Update app.py

app.py

@@ -555,12 +555,12 @@ with gr.Blocks(css=css) as demo:
             - **GemliteUIntXWeightOnly**: uintx gemlite quantization (default to 4 bit only for now)
             - **Int8WeightOnly**: 8-bit weight-only quantization
             - **Int8DynamicActivationInt8Weight**: 8-bit quantization for both weights and activations
-            - **Float8WeightOnly**: float8-bit weight-only quantization
-            - **Float8DynamicActivationFloat8Weight**: float8-bit quantization for both weights and activations
+            - **Float8WeightOnly**: float8 weight-only quantization
+            - **Float8DynamicActivationFloat8Weight**: float8 quantization for both weights and activations
             - **autoquant**: automatic quantization (uses the best quantization method for the model)
 
             ### Group Size
-            - Only applicable for int4_weight_only and int8_weight_only quantization
+            - Only applicable for Int4WeightOnly and GemliteUIntXWeightOnly quantization
             - Default value is 128
             - Affects the granularity of quantization