rephrase description

app.py

@@ -95,11 +95,13 @@ def quantize_model(
             calibration_dataset = None
 
         is_int8 = dtype == "int8"
-        if library_name == "diffusers":
-            quant_method = "hybrid"
-        elif not is_int8 and calibration_dataset is not None:
+        # if library_name == "diffusers":
+        # quant_method = "hybrid"
+        if not is_int8 and calibration_dataset is not None:
             quant_method = "awq"
         else:
+            if calibration_dataset is not None:
+                print("Default quantization was selected, calibration dataset won't be used")
             quant_method = "default"
 
         quantization_config = OVWeightQuantizationConfig(
@@ -112,7 +114,7 @@ def quantize_model(
 
         api = HfApi(token=oauth_token.token)
         if api.repo_exists(new_repo_id) and not overwritte:
-            return f"Model {new_repo_id} already exist, please set overwritte=True to push on an existing repo"
+            return f"Model {new_repo_id} already exist, please set overwritte=True to push on an existing repository"
 
         with TemporaryDirectory() as d:
             folder = os.path.join(d, repo_folder_name(repo_id=model_id, repo_type="models"))
@@ -130,7 +132,7 @@ def quantize_model(
                 ov_model.save_pretrained(folder)
                 new_repo_url = api.create_repo(repo_id=new_repo_id, exist_ok=True, private=private_repo)
                 new_repo_id = new_repo_url.repo_id
-                print("Repo created successfully!", new_repo_url)
+                print("Repository created successfully!", new_repo_url)
 
                 folder = Path(folder)
                 for dir_name in (
@@ -169,9 +171,9 @@ def quantize_model(
                 card.data.base_model = model_id
                 card.text = dedent(
                     f"""
-                    This model is a quantized version of [`{model_id}`](https://huggingface.co/{model_id}) and was exported to the OpenVINO format using [optimum-intel](https://github.com/huggingface/optimum-intel) via the [nncf-quantization](https://huggingface.co/spaces/echarlaix/nncf-quantization) space.
+                    This model is a quantized version of [`{model_id}`](https://huggingface.co/{model_id}) and is converted to the OpenVINO format. This model was obtained via the [nncf-quantization](https://huggingface.co/spaces/echarlaix/nncf-quantization) space with [optimum-intel](https://github.com/huggingface/optimum-intel).
 
-                    First make sure you have optimum-intel installed:
+                    First make sure you have `optimum-intel` installed:
 
                     ```bash
                     pip install optimum[openvino]
@@ -195,16 +197,16 @@ def quantize_model(
                     path_in_repo="README.md",
                     repo_id=new_repo_id,
                 )
-                return f"This model was successfully quantized, find it under your repo {new_repo_url}"
+                return f"This model was successfully quantized, find it under your repository {new_repo_url}"
             finally:
                 shutil.rmtree(folder, ignore_errors=True)
     except Exception as e:
         return f"### Error: {e}"
 
 DESCRIPTION = """
-This Space uses [Optimum Intel](https://huggingface.co/docs/optimum/main/en/intel/openvino/optimization) to automatically apply NNCF weight only quantization on a model hosted on the [Hub](https://huggingface.co/models) and convert it to the [OpenVINO format](https://docs.openvino.ai/2024/documentation/openvino-ir-format.html) if not already.
+This Space uses [Optimum Intel](https://github.com/huggingface/optimum-intel) to automatically apply NNCF [Weight Only Quantization](https://huggingface.co/docs/optimum/main/en/intel/openvino/optimization) (WOQ) on your model and convert it to the [OpenVINO format](https://docs.openvino.ai/2024/documentation/openvino-ir-format.html) if not already.
 
-The resulting model will then be pushed under your HF user namespace. For now we only support conversion for models that are hosted on public repositories.
+After conversion, a repository will be pushed under your namespace with the resulting model.
 
 The list of the supported architectures can be found in the [documentation](https://huggingface.co/docs/optimum/main/en/intel/openvino/models)
 """
@@ -215,9 +217,9 @@ model_id = HuggingfaceHubSearch(
     search_type="model",
 )
 dtype = gr.Dropdown(
-    ["int8", "int4"],
-    value="int8",
-    label="Precision data types",
+    ["8-bit", "4-bit"],
+    value="8-bit",
+    label="Weights precision",
     filterable=False,
     visible=True,
 )
@@ -255,13 +257,13 @@ ratio = gr.Slider(
 )
 private_repo = gr.Checkbox(
     value=False,
-    label="Private Repo",
-    info="Create a private repo under your username",
+    label="Private repository",
+    info="Create a private repository instead of a public one",
 )
 overwritte = gr.Checkbox(
     value=False,
-    label="Overwrite repo content",
-    info="Enable pushing files on existing repo, potentially overwriting existing files",
+    label="Overwrite repository content",
+    info="Enable pushing files on existing repositories, potentially overwriting existing files",
 )
 interface = gr.Interface(
     fn=quantize_model,