-
-
Request for report from gh-20933. On latest `main` (d304e9575) on MacOS ARM with Accelerate, Python `3.12`, NumPy `2.0.0rc2`. Some failures are `linalg` things that may not be surprising with a backen…
-
## Issue
Creating new databases, tables and columns take some time because of the data files creation.
## Ideas
### 1) Sparse files
Creating sparse files may dramatically accelerate data f…
-
-
I think GPU can run verilog code much fast that cpu
-
Because models are quite big with GB size, we should think of an efficient way to load models, e.g.
- localPath for experiment, but won't work in a multi nodes cluster
- fluid, which can help to spe…
-
### System Info
```Shell
- `Accelerate` version: 0.29.3
- Platform: Linux-3.10.0-1160.el7.x86_64-x86_64-with-glibc2.35
- `accelerate` bash location: /opt/conda/bin/accelerate
- Python version: 3.1…
-
I am a majaro xfce user, when megasync tries to run it obtains:
```
>>megasync
Setting QT_SCREEN_SCALE_FACTORS= "eDP1=1"
Error unsetting QT_QPA_PLATFORMTHEME vble
fish: 'megasync' terminated …
-
```
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
from accelerate import Accelerator
accelerator = Accelerator()
tokenizer = AutoTokenizer.from_pretrained("glm-large-chinese", trus…
-
Add feature suggestions below:
Current Goals:
- [x] Huber Loss
- [x] LPIPS Loss
- [ ] Modularize Code
- [ ] Add Sampling To Model
- [x] Add Unet & Required Modules
- [ ] Add conditioning …