-
Currently, the `Bidirectional` class uses the same positional and keywords arguments for its two children. As such, non-recurrent weights and biases will generally be shared between the two networks, …
-
**OPT:**
- [x] a specific positional embedding for opt
- [x] a bias support in linear layer
- [ ] some masks inside attention layer
**Bloom:**
- [ ] a permute kernel
- [ ] some changes inside …
-
Hi, Thank you for your sharing. When I looked through this repo, I found there is no relative positional embedding or relative bias, which have been mentioned in equations?.
-
In your code file `ViTAE-VSA\Image-Classification\vitaev2_vsa\NormalCell.py` L130:
`self.pos = nn.Conv2d(dim, dim, window_size//2*2+1, 1, window_size//2, groups=dim, bias=True)`
your `window_size`i…
-
It is expected that bias vector can be initialized not only with zeros but the following code fails.
```python
import jax, jax.numpy as jnp
import flax.linen as nn
layer = nn.Conv(features=32,…
-
File "/home/zpf/anaconda3/envs/humannorm/lib/python3.8/site-packages/transformers/models/clip/modeling_clip.py", line 185, in forward
patch_embeds = self.patch_embedding(pixel_values.to(dtype=tar…
-
- [x] -f
- [x] -r
- [ ] -b
- [x] -o
- [x] -l
- [ ] -G
- [x] --one-based
- [x] --num-threads
- [x] --dry-run
- [x] --verbose
- [ ] --strand
- [ ] --ko-strain
- [ ] --replicate
- [x] --min-mutation-fre…
soh-i updated
10 years ago
-
Cause the reason of internet,I have to download the pretrained parameters in local an load it. But it does not match.
code:
import open_clip
import torch
model, _, preprocess = open_clip.create…
-
run: ./libtorch-yolov5 --source ../images/bus.jpg --weights ../weights/yolov5s.torchscript.pt --gpu --view-img
error:
terminate called after throwing an instance of 'torch::jit::ErrorReport'
wha…
-
### 🐛 Describe the bug
When converting a model to ExecuTorch using `aten_dialect = export(model, (dummy_input_text, dummy_input_label, dummy_input_midi, dummy_input_duration_phn, dummy_input_duration…