Vadim Kantorov

Hi! Before doing more flexible learning, I'd like to learn hyperbolic embeddings for leaves of an existing tree (a tree from image region merging agglomerative clustering procedure). Such embeddings that...

Hi @ines-chami! At https://github.com/HazyResearch/HypHC/blob/master/model/hyphc.py#L42 : ```python init_size=1e-3 # in config.py also "init_size": 1e-3 max_scale=1. - 1e-3 # in config.py also "max_scale": 1 - 1e-3 self.scale = nn.Parameter(torch.Tensor([init_size]), requires_grad=True) min_scale =...

Hi Kyle! Thanks a lot for your video explainer, and also for the MIT license for this code example. I'm a machine learning engineer, but have to hack up some...

Discrete diffusion in the code

1

Hi! Thanks a lot for releasing the codes :) Do I understand correctly that the main model is in https://github.com/OpenGVLab/DDPS/blob/main/mmseg_custom/models/decode_heads/segformer_head_unet_fc_head_single_step.py#L128 ? Am I correct that first the discrete class labels...

AFAIK https://github.com/Dao-AILab/flash-attention/ did not have the bandwidth to support custom `attn_bias` (needed for relpos) - I think it's supported for the Triton version there, but I saw reports that it's...

[Feature Request] Array of dicts of tensors structure

3

Storing them in "columnar format" can be more compact in some circumstances (e.g. they are copy-on-write safe in multiprocessing context as they are stored as a very small number of...

plt.figure() throws if not immediately followed by more plotting

3

I need this to set up the report markup (ideally I'd just have `` in html and then fill them up later) and do the plotting later. ``` %% py...

[question] Myelin: attention fusion and FlashAttention

14

Hi! When attention op gets fused in a single op with Myelin, it's not written in trex-tooltip if it's using FlashAttention / proper fusion or not (and if it's using...

As Lexical and Playground is quite powerful, its toolbar already has a lot of actions that Google Docs provides. Google Docs also has a relatively lightweight toolbar, mostly focusing on...

Bug: inline images are not being exported to Markdown at all (verified in playground) and image captions are not exported to Markdown/DOM (instead some other text is used in alt)

4

I went to https://playground.lexical.dev/, inserted an inline image, it gets displayed okay. I then press the button for converting to Markdown. The markdown string contains no info about the image....