-
Hi! It's an intriguing work!
I have a question about K400 pre-trained models.
Did you use ImageNet-1K pre-trained models to pre-train K400 datasets for downstream tasks (e.g., Breakfast, COIN, etc) …
-
- https://arxiv.org/abs/2109.12178
- 2021
視覚と言語の事前学習（VLP）は，画像やテキストの入力を必要とする下流のタスクのモデル性能を向上させる．
現在のVLPアプローチは、
(i)モデルアーキテクチャ（特に画像エンベッダー）、
(ii)損失関数、
(iii)マスキングポリシーによって異なります。
画像エンベッダーは、ResNet…
e4exp updated
3 years ago
-
When building models for teacher and student in this [code](https://github.com/facebookresearch/dinov2/blob/main/dinov2/models/__init__.py#L15), the parameter args.arch is used for both student and te…
-
### Links
- Paper : https://arxiv.org/abs/2111.06377
- Github : https://github.com/facebookresearch/mae
### 한 줄 요약
- Self-supervised learning 중 masked image modeling 개념을 적용한 논문으로, NLP 도메인과의 차이점을…
-
### Describe the bug
I'm trying to load an SQLite database that's around 100MB.
Seems like I'm hitting this line when trying to access a table in the db that's bigger than 32MB:
https://github.com/e…
-
Hello! I notice in your code that the model's input remains consistent during training and inference, i.e., paired images `imgs`, paired labels `tgts`, and mask `bool_masked_pos`. During `forward()`, …
-
Hi, thank you for releasing this GitHub repository. I am trying to reproduce the stage 1 training on ImageNet. Could you please share the W&B log or let me know the initial and final loss values for t…
-
### 软件环境
```Markdown
- paddlepaddle:
- paddlepaddle-gpu: 3.0.0b1
- paddlenlp: 3.0.0b1.post20241009
```
### 重复问题
- [X] I have searched the existing issues
### 错误描述
```Markdown
对 Qwen/Qwen2-…
-
AssertionError: Could not infer task type from {'_name': 'av_hubert_pretraining', 'is_s2s': True, 'data': '/checkpoint/bshi/data/lrs3//exp/ls-hubert/tune-modality/all_tsv/', 'label_dir': '/checkpoint/…
-
The goal of this issue is to track new self-supervised methods and to which extent they are implemented in Lightly.
See also our [#papers channel](https://discord.com/channels/752876370337726585/81…