-
I'm curious, is there good evidence in paper that TreeGen is better than regular transformers?
I've noticed that other papers and my own experiments, if I increase the data set size then the extra …
-
When attempting to generate the third merkle tree interval using v1.0.5:
```
phiz@animal:~/treegen$ ./treegen-linux-amd64 -i 3
2022/12/08 19:06:25 Beacon node is configured for Mainnet.
2022/1…
-
* [x] Write an initial implementation
* [x] Get the result similar to the paper
* [ ] Make training stable (the loss becomes NaN after 40 epochs with Hearthstone dataset)
-
As per our discussion in the hangouts chat, here's a place we can discuss ideas for a Calder DSL, as well as pain points with the way we currently use the library.
### Current Calder Code
```js…
-
Hi,
I was wondering, why was TreeGen not trained with the standard warm up scheduler (or RAdam)? It seems to be an essential piece for training most NLP transfromers so I was curious is this was tr…
-
Hi Authors of TreeGen!
First thanks again for sharing your code and for the very interesting work!
I was reading your work and noticed that you predict the grammar rules all in one go with a cla…
-
Could you wrap the treegen in a native time function to let the user know how long the generation took? Bonus points if it gave a report:
```
Calculating collateral rewards took 20 seconds
Smooth…
-
TreeGen learns by predicting the rules from the target program. Thus, I assume you make the target program into an AST and in that process you get a sequence (that is padded) indicating which rule was…
-
**Description:**
$subject.
We can have the two nodes with an is-a relationship.
e.g. Renaming `BracedExpressionNode` to `ParenthesisedExpressionNode`
1. Extend the old node with the new one. `Br…
-
Hi Authors,
My understaning is that TreeGen learns by predicting the rules from the target program and then computing the CE loss with the ground truth grammar rules (as a sequence). Thus, I assume…