David Thrower

Kind of issue: kind/enhancement

Kind of issue: kind/enhancement R and D

Try--dropout-embedding-with-gpt-tokenizer-best-run

1

Kind of issue: kind/enhancement; R and D Hybridize Alex's proof of concept for Droupout(0.75) -> Embedding(15 dimensions) with the parameters in [f2fdcf708269fc9c9fd29ababd7b93cdc6f8f834](https://github.com/david-thrower/cerebros-core-algorithm-alpha/commit/f2fdcf708269fc9c9fd29ababd7b93cdc6f8f834) Suggested Labels (If you don't know, that's ok):...

Kind of issue: The botteck on the tandem embeddings may be that the embedding converges to an optima well before dense layers do. Consequently, the embedding gradients will zero out....

try-conv-1d-skip-connection-junctions

1

Kind of issue: Enhancements Suggested Labels (If you don't know, that's ok): kind/enhancement kind/r-and-d

Kind of issue: Enhancement Additional context: Alex developed a custom enbedding: ```python3 class CustomEmbedding(tf.keras.layers.Layer): def __init__(self, input_dim, output_dim, **kwargs): super(CustomEmbedding, self).__init__(**kwargs) self.input_dim = input_dim self.output_dim = output_dim def build(self, input_shape):...

Kind of issue: Enhancement R and D Additional context merge of work in #77 and #136 Suggested Labels (If you don't know, that's ok): kind/enhancement

Kind of issue:continued R and D for #135 Suggested Labels (If you don't know, that's ok): kind/enhancement

Kind of issue: enhancement Expected behavior A clear description of what you expected to happen. Additional context: Alex suggested to try injecting embedding layers into the NAS. Noting that embedding...

Kind of issue: Enhancement / R and D Additional context: Enhancement to work in branch 131, merge candidate: 9e92a89a7b89c32e4c3244a97a1c323cff55bccc Add any other context about the problem here. On the test...