ggml webgpu: support for backend sampling

ggml/src/ggml-webgpu/ggml-webgpu.cpp

@@ -2043,6 +2043,8 @@ static std::optional<webgpu_command> ggml_webgpu_encode_node(webgpu_context ctx,
             return ggml_webgpu_unary_op(ctx, src0, node);
         case GGML_OP_FILL:
             return ggml_webgpu_unary_op(ctx, src0, node);
+        case GGML_OP_LOG:
+            return ggml_webgpu_unary_op(ctx, src0, node);
         case GGML_OP_PAD:
             return ggml_webgpu_pad(ctx, src0, node);
         case GGML_OP_ARGMAX:
@@ -3069,6 +3071,9 @@ static bool ggml_backend_webgpu_device_supports_op(ggml_backend_dev_t dev, const
         case GGML_OP_FILL:
             supports_op = op->type == GGML_TYPE_F32 && src0->type == GGML_TYPE_F32;
             break;
+        case GGML_OP_LOG:
+            supports_op = (op->type == GGML_TYPE_F32 || op->type == GGML_TYPE_F16) && (src0->type == op->type);
+            break;
         case GGML_OP_PAD:
             supports_op = op->type == GGML_TYPE_F32 && src0->type == GGML_TYPE_F32;
             break;

ggml/src/ggml-webgpu/wgsl-shaders/unary.wgsl

@@ -102,6 +102,9 @@ fn main(@builtin(global_invocation_id) gid: vec3<u32>) {
 #ifdef EXP
     let res = exp(src[params.offset_src + src_idx]);
 #endif
+#ifdef LOG
+    let res = log(src[params.offset_src + src_idx]);
+#endif
 #ifdef CLAMP
     let res = clamp(src[params.offset_src + src_idx], TYPE(params.clamp_min), TYPE(params.clamp_max));
 #endif