-
Nodes doesnt work when reaching the samplers, i get this error :
Simple one :
"
TypeError: All input tensors need to be on the same GPU, but found some tensors to not be on a GPU:
[(torch.Siz…
-
I am using this command to try inference:
```
python generate.py --load pytorch_model.bin --tokenizer-type HFTokenizer --vocab-file tokenizer.json --text_gen_type interactive --temperature 0.0 --max…
-
### What is your question?
## How to Pass Weights as Parameters in Flower?
I’m trying to use the Flower framework to train a YOLO model in a federated learning setting. I’m having trouble figuring…
-
### Describe the bug
With the 0.7.4 release, augmentations on the GPU are not possible anymore because the output of the input tensor is always explicitly moved to the CPU.
The problem is that `…
-
### 🐛 Describe the bug
Building wheel torch-2.6.0a0+gitb86b534
-- Building version 2.6.0a0+gitb86b534
cmake -GNinja -DBUILD_PYTHON=True -DBUILD_TEST=True -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_…
-
I'm attempting to run the train_DESI.py script in the following way
$ python train_DESI.py -b 256 -l 10 -zmax 0.45 -v
And the following error is raised:
File "/opt/SIE/ML/miniconda3/envs/…
-
I tried to compile single_prefill_with_kv_cache using torch.compile.
```Python
import torch
from flashinfer import single_prefill_with_kv_cache
data_type = torch.bfloat16
QH=64
KH=8
S=1024
…
-
I’m working on some training code that computes the total log probabilities of prediction sequences (i.e. outputs from a language model).
I had previously implemented this using `F.log_softmax` fol…
-
hi, i'm try to run the notebook, but i meet some problems.
it look's like the download model file is not correct, because it's too small：
:
"""
Expands attention_mask from `[bs…