💬 Commit message: Update 2026-02-15 05:04:35, 10 files, 374 lines

📁 Files changed: 10
📝 Lines changed: 374

  • convert.sh
  • index-BQdjF_w0.css
  • index-CKJOpgtQ.js
  • index-DEHUU-Zz.js
  • index-Ljwp9hgM.css
  • index.html
  • App.vue
  • DownloadsPanel.vue
  • ModelCard.vue
  • app.ts

scripts/convert.sh

@@ -0,0 +1,34 @@
+#!/usr/bin/env bash
+# Convert a safetensors model to q8_0 GGUF format
+# Usage: ./scripts/convert.sh <input.safetensors>
+set -euo pipefail
+
+if [[ $# -lt 1 ]]; then
+    echo "Usage: $0 <input.safetensors> [quantization]"
+    echo "  quantization: f32, f16, q4_0, q5_0, q8_0 (default: q8_0)"
+    exit 1
+fi
+
+INPUT="$1"
+QUANT="${2:-q8_0}"
+
+if [[ ! -f "$INPUT" ]]; then
+    echo "Error: File not found: $INPUT"
+    exit 1
+fi
+
+# Derive output filename: model.safetensors -> model-q8_0.gguf
+BASENAME=$(basename "$INPUT" .safetensors)
+DIRNAME=$(dirname "$INPUT")
+OUTPUT="${DIRNAME}/${BASENAME}-${QUANT}.gguf"
+
+echo "==> Converting: $INPUT"
+echo "    Output: $OUTPUT"
+echo "    Quantization: $QUANT"
+echo ""
+
+sd -M convert -m "$INPUT" -o "$OUTPUT" --type "$QUANT"
+
+echo ""
+echo "==> Done: $OUTPUT"
+ls -lh "$OUTPUT"