some cleanup

microgopt.go

@@ -104,7 +104,7 @@ func Run(docs []string) {
 		n := min(blockSize, len(tokens)-1)
 
 		// Forward the token sequence through the model, building up the computation graph all the way to the loss
-		keys, values := mkDeepSlice(nLayer), mkDeepSlice(nLayer)
+		keys, values := mkDeepSlice(), mkDeepSlice()
 		losses := []*value{}
 		for posId := range n {
 			tokenId, targetId := tokens[posId], tokens[posId+1]
@@ -138,7 +138,7 @@ func Run(docs []string) {
 	temperature := 0.5 // in (0, 1], control the "creativity" of generated text, low to high
 	fmt.Println("\n--- inference (new, hallucinated names) ---")
 	for sampleIdx := range 20 {
-		keys, values := mkDeepSlice(nLayer), mkDeepSlice(nLayer)
+		keys, values := mkDeepSlice(), mkDeepSlice()
 		tokenId := BOS
 		sample := []rune{}
 		for posId := range blockSize {
@@ -390,7 +390,7 @@ func (v *value) Backward() {
 	}
 }
 
-func mkDeepSlice(size int) [][][]*value {
+func mkDeepSlice() [][][]*value {
 	a := make([][][]*value, 1, 10)
 	a[0] = make([][]*value, 0, 10)
 	return a

readme.md

@@ -13,3 +13,4 @@ Differences between the Go and the Python, as well as notes more generally:
 * The Value struct has actual tests confirming the backward propagation logic.
 * When writing the Value struct and its methods, I accidentally swapped the order of the values in the `localGrads` slice in `Mul` and tore my hair out trying to figure out where the bug was. When I broke down and asked copilot to "compare these two implementations and tell me how they differ," it managed to find the error -- but also reported three non-existent differences and told me that `slices.Backward()` doesn't exist.
 * Initial pass translating the linear algebra functions has me worried that all those value structs aren't going to be very fast...
+* Had to implement weighted random choice. <https://cybernetist.com/2019/01/24/random-weighted-draws-in-go/> made that relatively straightforward; it's a neat algorithm.