update readme, add starcoder2, cosmopedia

Former-commit-id: 1ae7c183640146bb9b06c98942985a1721d2b9c9

src/llmtuner/extras/constants.py

@@ -743,6 +743,21 @@ register_model_group(
 )
 
 
+register_model_group(
+    models={
+        "StarCoder2-3B": {
+            DownloadSource.DEFAULT: "bigcode/starcoder2-3b",
+        },
+        "StarCoder2-7B": {
+            DownloadSource.DEFAULT: "bigcode/starcoder2-7b",
+        },
+        "StarCoder2-15B": {
+            DownloadSource.DEFAULT: "bigcode/starcoder2-15b",
+        },
+    }
+)
+
+
 register_model_group(
     models={
         "Vicuna1.5-7B-Chat": {

src/llmtuner/webui/components/export.py

@@ -28,6 +28,7 @@ def save_model(
     export_quantization_dataset: str,
     export_legacy_format: bool,
     export_dir: str,
+    export_hub_model_id: str,
 ) -> Generator[str, None, None]:
     error = ""
     if not model_name:
@@ -59,6 +60,7 @@ def save_model(
         finetuning_type=finetuning_type,
         template=template,
         export_dir=export_dir,
+        export_hub_model_id=export_hub_model_id or None,
         export_size=max_shard_size,
         export_quantization_bit=int(export_quantization_bit) if export_quantization_bit in GPTQ_BITS else None,
         export_quantization_dataset=export_quantization_dataset,
@@ -77,7 +79,10 @@ def create_export_tab(engine: "Engine") -> Dict[str, "Component"]:
         export_quantization_dataset = gr.Textbox(value="data/c4_demo.json")
         export_legacy_format = gr.Checkbox()
 
-    export_dir = gr.Textbox()
+    with gr.Row():
+        export_dir = gr.Textbox()
+        export_hub_model_id = gr.Textbox()
+
     export_btn = gr.Button()
     info_box = gr.Textbox(show_label=False, interactive=False)
 
@@ -95,6 +100,7 @@ def create_export_tab(engine: "Engine") -> Dict[str, "Component"]:
             export_quantization_dataset,
             export_legacy_format,
             export_dir,
+            export_hub_model_id,
         ],
         [info_box],
     )
@@ -105,6 +111,7 @@ def create_export_tab(engine: "Engine") -> Dict[str, "Component"]:
         export_quantization_dataset=export_quantization_dataset,
         export_legacy_format=export_legacy_format,
         export_dir=export_dir,
+        export_hub_model_id=export_hub_model_id,
         export_btn=export_btn,
         info_box=info_box,
     )

src/llmtuner/webui/locales.py

@@ -1019,6 +1019,20 @@ LOCALES = {
             "info": "保存导出模型的文件夹路径。",
         },
     },
+    "export_hub_model_id": {
+        "en": {
+            "label": "HF Hub ID (optional)",
+            "info": "Repo ID for uploading model to Hugging Face hub.",
+        },
+        "ru": {
+            "label": "HF Hub ID (опционально)",
+            "info": "Идентификатор репозитория для загрузки модели на Hugging Face hub.",
+        },
+        "zh": {
+            "label": "HF Hub ID（非必填）",
+            "info": "用于将模型上传至 Hugging Face Hub 的仓库 ID。",
+        },
+    },
     "export_btn": {
         "en": {
             "value": "Export",