-
# Summary
This can have large performance impact in real Attention modules.
The most common pattern (derived from nano-gpt)
```Python
import torch
import torch.nn as nn
import torch.nn.funct…
-
Provided code calculates matrix product of q and k.
https://github.com/YuchuanTian/DiJiang/blob/main/modeling/pythia-2.8B-dijiang/modeling_gpt_neox_dijiang.py#L286
That means it has computational …
-
"E:\software\Umi\GPT-SoVITS-beta0217fix3\runtime\python.exe" GPT_SoVITS/s2_train.py --config "E:\software\Umi\GPT-SoVITS-beta0217fix3\TEMP/tmp_s2.json"
INFO:zhouwensha:{'train': {'log_interval': 100,…
-
## Application
- [x] Highlighted menu item icon is wrong color
- [x] history search not working
- [ ] BlankSlate flickers, image size?
- [ ] Documents upload page doesn't update No of Chunks aft…
-
[dali.zip](https://github.com/user-attachments/files/17500253/dali.zip) (279000 steps)
This is a synthesized 16kHz Chinese audio, where noise is consistently present at specific frequencies.
![i…
-
## 🐛 Bug
### To Reproduce
```python
import thunder
from thunder.tests.litgpt_model import GPT
import torch
device = torch.device("cuda")
with device:
model = GPT.from_name("llama2-li…
-
Forge worked just fine yesterday. Today I installed requirement.txt for the mixlab node for ComfyUI and ALL my Forge-"installations" gpt messed up 💩 How is that even possible.
Both my venv-one and my…
-
开启SoVITS训练时报错：
```
D:\python\lib\site-packages\torch\functional.py:650: UserWarning: stft with return_complex=False is deprecated. In a future pytorch release, stft will return complex tensors for…
-
Does vLLM support 8 bit quantization? We need to use vLLM with large context window (>1K tokens). We tried AWQ but the generation quality is not good. Any pointer will be greatly appreciated.
-
Is there any example code to do this? Should I generate new BlockMask everytime?
Thanks!
------------------------------
Essentially, I have problem of slicing BlockMask. For exmaple, if we have…