configure github pages deployer

.github/workflows/deploy.yaml

@@ -0,0 +1,47 @@
+name: Build and Deploy
+on:
+  push:
+    branches:
+      - "master"
+  workflow_dispatch:
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
+      - uses: actions/setup-node@v3
+        with:
+          node-version: "20"
+      - name: build
+        run: |
+          cd web
+          npm install
+          npm run build
+      - name: archive web build
+        uses: actions/upload-artifact@v3
+        with:
+          name: github-pages
+          path: ./web/dist
+
+  deploy:
+    needs: build
+
+    # Grant GITHUB_TOKEN the permissions required to make a Pages deployment
+    permissions:
+      pages: write # to deploy to Pages
+      id-token: write # to verify the deployment originates from an appropriate source
+
+    # Deploy to the github-pages environment
+    environment:
+      name: github-pages
+      url: ${{ steps.deployment.outputs.page_url }}
+
+    # Specify runner + deployment step
+    runs-on: ubuntu-latest
+    steps:
+      - name: Deploy to GitHub Pages
+        id: deployment
+        uses: actions/deploy-pages@v2

chessbot/src/evaluator/nneteval.rs

@@ -101,6 +101,12 @@ impl NnetEval {
 
 impl Evaluator for NnetEval {
     fn evaluate(&self, board: &chess::Board) -> anyhow::Result<f32> {
+        match board.status() {
+            chess::BoardStatus::Ongoing => {}
+            chess::BoardStatus::Checkmate => return Ok(-10000.0),
+            chess::BoardStatus::Stalemate => return Ok(-9000.0),
+        }
+
         let tensor =
             NnetEval::board_to_input(board).context("failed to create tensor for board")?;
         let output = self

modeltraining/training.py

@@ -52,7 +52,8 @@ def __getitem__(self, idx):
     assert(len(bitvec) == lib.game.tensor_packed_len)
     bin = np.frombuffer(bitvec, dtype=np.uint8)
     bin = np.unpackbits(bin, axis=0).astype(np.single)[0:lib.game.tensor_len]
-    score =  math.copysign(math.log2(abs(score / 10.0) + 1), score) # the more extreme the win the less we weight it, we just want to bias towards winning.
+    score = score / 100.0
+    # score =  math.copysign(math.log2(abs(score / 10.0) + 1), score) # the more extreme the win the less we weight it, we just want to bias towards winning.
     return {
       'binary': bin,
       'eval': np.array([score]).astype(np.single)