-
Hi, I'm unsure about this piece of code in `MaskedCrossAttention` inside [`gated_cross_attention.py`](https://github.com/dhansmair/flamingo-mini/blob/main/flamingo_mini/gated_cross_attention.py)
```
…
-
If you are submitting a bug report, please fill in the following details and use the tag [bug].
**Describe the bug**
The generations from huggingface model (LlamaForCausalLM) and HookedTransformer…
-
When running on an ARM64 host system running either Linux or macOS, BuildKit will default to building ARM64 container images unless the flags `--platform linux/amd64` are explicitly specified. At the …
-
Random idea for "Culture" category of roadmap:
Introduction:
The purpose of culture, like Civilization (the game), it could be a factor in territory (since we are most likely competing with sever…
-
Hello ! I tried something a bit out of scope but still:
## context
Use conditional compiling in `rid` plugin crate, gated behind a `feature flag`.
## steps to reproduce
Just use a feature …
-
## Issue Report
Licensee and Linguist are used as axioms to allow some linting rules to only apply when the output of those tools match the license or language. Unfortunately, both are written in R…
-
使用经过脚本转换后的huggingface上的mengzi-t5-base模型时报错:
```
RuntimeError: Error(s) in loading state_dict for Model:
size mismatch for embedding.word_embedding.weight: copying a param with shape torch.S…
-
**Paper**: http://openaccess.thecvf.com/content_ICCV_2019/papers/Choi_Looking_to_Relations_for_Future_Trajectory_Forecast_ICCV_2019_paper.pdf
**Summary**: Predict future trajectories of all objects…
-
### Feature request
In [`Transformers 4.36`](https://github.com/huggingface/transformers/releases/tag/v4.36.0), we started adding native support of [torch.nn.functional.scaled_dot_product_attention](…
-