auto dtype selection

Author: xenova

src/models.js

@@ -182,6 +182,17 @@ async function getSession(pretrained_model_name_or_path, fileName, options) {
         }
     }
 
+    if (dtype === 'auto') {
+        const config_dtype = custom_config.dtype?.[fileName];
+        if (config_dtype === 'auto') {
+            // Choose default dtype based on device, falling back to fp32
+            dtype = DEFAULT_DEVICE_DTYPE_MAPPING[selectedDevice] ?? DATA_TYPES.fp32;
+        } else {
+            // Defined by the custom config, and is not "auto"
+            dtype = config_dtype;
+        }
+    }
+
     const selectedDtype = /** @type {import("./utils/dtypes.js").DataType} */(dtype);
 
     if (!DEFAULT_DTYPE_SUFFIX_MAPPING.hasOwnProperty(selectedDtype)) {

src/utils/dtypes.js

@@ -31,6 +31,7 @@ export const isWebGpuFp16Supported = (function () {
 })();
 
 export const DATA_TYPES = Object.freeze({
+    auto: 'auto', // Auto-detect based on environment
     fp32: 'fp32',
     fp16: 'fp16',
     q8: 'q8',
@@ -47,7 +48,7 @@ export const DEFAULT_DEVICE_DTYPE_MAPPING = Object.freeze({
     [DEVICE_TYPES.wasm]: DATA_TYPES.q8,
 });
 
-/** @type {Record<DataType, string>} */
+/** @type {Record<Exclude<DataType, "auto">, string>} */
 export const DEFAULT_DTYPE_SUFFIX_MAPPING = Object.freeze({
     [DATA_TYPES.fp32]: '',
     [DATA_TYPES.fp16]: '_fp16',