-
Hello dear HuggingFace team!
According to the original paper, data2vec is not an actual model but more of a self-distilling training strategy. It takes an encoder model as backbone (RoBERTa for text,…
-
Comment
-
Hi man. Thanks for your work on Data2Vec. I just reviewed your implementation and noticed that you use the actual image as targets but I think the targets must be discrete as in BEiT. Although I didn'…
-
BEIT模型在paddle混合精度训练下 前向出现nan。
-
## Environment info
- `transformers` version: NA
- Platform: Windows 10 (64 bit)
- Python version: 3.6 / 3.10
- PyTorch version (GPU?): NA
- Tensorflow version (GPU?): NA
- Using GPU in scri…
-
First of all, thank you for the great work @AryanShekarlaban and @kabouzeid!
quick question: I have less experience with training big models like Transformers. I see that there are many frameworks …
-
## Job Summary
We are looking for a Java Senior Software Engineer that will be in charge of building and evolving the search domain microservices platform which powers more than 80 millions searche…
-
I've ported RepLKNet model to TF (will be published soon).
But during tests i've found that one particular checkpoint - RepLKNet-31L for imagenet21k - produces wrong predictions.
For image https:/…
-
### System Info
```shell
- `transformers` version: 4.19.2
- Platform: macOS-12.3.1-arm64-arm-64bit
- Python version: 3.9.9
- Huggingface_hub version: 0.4.0
- PyTorch version (GPU?): 1.11.0 (False…
-
Although pretraining these models requires a lot of hardware resources and is almost impossible for an individual like me to do, there is the possibility to port the weights from HuggingFace models th…