-
Hi,
Great notebook! Just wanted to mention that there is no need to pass the `generator` in the constructor of the `EncoderDecoder` class. It makes it a bit confusing as looking at the model descri…
-
One of the features in the getting started, and of notebooks in general I thought, it using the previous cell's output.
I naively tried with a random number generator in one cell, then a second cel…
-
```model_id = "black-forest-labs/FLUX.1-schnell" # you can also use `black-forest-labs/FLUX.1-dev
pipe = FluxPipeline.from_pretrained(
"black-forest-labs/FLUX.1-schnell", torch_dtype=torch.bf…
-
Since google colab transitioned to ubuntu 22, the setup code no longer works. Below is the correct setup code so that the notebook runs as expected.
```
# install all tools and dependencies
import…
-
When running the `ORPO Unsloth Example.ipynb` notebook, I encountered an error during the execution of `orpo_trainer.train()`. The error occurs consistently across different GPU types and persists eve…
-
Battle-Plan
- phase1: the Barcode,
- [x] Proof-of-Concept Notebook that we can go till there : we can generate a full GS1_128 Barcode on Jupyterlite
- [x] workaround till 'true' GS1_128: legend…
-
Found this interesting library for Kotlin/Java code generation from OpenApi schemas. While we already have support, there are still some issues with the code generation (https://github.com/Kotlin/data…
-
I followed [this notebook](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/SAM/Fine_tune_SAM_(segment_anything)_on_a_custom_dataset.ipynb) for finetuning the SAM on custom data, and s…
unmo updated
3 months ago
-
Hi, thanks for the amazing work. I understood that SAM2 is tracking single or multiple objects in a video. Is there any option to predict all the objects of a video without any prompting? If yes, how …
-
Hi, I am working on segmenting car bodies in images using the Meta SAM model. I am facing a significant difference in performance between the UI demo on the official website and the code provided on t…