fix: 512px VAE decode OOM - effective half-size tiling + unconditional retry

docs/backend.md

@@ -188,8 +188,9 @@ weights, compute buffers and caches must
 still fit the runner's capacity checks. Offloading weights does not guarantee
 that every resolution or frame count will fit, and auto-fit does not change a
 component to CPU computation solely because its full weights exceed VRAM.
-If a VAE decode fails, auto-fit retries with spatial tiling; supported video
-decoders try temporal tiling first and can then add spatial tiling.
+If a VAE decode fails, decoding retries with spatial tiling even when `--auto-fit`
+is off; supported video decoders try temporal tiling first and can then add
+spatial tiling. Spatial retries use half-size tiles along each latent dimension.
 
 ## Modules
 

src/core/backend_fit.cpp

@@ -390,6 +390,8 @@ namespace sd::backend_fit {
             retry_mode                    = tiling_params.enabled ? "spatial+temporal" : "temporal";
         } else if (!tiling_params.enabled) {
             tiling_params.enabled = true;
+            tiling_params.rel_size_x = 0.5f;
+            tiling_params.rel_size_y = 0.5f;
             if (tiling_params.tile_size_x <= 0) {
                 tiling_params.tile_size_x = 256;
             }
@@ -401,7 +403,7 @@ namespace sd::backend_fit {
             return false;
         }
 
-        LOG_WARN("auto-fit: VAE decode failed (likely out of memory); retrying with %s tiling",
+        LOG_WARN("VAE decode failed (likely out of memory); retrying with %s tiling",
                  retry_mode);
         return true;
     }

src/pipeline/diffusion_engine.cpp

@@ -2621,7 +2621,6 @@ sd::Tensor<float> StableDiffusionGGML::decode_first_stage(const sd::Tensor<float
     auto decoded                      = first_stage_model->decode(n_threads, latents, vae_tiling_params, decode_video, circular_x, circular_y);
     const bool prefer_temporal_tiling = decode_video && first_stage_model->can_temporal_tile_decode();
     while (decoded.empty() &&
-           auto_fit_enabled &&
            sd::backend_fit::prepare_vae_decode_retry_tiling(vae_tiling_params, prefer_temporal_tiling)) {
         decoded = first_stage_model->decode(n_threads, latents, vae_tiling_params, decode_video, circular_x, circular_y);
     }