-
I use the default train.sh as
accelerate launch --mixed_precision='bf16' scripts/train.py
--pretrained_model_name_or_path=$MODEL_NAME
--train_data_dir=$DATASET_NAME
--train_data_meta=$DATASET_META…
-
```
Multi word field names with non-breaking spaces between words get turned into
single word field names.
Replication instructions:
- If you use non-breaking spaces in the text frame and then use "…
-
### Content request
One of the new things that we have been experimenting with in the FRAME team is composite storage item. These are not a new abstraction that FRAME provides, but rather something t…
-
I would really like to have _rank_breakpoint(0) and multi_print(msg) (using color) to help debugging DDP. Maybe next PRs :)
_Originally posted by @tchaton in https://github.com/PyTorchLightning/pyt…
-
```
We should look into how the temporal denoiser is triggered when running in
multi-resolution encoding mode and make sure we do the best thing there. First
we should make sure the denoising code i…
-
In my OpenGL program, I first display a sphere, and then, on the fly, I read in an arbitrary amount of text from a file and then display it on top of the sphere. I never know in advance the (width, h…
-
Hello!
Will the source code be released? If so, is there a specific time
-
First of all I want to congratulate the work was very good.
I am writing a booklet in Portuguese, and would like to change the title box where the ToC is displayed. Changing the text to Portuguese.
-
Hi, Thanks for your great work
I am trying to launch two devices at same time. Instead of using livox hub, are there any ways to launch two devices which are directly connected to laptop ?
I am …
-
Hello,
I have been trying out this project for a while and have recently managed to work through many issues to achieve a desirable output. After many days of training with different seed values, I h…