M-Salti",
"html_url": "https://github.com/M-Salti",
"followers_url": "https://api.github.com/users/M-Salti/... | [] | closed | false | null | [] | [] | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | If I'm not mistaken, masked label ids should be set to `-100` not `-1`
| {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10148/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/huggingface/transformers/issues/10148/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10148",
"html_url": "https://github.com/huggingface/transformers/pull/10148",
"diff_url": "https://github.com/huggingface/transformers/pull/10148.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10148.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10147 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10147/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10147/comments | https://api.github.com/repos/huggingface/transformers/issues/10147/events | https://github.com/huggingface/transformers/issues/10147 | 806,702,135 | MDU6SXNzdWU4MDY3MDIxMzU= | 10,147 | BERT with regression head cannot fit one datapoint | {
"login": "lucky-bai",
"id": 123435,
"node_id": "MDQ6VXNlcjEyMzQzNQ==",
"avatar_url": "https://avatars.githubusercontent.com/u/123435?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/lucky-bai",
"html_url": "https://github.com/lucky-bai",
"followers_url": "https://api.github.com/users/luck... | [] | closed | false | null | [] | [
"Hello, thanks for opening an issue! We try to keep the github issues for bugs/feature requests.\r\nCould you ask your question on the [forum](https://discusss.huggingface.co) instead?\r\n\r\nThanks!"
] | 1,613 | 1,613 | 1,613 | NONE | null | Hi, I am trying to use BERT for a token-level regression task (predict a continuous value for each token), and I'm having trouble getting my model to train. As a debugging strategy, I'm trying to get it to overfit one datapoint, which should be easy, but it's failing that also.
Here is a minimal reproducing source c... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10147/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/huggingface/transformers/issues/10147/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10146 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10146/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10146/comments | https://api.github.com/repos/huggingface/transformers/issues/10146/events | https://github.com/huggingface/transformers/issues/10146 | 806,670,345 | MDU6SXNzdWU4MDY2NzAzNDU= | 10,146 | Model not training beyond 1st epoch | {
"login": "neel04",
"id": 11617870,
"node_id": "MDQ6VXNlcjExNjE3ODcw",
"avatar_url": "https://avatars.githubusercontent.com/u/11617870?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/neel04",
"html_url": "https://github.com/neel04",
"followers_url": "https://api.github.com/users/neel04/fo... | [] | closed | false | null | [] | [
"Could you please post this on the [forum](https://discuss.huggingface.co/), rather than here? The authors of HuggingFace like to keep this place for bugs or feature requests, and they're more than happy to help you on the forum.\r\n\r\nLooking at your code, this seems more like an issue with preparing the data cor... | 1,613 | 1,651 | 1,613 | NONE | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 4.4.0.dev0
- Platform: Linux-4.19.112+-x86_64-with-Ubuntu-18.04-bionic
- Python version: 3.6.9
- PyTorch v... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10146/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/huggingface/transformers/issues/10146/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10145 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10145/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10145/comments | https://api.github.com/repos/huggingface/transformers/issues/10145/events | https://github.com/huggingface/transformers/pull/10145 | 806,601,099 | MDExOlB1bGxSZXF1ZXN0NTcxOTkyMTUw | 10,145 | Add Fine-Tuning for Wav2Vec2 | {
"login": "patrickvonplaten",
"id": 23423619,
"node_id": "MDQ6VXNlcjIzNDIzNjE5",
"avatar_url": "https://avatars.githubusercontent.com/u/23423619?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/patrickvonplaten",
"html_url": "https://github.com/patrickvonplaten",
"followers_url": "https://... | [] | closed | false | null | [] | [
"This is really nice, and the piece that will make the wav2vec 2.0 stuff awesome and more readily available! let me know if I can assist in testing/whatnot :)"
] | 1,613 | 1,614 | 1,614 | MEMBER | null | # What does this PR do?
This PR adds the possibility to finetune Wav2Vec2 on a downstream task. I ran a couple of experiments and I think the training is pretty stable now, see this training run *e.g.*:
https://wandb.ai/patrickvonplaten/huggingface/reports/Project-Dashboard--Vmlldzo0OTI0OTc?accessToken=8azw8iyxnb... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10145/reactions",
"total_count": 19,
"+1": 9,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 8,
"rocket": 2,
"eyes": 0
} | https://api.github.com/repos/huggingface/transformers/issues/10145/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10145",
"html_url": "https://github.com/huggingface/transformers/pull/10145",
"diff_url": "https://github.com/huggingface/transformers/pull/10145.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10145.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10144 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10144/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10144/comments | https://api.github.com/repos/huggingface/transformers/issues/10144/events | https://github.com/huggingface/transformers/issues/10144 | 806,553,933 | MDU6SXNzdWU4MDY1NTM5MzM= | 10,144 | T5 Base length of Tokenizer not equal config vocab_size | {
"login": "ari9dam",
"id": 14134882,
"node_id": "MDQ6VXNlcjE0MTM0ODgy",
"avatar_url": "https://avatars.githubusercontent.com/u/14134882?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ari9dam",
"html_url": "https://github.com/ari9dam",
"followers_url": "https://api.github.com/users/ari9da... | [] | closed | false | {
"login": "patrickvonplaten",
"id": 23423619,
"node_id": "MDQ6VXNlcjIzNDIzNjE5",
"avatar_url": "https://avatars.githubusercontent.com/u/23423619?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/patrickvonplaten",
"html_url": "https://github.com/patrickvonplaten",
"followers_url": "https://... | [
{
"login": "patrickvonplaten",
"id": 23423619,
"node_id": "MDQ6VXNlcjIzNDIzNjE5",
"avatar_url": "https://avatars.githubusercontent.com/u/23423619?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/patrickvonplaten",
"html_url": "https://github.com/patrickvonplaten",
"follo... | [
"duplicate of https://github.com/huggingface/transformers/issues/4875 I think",
"This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing gu... | 1,613 | 1,619 | 1,619 | NONE | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: Installed from git
## Issue
The `len(AutoTokenizer.from_pretrained("t5-base"))` is `32100` but the `T5ForCo... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10144/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/huggingface/transformers/issues/10144/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10143 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10143/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10143/comments | https://api.github.com/repos/huggingface/transformers/issues/10143/events | https://github.com/huggingface/transformers/issues/10143 | 806,533,214 | MDU6SXNzdWU4MDY1MzMyMTQ= | 10,143 | context manager for seeding, or generating fixed random tensor. | {
"login": "sadakmed",
"id": 18331629,
"node_id": "MDQ6VXNlcjE4MzMxNjI5",
"avatar_url": "https://avatars.githubusercontent.com/u/18331629?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sadakmed",
"html_url": "https://github.com/sadakmed",
"followers_url": "https://api.github.com/users/sad... | [] | closed | false | null | [] | [
"This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md)... | 1,613 | 1,619 | 1,619 | CONTRIBUTOR | null | # 🚀 Feature request
context manager for torch random seed, where the seed is fixed only inside
## Motivation
in some integration test, an input required is very large to be hardcoded, and the ids_tensor provide only int32 examples.
However to fix this input either we can use NumPy seed, but probably it wi... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10143/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/huggingface/transformers/issues/10143/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10142 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10142/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10142/comments | https://api.github.com/repos/huggingface/transformers/issues/10142/events | https://github.com/huggingface/transformers/issues/10142 | 806,406,834 | MDU6SXNzdWU4MDY0MDY4MzQ= | 10,142 | T5 GPU Runtime Degradation | {
"login": "dsgissin",
"id": 20739375,
"node_id": "MDQ6VXNlcjIwNzM5Mzc1",
"avatar_url": "https://avatars.githubusercontent.com/u/20739375?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/dsgissin",
"html_url": "https://github.com/dsgissin",
"followers_url": "https://api.github.com/users/dsg... | [] | closed | false | {
"login": "patrickvonplaten",
"id": 23423619,
"node_id": "MDQ6VXNlcjIzNDIzNjE5",
"avatar_url": "https://avatars.githubusercontent.com/u/23423619?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/patrickvonplaten",
"html_url": "https://github.com/patrickvonplaten",
"followers_url": "https://... | [
{
"login": "patrickvonplaten",
"id": 23423619,
"node_id": "MDQ6VXNlcjIzNDIzNjE5",
"avatar_url": "https://avatars.githubusercontent.com/u/23423619?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/patrickvonplaten",
"html_url": "https://github.com/patrickvonplaten",
"follo... | [
"Thanks a lot for this issue @dsgissin! Will take a look this week!",
"Hey! \r\nDid you get a chance to look into the runtime degradation?\r\n\r\nThanks",
"Looking now! Sorry for the delay",
"Okey, I can reproduce the degradation! Will try to fix it today",
"I think this PR should fix it: https://github.com... | 1,613 | 1,614 | 1,614 | NONE | null | ## Environment info
<!-- You can run the command `transformers-cli env` and copy-and-paste its output below.
Don't forget to fill out the missing fields in that output! -->
- `transformers` version: 4.2.1 VS 3.4.0
- Platform: Colab (K80 GPU)
- Python version: 3.6.9
- PyTorch version (GPU?): 1.7.0+cu101
- ... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10142/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/huggingface/transformers/issues/10142/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10141 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10141/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10141/comments | https://api.github.com/repos/huggingface/transformers/issues/10141/events | https://github.com/huggingface/transformers/pull/10141 | 806,366,601 | MDExOlB1bGxSZXF1ZXN0NTcxNzk5NzA2 | 10,141 | Add AMP for TF Albert | {
"login": "jplu",
"id": 959590,
"node_id": "MDQ6VXNlcjk1OTU5MA==",
"avatar_url": "https://avatars.githubusercontent.com/u/959590?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jplu",
"html_url": "https://github.com/jplu",
"followers_url": "https://api.github.com/users/jplu/followers",
... | [] | closed | false | null | [] | [
"I can split this PR into two different ones, but the one on AMP will be very short (only two single line to update, see the review above). Are you agree with a that tiny PR? If it is still ok, I will split this one^^",
"It's ok, thanks for showing me the changes!",
"@patrickvonplaten feel free to merge if it l... | 1,613 | 1,613 | 1,613 | CONTRIBUTOR | null | # What does this PR do?
This PR adds the following features to TF Albert:
- AMP compliancy
- Loss computation for TFAlbertForPreTraining
- Cleaning source code | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10141/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/huggingface/transformers/issues/10141/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10141",
"html_url": "https://github.com/huggingface/transformers/pull/10141",
"diff_url": "https://github.com/huggingface/transformers/pull/10141.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10141.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10140 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10140/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10140/comments | https://api.github.com/repos/huggingface/transformers/issues/10140/events | https://github.com/huggingface/transformers/issues/10140 | 806,339,706 | MDU6SXNzdWU4MDYzMzk3MDY= | 10,140 | Direct way to apply different learning rate for different group of parameters in Trainer. | {
"login": "liyucheng09",
"id": 27999909,
"node_id": "MDQ6VXNlcjI3OTk5OTA5",
"avatar_url": "https://avatars.githubusercontent.com/u/27999909?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/liyucheng09",
"html_url": "https://github.com/liyucheng09",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [
"What if instead you derive from `Trainer` and override `create_optimizer_and_scheduler()` and have that function set your different learning rates?"
] | 1,613 | 1,613 | 1,613 | NONE | null | # 🚀 Feature request
<!-- A clear and concise description of the feature proposal.
Please provide a link to the paper and code in case they exist. -->
For now, if I want to specify learning rate to different parameter groups, I need to define an AdamW optimizer in my main function like the following:
```
... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10140/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/huggingface/transformers/issues/10140/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10139 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10139/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10139/comments | https://api.github.com/repos/huggingface/transformers/issues/10139/events | https://github.com/huggingface/transformers/issues/10139 | 806,289,744 | MDU6SXNzdWU4MDYyODk3NDQ= | 10,139 | ValueError: `Checkpoint` was expecting a trackable object (an object derived from `TrackableBase`), got GPT2LMHeadModel | {
"login": "George-Ogden",
"id": 38294960,
"node_id": "MDQ6VXNlcjM4Mjk0OTYw",
"avatar_url": "https://avatars.githubusercontent.com/u/38294960?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/George-Ogden",
"html_url": "https://github.com/George-Ogden",
"followers_url": "https://api.github.c... | [] | closed | false | null | [] | [
"Oh no! I was using the PyTorch model with the TF trainer! I have fixed it now."
] | 1,613 | 1,613 | 1,613 | NONE | null | I'm having this issue, and I think it's my fault, but can someone, please, advise me in case this is a bug rather than a mistake:
```
from transformers import TFTrainer, TFTrainingArguments, GPT2Tokenizer, GPT2LMHeadModel
training_args = TFTrainingArguments(
do_train=True,
output_dir="results",
overwr... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10139/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/huggingface/transformers/issues/10139/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10138 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10138/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10138/comments | https://api.github.com/repos/huggingface/transformers/issues/10138/events | https://github.com/huggingface/transformers/issues/10138 | 806,279,502 | MDU6SXNzdWU4MDYyNzk1MDI= | 10,138 | Back Translation | {
"login": "chaituValKanO",
"id": 8213640,
"node_id": "MDQ6VXNlcjgyMTM2NDA=",
"avatar_url": "https://avatars.githubusercontent.com/u/8213640?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/chaituValKanO",
"html_url": "https://github.com/chaituValKanO",
"followers_url": "https://api.github.... | [] | closed | false | null | [] | [
"Hi @chaituValKanO \r\n\r\nPlease use the [forum](https://discuss.huggingface.co/) to ask such questions. Issues are for bugs, feature requests etc. \r\n\r\nAnd to answer your question you could use the `MarianMT` models for this purpose. Here's a nice blog-post about that\r\nhttps://amitness.com/back-translation/... | 1,613 | 1,613 | 1,613 | NONE | null | # 🚀 Feature request
I want to perform Back translation as a text data augmentation technique using TensorFlow
I want to augment data using translation techniques. I want to perform the below operation English ---> French ---> English. so that resulting english statement might be a new one. These sentences can be p... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10138/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/huggingface/transformers/issues/10138/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10137 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10137/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10137/comments | https://api.github.com/repos/huggingface/transformers/issues/10137/events | https://github.com/huggingface/transformers/issues/10137 | 806,261,648 | MDU6SXNzdWU4MDYyNjE2NDg= | 10,137 | Text to Speech Generalized End-To-End Loss for Speaker Verification, Real Time Voice Cloning | {
"login": "BirgerMoell",
"id": 1704131,
"node_id": "MDQ6VXNlcjE3MDQxMzE=",
"avatar_url": "https://avatars.githubusercontent.com/u/1704131?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/BirgerMoell",
"html_url": "https://github.com/BirgerMoell",
"followers_url": "https://api.github.com/us... | [
{
"id": 1843244711,
"node_id": "MDU6TGFiZWwxODQzMjQ0NzEx",
"url": "https://api.github.com/repos/huggingface/transformers/labels/New%20model",
"name": "New model",
"color": "fbca04",
"default": false,
"description": ""
},
{
"id": 2648621985,
"node_id": "MDU6TGFiZWwyNjQ4NjIxOTg... | open | false | null | [] | [
"@patrickvonplaten This is a suggestion but there are several models available and I think the best first step would be to look into getting a Text-To-Speech model working.\r\n\r\nI explored the Real-Time-Voice-Cloning the other day and noticed it had several issues (since the project is no longer maintained) so it... | 1,613 | 1,650 | null | NONE | null | # 🌟 New model addition
## Model description
Generalized End-To-End Loss for Speaker Verification implements Real time voice cloning, a way to generate a Text-To-Speech model adapted to a certain speaker with a short audio sample. The model implements the following paper.
https://arxiv.org/pdf/1806.04558.pdf and t... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10137/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/huggingface/transformers/issues/10137/timeline | null | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10136 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10136/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10136/comments | https://api.github.com/repos/huggingface/transformers/issues/10136/events | https://github.com/huggingface/transformers/pull/10136 | 806,208,572 | MDExOlB1bGxSZXF1ZXN0NTcxNjc0NTgz | 10,136 | [WIP][examples/seq2seq] move old s2s scripts to legacy | {
"login": "patil-suraj",
"id": 27137566,
"node_id": "MDQ6VXNlcjI3MTM3NTY2",
"avatar_url": "https://avatars.githubusercontent.com/u/27137566?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/patil-suraj",
"html_url": "https://github.com/patil-suraj",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [
"Thanks a lot Stas and Sylvain :)"
] | 1,613 | 1,613 | 1,613 | MEMBER | null | # What does this PR do?
Move the `finetune_trainer.py` and related utils, tests, bash scripts to `examples/legacy/seq2seq` | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10136/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/huggingface/transformers/issues/10136/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10136",
"html_url": "https://github.com/huggingface/transformers/pull/10136",
"diff_url": "https://github.com/huggingface/transformers/pull/10136.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10136.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10135 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10135/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10135/comments | https://api.github.com/repos/huggingface/transformers/issues/10135/events | https://github.com/huggingface/transformers/issues/10135 | 806,190,544 | MDU6SXNzdWU4MDYxOTA1NDQ= | 10,135 | Adding end-to-end retriever training to RAG with RAY implementation. | {
"login": "shamanez",
"id": 16892570,
"node_id": "MDQ6VXNlcjE2ODkyNTcw",
"avatar_url": "https://avatars.githubusercontent.com/u/16892570?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/shamanez",
"html_url": "https://github.com/shamanez",
"followers_url": "https://api.github.com/users/sha... | [] | closed | false | null | [] | [
"This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/master/CONTRIBUTING.md)... | 1,613 | 1,619 | 1,619 | CONTRIBUTOR | null | # 🚀 Feature request
Use of RAY to run separate processes for retrieve document indexes, training the system, and re-initialize the indexes with an updated context encoder.
## Motivation
Recent [papers](https://arxiv.org/abs/2101.00408) have shown that fine-tuning the entire retriever gives huge gains for QA t... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10135/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/huggingface/transformers/issues/10135/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10134 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10134/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10134/comments | https://api.github.com/repos/huggingface/transformers/issues/10134/events | https://github.com/huggingface/transformers/issues/10134 | 806,188,442 | MDU6SXNzdWU4MDYxODg0NDI= | 10,134 | cant install from source | {
"login": "prathameshk",
"id": 4078857,
"node_id": "MDQ6VXNlcjQwNzg4NTc=",
"avatar_url": "https://avatars.githubusercontent.com/u/4078857?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/prathameshk",
"html_url": "https://github.com/prathameshk",
"followers_url": "https://api.github.com/us... | [] | closed | false | null | [] | [
"Hi,\r\n\r\nmaybe `pip` is not connected to a proper Python version (f-strings work in >= 3.6). To make sure that the right version gets called, you can execute the following command:\r\n```\r\npython -m pip install git+https://github.com/huggingface/transformers\r\n```\r\n\r\nThis is why it's a good idea to use a ... | 1,613 | 1,613 | 1,613 | NONE | null | ## Environment info
transformers-cli env
Copy-and-paste the text below in your GitHub issue and FILL OUT the two last points.
- `transformers` version: 3.3.1
- Platform: Linux-4.19.0-13-cloud-amd64-x86_64-with-debian-10.7
- Python version: 3.7.8
- PyTorch version (GPU?): 1.7.1 (True)
- Tensorflow version (GP... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10134/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/huggingface/transformers/issues/10134/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10133 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10133/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10133/comments | https://api.github.com/repos/huggingface/transformers/issues/10133/events | https://github.com/huggingface/transformers/pull/10133 | 806,171,948 | MDExOlB1bGxSZXF1ZXN0NTcxNjQ1ODE0 | 10,133 | [examples/run_s2s] remove task_specific_params and update rouge computation | {
"login": "patil-suraj",
"id": 27137566,
"node_id": "MDQ6VXNlcjI3MTM3NTY2",
"avatar_url": "https://avatars.githubusercontent.com/u/27137566?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/patil-suraj",
"html_url": "https://github.com/patil-suraj",
"followers_url": "https://api.github.com/... | [] | closed | false | null | [] | [
"**Context:**\r\n\r\nHere some context on the `task_specific_params` config param. In the beginning, we had T5 as the only model that was used for both the translation and summarization pipeline. The problem was that we had **one** model that we used as a default for both pipelines. At that time @thomwolf and I tho... | 1,613 | 1,613 | 1,613 | MEMBER | null | # What does this PR do?
- correctly handle `task_specific_params` and `prefix`
The current script tries to access the `prefix` from `config.task_specific_params.prefix`, which is always going to be `None` as `task_specific_params` is a nested `dict` with each key being a task name. This PR retrieves the `task_spe... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10133/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/huggingface/transformers/issues/10133/timeline | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/10133",
"html_url": "https://github.com/huggingface/transformers/pull/10133",
"diff_url": "https://github.com/huggingface/transformers/pull/10133.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/10133.patch",
"merged_at... |
https://api.github.com/repos/huggingface/transformers/issues/10132 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10132/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10132/comments | https://api.github.com/repos/huggingface/transformers/issues/10132/events | https://github.com/huggingface/transformers/issues/10132 | 806,147,244 | MDU6SXNzdWU4MDYxNDcyNDQ= | 10,132 | Where the helsinki models downloaded to? when using the pretrained models | {
"login": "vishnu3741",
"id": 37154661,
"node_id": "MDQ6VXNlcjM3MTU0NjYx",
"avatar_url": "https://avatars.githubusercontent.com/u/37154661?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/vishnu3741",
"html_url": "https://github.com/vishnu3741",
"followers_url": "https://api.github.com/use... | [] | closed | false | null | [] | [
"You can also just clone the repo: \r\n\r\n```\r\ngit clone https://huggingface.co/Helsinki-NLP/opus-mt-es-en\r\n```\r\n\r\nand then load the model and tokenizer locally from the cloned repo:\r\n\r\n```python\r\nmodel = MarianMTModel.from_pretrained(\"/path/to/cloned/repo\")\r\ntokenizer = MarianTokenizer.from_pret... | 1,613 | 1,613 | 1,613 | NONE | null | src_text=['No, los préstamos existentes continuarán por debajo de la tasa de referencia existente.']
model_name='Helsinki-NLP/opus-mt-es-en'
tokenizer=MarianTokenizer.from_pretrained(model_name)
model=MarianMTModel.from_pretrained(model_name)
translated=model.generate(**tokenizer.prepare_seq2seq_batch(src_text, r... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10132/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/huggingface/transformers/issues/10132/timeline | completed | null | null |
https://api.github.com/repos/huggingface/transformers/issues/10131 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/10131/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/10131/comments | https://api.github.com/repos/huggingface/transformers/issues/10131/events | https://github.com/huggingface/transformers/issues/10131 | 806,104,611 | MDU6SXNzdWU4MDYxMDQ2MTE= | 10,131 | Trainer Evaluates at every step | {
"login": "Megh-Thakkar",
"id": 22455368,
"node_id": "MDQ6VXNlcjIyNDU1MzY4",
"avatar_url": "https://avatars.githubusercontent.com/u/22455368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Megh-Thakkar",
"html_url": "https://github.com/Megh-Thakkar",
"followers_url": "https://api.github.c... | [] | closed | false | {
"login": "sgugger",
"id": 35901082,
"node_id": "MDQ6VXNlcjM1OTAxMDgy",
"avatar_url": "https://avatars.githubusercontent.com/u/35901082?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sgugger",
"html_url": "https://github.com/sgugger",
"followers_url": "https://api.github.com/users/sgugge... | [
{
"login": "sgugger",
"id": 35901082,
"node_id": "MDQ6VXNlcjM1OTAxMDgy",
"avatar_url": "https://avatars.githubusercontent.com/u/35901082?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/sgugger",
"html_url": "https://github.com/sgugger",
"followers_url": "https://api.git... | [
"It's hard to know without seeing your code. This combination of arguments should evaluate every 6 steps.\r\nAlso how do you know it's evaluation every step instead of every 6 steps?",
"Hi, thank you for the reply. I have actually overridden the 'evaluate' function of the trainer, and have certain print statement... | 1,613 | 1,619 | 1,619 | NONE | null | Hi, thanks for the amazing and easy to use library. While using the Trainer with Training Arguments, the trainer is evaluating at every step, instead of eval_steps.
Version: 4.3.0
```python
training_args = TrainingArguments(output_dir='outputs', per_device_train_batch_size=1,per_device_eval_batch_size=2,
... | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/10131/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/huggingface/transformers/issues/10131/timeline | completed | null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.