docs: tutorials + docstrings#63
Conversation
|
@MichielStock, I'd be happy if you could take a look; otherwise, we can merge it, and you can work from there. As you prefer. |
| # and manipulates information using high-dimensional vectors called **hypervectors**. These | ||
| # vectors typically have thousands of dimensions (often 1,000-10,000), making them | ||
| # "hyperdimensional." | ||
| # vectors typically have thousands of dimensions (often 1.000-10.000), making them |
There was a problem hiding this comment.
should be 1000-10,000
|
|
||
| # Alternatively, one can create a hypervector directly from a `Vector{T}` where `{T}` is an | ||
| # appropiate data type, e.g. integers for BipolarHV: | ||
| H = BinaryHV |
There was a problem hiding this comment.
Is correct, but is this not quite opaque for a tutorial?
There was a problem hiding this comment.
This is more to exemplify, not to be too verbose. I tested it without that line and it feels like you are reading "BinaryHV" along all the tutorial (when in reality its not SO important for an introductory tutorial)
| h₁ = BipolarHV(; D = 8) | ||
| h₂ = BipolarHV(; D = 8) | ||
| h₃ = BipolarHV(; D = 8); | ||
| # ...and there are a lot of them -- this is where the "hyper" comes from: |
There was a problem hiding this comment.
chances are astronomically low that two randomly drawn vectors are the same
| # We can now encode a single flower: bind each feature value to its key and bundle the pairs with | ||
| # [`hashtable`](@ref). | ||
|
|
||
| encode(features::AbstractVector{Float64}) = hashtable(cm2hv.(features), H_features) |
There was a problem hiding this comment.
maybe merge the stateful encoders #64 and then use that function. I can also update this automatically with the agent
There was a problem hiding this comment.
Makes sense, I'll do the merging tomorrow
There was a problem hiding this comment.
Unsure whether to continue this, maybe too gimicky ?
There was a problem hiding this comment.
Having more representations for the hypervectors? I think what we have now is good enough
2 participants
This PR improves on the existing tutorials and covers the package's functionality as much as possible.
Closes #32
Closes #36
Closes #38