3 ms·
> employs a single-layer vector quantizer (VQ) codec and a single Transformer architecture to fully align I really wish when new models were released that they
by dheera 1y ago
> employs a single-layer vector quantizer (VQ) codec and a single Transformer architecture to fully align
I really wish when new models were released that they would draw a diagram of all the layers and the tensor input and output sizes at each layer, with zoom in/out capabilities if needed using D3.js or whatever visualization framework if needed. Every single layer should be on there with its input and output sizes.
These one-sentence descriptions, and approximate block diagrams with arrows pointing at each other are never enough to understand how something is actually implemented.
- exe34 1y agoSounds like a solid SaaS business plan!
- dr_kiszonka 1y agoThat might be intentional.
- imtringued 1y agoThis already exists in Transformer Lab and ONNX (not recommended for transformers). You can also build a custom version of llama.cpp that writes out the ggml compute graph. What's irritating is that hugging face didn't add it to their GGUF file viewer.
- dheera 1y agoOh, sure, for the well-known models that are already on there. I just wish that new research would always spell it out in full instead of these silly block diagrams labelled with just e.g. "Cross Attention" and not the exact parameters, number of heads, layer sizes, etc. Also some of these diagrams use a + for concatenation and some use it for addition, that's another headache to figure out, having layer sizes would make it clear.