diff --git a/README.md b/README.md
index 806ea6e..9b128e5 100644
--- a/README.md
+++ b/README.md
@@ -37,13 +37,13 @@ Endpoint decodability 示意：上排是直接解码中间 noisy latent，早期
 
 <img src="assets/endpoint_decodability.png" alt="Endpoint decodability" width="760">
 
-不同 NFE 下的生成质量变化：
+Z-Image-Turbo 在不同 NFE 下的生成质量变化：
 
-<img src="assets/tjs_low_nfe_gallery.png" alt="TJS low NFE gallery" width="760">
+<img src="assets/z_image_turbo_low_nfe_gallery.png" alt="Z-Image-Turbo low NFE gallery" width="760">
 
-更长步数范围下的 NFE 渐进对比：
+SDXL 30 步采样设置下的 NFE 渐进对比：
 
-<img src="assets/tjs_nfe_progression.png" alt="TJS NFE progression" width="760">
+<img src="assets/sdxl_30step_nfe_progression.png" alt="SDXL 30-step NFE progression" width="760">
 
 FID 与 NFE 的趋势示意：
 
diff --git a/README_EN.md b/README_EN.md
index 09a4aaf..bfe4eaa 100644
--- a/README_EN.md
+++ b/README_EN.md
@@ -45,13 +45,13 @@ endpoint.
 
 <img src="assets/endpoint_decodability.png" alt="Endpoint decodability" width="760">
 
-Generation quality at different NFE budgets:
+Z-Image-Turbo generation quality at different NFE budgets:
 
-<img src="assets/tjs_low_nfe_gallery.png" alt="TJS low NFE gallery" width="760">
+<img src="assets/z_image_turbo_low_nfe_gallery.png" alt="Z-Image-Turbo low NFE gallery" width="760">
 
-Progression across a wider NFE range:
+SDXL 30-step NFE progression:
 
-<img src="assets/tjs_nfe_progression.png" alt="TJS NFE progression" width="760">
+<img src="assets/sdxl_30step_nfe_progression.png" alt="SDXL 30-step NFE progression" width="760">
 
 FID versus NFE trend:
 
diff --git a/assets/tjs_nfe_progression.png b/assets/sdxl_30step_nfe_progression.png
similarity index 100%
rename from assets/tjs_nfe_progression.png
rename to assets/sdxl_30step_nfe_progression.png
diff --git a/assets/tjs_low_nfe_gallery.png b/assets/z_image_turbo_low_nfe_gallery.png
similarity index 100%
rename from assets/tjs_low_nfe_gallery.png
rename to assets/z_image_turbo_low_nfe_gallery.png