Gordon Lim
Gordon Lim
May I ask what were you last outputs from your training loop? I trained using your weights but got `[781/782] Loss_D: 0.0267 Loss_G: 8.2782 D(x): 0.9833 D(G(z)): 0.0072 / 0.0008`...
Using `ObjectClassMasks = (R/10).astype(np.int32)*256+(G.astype(np.int32));` I got the id 165 for a particular pixel by which `index_ade20k['objectnames'][165]` gives 'bed column'. However according to the accompanying json to this image, the object...
Thank you authors of this package. It is very relevant to my undergraduate research project. I have clean document images scraped from the Internet such as the following which I...
By my understanding, `cleanlab.benchmarking.noise_generation.generate_noise_matrix_from_trace` is used to generate a set percentage of incorrect labels. In the code below, I've tried using a 0.2 noise amount (20% label noise) however the...
Awesome job on the python implementation! I just have one suggestion. I feel a quick way for anyone to know what this project is about would be to provide an...
In examples/paper_replication/runner.py: `:param bool threshold_samples_set_idx: (default 1) Which set of threshold samples to use (based on index)` However in examples/paper_replication/small_dataset_aum.sh, the value of threshold_samples_set_idx was provided as 2.
Dear author, Please allow me to clarify some details regarding the implementation. To run inference on an image using your proposed "MLNN based stacking of holistic & region-based models with...
With reference to this [tutorial ](https://docs.wandb.ai/tutorials/tables)on visualizing predictions, I'm wondering if per-sample logging as in the following block would bottleneck batching since you are still linearly going through each sample:...
Wonderful project! I want to pitch an idea for an Arcade mode in which you can choose git commands you want more practice with beyond the vanilla levels. Example: Say...
Why do we need the following in the Fine-tuning LayoutLMv2ForSequenceClassification on RVL-CDIP notebook? ``` # we need to define custom features features = Features({ 'image': Array3D(dtype="int64", shape=(3, 224, 224)), 'input_ids':...