Xingjian Shi

Yeah, we had the discussion around customized loss function (the purpose of adding the key: https://github.com/autogluon/autogluon/blob/22ea482e52067b4958f897b4628079249bc4f583/multimodal/src/autogluon/multimodal/configs/optimization/adamw.yaml#L37). (Added by @Linuxdex ) May be we should add a registry about loss functions.

We support loading models that are compatible to the Huggingface API. CC @suzhoum @zhiqiangdon We can add additional documents.

I think we allow customized `eval_metric`, but not an easy way of customizing the `validation_metric` (a different concept). The main logic is here https://github.com/autogluon/autogluon/blob/22ea482e52067b4958f897b4628079249bc4f583/multimodal/src/autogluon/multimodal/utils/metric.py#L45-L129 . We need some documentation in...

@zleyk22 Yes, it's a bug and also exists in numpy. ```python import mxnet as mx mx.npx.set_np() a = mx.np.array([1,0]) a.attach_grad() with mx.autograd.record(): b = mx.np.prod(a) b.backward() print(a.grad) ``` Output: ```...

@zleyk22 I think currently no one is looking at the issue. Would you try to solve it? Thanks for pointing out the problem!

@zleyk22 I can think of two possible ways to solve this problem: 1) Use two cumsums - `[\log a_0, \log a_0 + \log a_1, ..., \log a_0 + \log a_1...

I ran into similar issue with the latest master.

Thanks for your interest and thanks for the response from @gaozhihan . The team here is still improving the Earthformer model and will release Earthformer AR next year.

Thanks! This is nice suggestion. We will try to track it.

Checking the CI failure about ``` E TypeError: __init__() got an unexpected keyword argument 'checkpoint_callback' ``` It might be related to these two lines: - https://github.com/ray-project/ray_lightning/blob/13468d753940a5c4557e53c7e662addf6d3e315c/ray_lightning/tests/utils.py#L231 - https://github.com/ray-project/ray_lightning/blob/13468d753940a5c4557e53c7e662addf6d3e315c/ray_lightning/tests/test_tune.py#L34