Songlin Li
Songlin Li
Hey folks had a similar issue, I'm running with offline inference mode. I was able to clear the resource with `ray stop` But when I try to reload the resource...
But how to run the MeamSum model on the testing yelp dataset? Could you please share some insights about it? I suppose we should use the test mode in train_sum.py?
Hi, I have the same questions. Have you figured out how? Any help is appreciated.
> A new bug was introduced in 0.4.2, but fixed in #4737. Please try with that PR or as a workaround you can also install `tensorizer`. > > This should...
If you are using a huggingface ckpt and this is a lm model this might not be possible. The reason behind is that the embedding layer IS the lm_head. You...
Hi @Muennighoff, amazing work! I have a similar confusing as @yonxie. I can see [here](https://github.com/ContextualAI/gritlm/blob/47b7fe6c7109ba46b82b68c37d32aa9a8bf010c5/gritlm/gritlm.py#L209) that you did a final pooling. You mentioned that "The last hidden state is produced...
这里的xxxx是你的main node地址吗