update

data/eval_data/byteplus/byteplus/skylark-lite-250215/shenzhen_sase/byteplus_api_key/20250819_180217/agent-bingoplus-ph-200-chat.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e3a2567a913b1dc14fa6ea70adc8cab8f21de50fdb7004c392f021cb3aae218
+size 2448697

data/eval_data/byteplus/byteplus/skylark-lite-250215/shenzhen_sase/byteplus_api_key/20250819_180217/agent-bingoplus-ph-90-choice.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b2603d2a05a92ea2638f37542987c7f402bd8f049a92c7530b5db3a4eec13344
+size 258528

data/eval_data/byteplus/byteplus/skylark-lite-250215/shenzhen_sase/byteplus_api_key/20250819_180217/agent-cod-zh-70-chat.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1351fab4285ddb5f9157a53a7418a38dd69f45de308a9875efa03827ca9670c
+size 308863

data/eval_data/byteplus/byteplus/skylark-lite-250215/shenzhen_sase/byteplus_api_key/20250819_180217/agent-lingoace-zh-400-choice.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fe3ef2bb2dbde07ced2c9fd6b4621fd6e5126e859077f11aaea4560862dff01
+size 1211263

data/eval_data/byteplus/byteplus/skylark-lite-250215/shenzhen_sase/byteplus_api_key/20250819_180217/agent-lingoace-zh-80-chat.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6848f25727166e024e2b88e6f8a256da2363a219700c371279bb05abcfea7613
+size 876960

data/eval_data/byteplus/byteplus/skylark-pro-250415/shenzhen_sase/byteplus_api_key/20250819_171424/agent-bingoplus-ph-200-chat.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8bb57b8e92f77983202f69ecaa498e7482881f843b9be6469796998e86ec2ec3
+size 2437685

data/eval_data/byteplus/byteplus/skylark-pro-250415/shenzhen_sase/byteplus_api_key/20250819_171424/agent-bingoplus-ph-90-choice.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d57411a307c61095f919310595ba039eaf080698657038ad469d715e9592007
+size 258551

data/eval_data/byteplus/byteplus/skylark-pro-250415/shenzhen_sase/byteplus_api_key/20250819_171424/agent-cod-zh-70-chat.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a1a25a4c54e498a2ef46c863bf0d341fdcbfb210b5d80af4be36f86b20265b2
+size 306012

data/eval_data/byteplus/byteplus/skylark-pro-250415/shenzhen_sase/byteplus_api_key/20250819_171424/agent-lingoace-zh-400-choice.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cdc386e8b56d6f9f098bd954067b91fb089d9aeaa53a356231155930eaaee1be
+size 1211294

data/eval_data/byteplus/byteplus/skylark-pro-250415/shenzhen_sase/byteplus_api_key/20250819_171424/agent-lingoace-zh-80-chat.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e44cf602fbf5ad5a7d1a4efb9394de7b3fb34aef61f55f309920e90ab4f547f
+size 877036

examples/make_raw_dataset/step_3_filter_by_keywords.py

@@ -15,8 +15,8 @@ def get_args():
     parser = argparse.ArgumentParser()
     parser.add_argument(
         "--dataset_dir",
-        # default=(project_path / "data/llm-log-hk/extract-dataset").as_posix(),
-        default=(project_path / "data/llm-log-hk/extract-dataset/20250804").as_posix(),
+        default=(project_path / "data/llm-log-hk/extract-dataset").as_posix(),
+        # default=(project_path / "data/llm-log-hk/extract-dataset/20250804").as_posix(),
         # default=(project_path / "data/llm-log-usa/extract-dataset").as_posix(),
         type=str
     )
@@ -53,51 +53,40 @@ def main():
             with open(response_file.as_posix(), "r", encoding="utf-8") as f:
                 response = f.read()
 
-            for key_str in [
-                # "LingoAce",
-                # "BingoPlus",
-                # " COD ",
-                # "机器狗",
-                # "andy",
-                "HomePro",
-                # "NXPay",
-                # "NX Money",
-                # "Exodus Bank",
-                # "Exodus Telecom",
-                # "Exodus Retail",
-                # "Exodus Automotive",
-                # "kta kilat", "KTA KILAT",
-                # "NXCloud",
-                # "作为VIP客户",
-                # "FedEx",
-                # "Chinese laser cutting",
-            ]:
-                if system_prompt.__contains__(key_str) or user_prompt.__contains__(key_str):
-                    print(f"process: {sample_dir.as_posix()}")
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-lingoace"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-bingoplus"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-cod"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-andy"
-                    tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-homepro"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-nxpay"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-nxmoney"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-exodus-bank"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-exodus-telecom"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-exodus-retail"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-exodus-automotive"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-kta"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-nxcloud"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-vip"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-fedex"
-                    # tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-laser"
+            filter_map = [
+                (["LingoAce"], "lingoace"),
+                (["BingoPlus"], "bingoplus"),
+                ([" COD "], "cod"),
+                (["HomePro"], "homepro"),
+                (["andy"], "andy" ),
+                (["NXPay"], "nxpay"),
+                (["NX Money"], "nxmoney"),
+                (["Exodus Bank"], "exodus-bank"),
+                (["Exodus Telecom"], "exodus-telecom"),
+                (["Exodus Retail"], "exodus-retail"),
+                (["Exodus Automotive"], "exodus-automotive"),
+                (["kta kilat", "KTA KILAT"], "kta"),
+                (["作为VIP客户"], "vip"),
+                (["FedEx"], "fedex"),
+                (["Chinese laser cutting"], "laser"),
+            ]
 
-                    tgt_dir.mkdir(parents=True, exist_ok=True)
-                    shutil.move(
-                        sample_dir.as_posix(),
-                        tgt_dir.as_posix(),
-                    )
+            flag = False
+            for keys, suffix in filter_map:
+                if flag:
                     break
-
+                for key_str in keys:
+                    if flag:
+                        break
+                    if system_prompt.__contains__(key_str) or user_prompt.__contains__(key_str):
+                        print(f"process: {sample_dir.as_posix()}")
+                        tgt_dir = dataset_dir / f"{data_dir.parts[-1]}-{suffix}"
+                        tgt_dir.mkdir(parents=True, exist_ok=True)
+                        shutil.move(
+                            sample_dir.as_posix(),
+                            tgt_dir.as_posix(),
+                        )
+                        flag = True
     return
 
 

examples/test_metrics/bingoplus_chat_metric.py

@@ -38,12 +38,12 @@ python3 azure_openai.py --model_name gpt-4o-mini \
     )
     parser.add_argument(
         "--eval_data_file",
-        default=(project_path / "data/eval_data/siliconflow/siliconflow/Tongyi-Zhiwen#QwenLong-L1-32B/shenzhen_sase/siliconflow_api_key/20250814_134104/agent-bingoplus-ph-200-chat.jsonl.raw").as_posix(),
+        default=(project_path / "data/eval_data/byteplus/byteplus/skylark-pro-250415/shenzhen_sase/byteplus_api_key/20250819_171424/agent-bingoplus-ph-200-chat.jsonl.raw").as_posix(),
         type=str
     )
     parser.add_argument(
         "--output_file",
-        default=(project_path / "data/eval_data/siliconflow/siliconflow/Tongyi-Zhiwen#QwenLong-L1-32B/shenzhen_sase/siliconflow_api_key/20250814_134104/agent-bingoplus-ph-200-chat.jsonl").as_posix(),
+        default=(project_path / "data/eval_data/byteplus/byteplus/skylark-pro-250415/shenzhen_sase/byteplus_api_key/20250819_171424/agent-bingoplus-ph-200-chat.jsonl").as_posix(),
         type=str
     )
     parser.add_argument(

examples/test_metrics/cod_chat_metric.py

@@ -38,16 +38,12 @@ python3 azure_openai.py --model_name gpt-4o-mini \
     )
     parser.add_argument(
         "--eval_data_file",
-        default=(project_path / "data/eval_data/azure_openai/azure/gpt-4o-mini/shenzhen_sase/west_us_chatgpt_openai_azure_com/20250806_114802/agent-cod-zh-70-chat.jsonl.raw").as_posix(),
-        # default=(project_path / "data/eval_data/byteplus/byteplus/seed-1-6-flash-250615/shenzhen_sase/byteplus_api_key/20250728_113641/agent-cod-zh-70-chat.jsonl.raw").as_posix(),
-        # default=(project_path / "data/eval_data/gemini_google/google/gemini-2.5-flash/shenzhen_sase/google_potent_veld_462405_t3/20250808_160530/agent-cod-zh-70-chat.jsonl.raw").as_posix(),
+        default=(project_path / "data/eval_data/byteplus/byteplus/skylark-pro-250415/shenzhen_sase/byteplus_api_key/20250819_171424/agent-cod-zh-70-chat.jsonl.raw").as_posix(),
         type=str
     )
     parser.add_argument(
         "--output_file",
-        default=(project_path / "data/eval_data/azure_openai/azure/gpt-4o-mini/shenzhen_sase/west_us_chatgpt_openai_azure_com/20250806_114802/agent-cod-zh-70-chat.jsonl").as_posix(),
-        # default=(project_path / "data/eval_data/byteplus/byteplus/seed-1-6-flash-250615/shenzhen_sase/byteplus_api_key/20250728_113641/agent-cod-zh-70-chat.jsonl").as_posix(),
-        # default=(project_path / "data/eval_data/gemini_google/google/gemini-2.5-flash/shenzhen_sase/google_potent_veld_462405_t3/20250808_160530/agent-cod-zh-70-chat.jsonl").as_posix(),
+        default=(project_path / "data/eval_data/byteplus/byteplus/skylark-pro-250415/shenzhen_sase/byteplus_api_key/20250819_171424/agent-cod-zh-70-chat.jsonl").as_posix(),
         type=str
     )
     parser.add_argument(

examples/test_metrics/lingoace_chat_metric.py

@@ -43,12 +43,12 @@ python3 azure_openai.py --model_name gpt-4o-mini \
     )
     parser.add_argument(
         "--eval_data_file",
-        default=(project_path / "data/eval_data/siliconflow/siliconflow/Tongyi-Zhiwen#QwenLong-L1-32B/shenzhen_sase/siliconflow_api_key/20250814_134104/agent-lingoace-zh-80-chat.jsonl.raw").as_posix(),
+        default=(project_path / "data/eval_data/byteplus/byteplus/skylark-pro-250415/shenzhen_sase/byteplus_api_key/20250819_171424/agent-lingoace-zh-80-chat.jsonl.raw").as_posix(),
         type=str
     )
     parser.add_argument(
         "--output_file",
-        default=(project_path / "data/eval_data/siliconflow/siliconflow/Tongyi-Zhiwen#QwenLong-L1-32B/shenzhen_sase/siliconflow_api_key/20250814_134104/agent-lingoace-zh-80-chat.jsonl").as_posix(),
+        default=(project_path / "data/eval_data/byteplus/byteplus/skylark-pro-250415/shenzhen_sase/byteplus_api_key/20250819_171424/agent-lingoace-zh-80-chat.jsonl").as_posix(),
         type=str
     )
     parser.add_argument(

examples/tokenization/byteplus/step_1_get_by_api.py

@@ -66,12 +66,13 @@ def main():
         "Content-Type": "application/json",
         "Authorization": f"Bearer {api_key}"
     }
+
     data = {
         # "model": args.model_name,
-        # "model": "doubao-pro-32k-241215",
+        "model": "doubao-pro-32k-241215",
         # "model": "seed-1-6-250615",
         # "model": "seed-1-6-flash-250615",
-        "model": "seed-1-6-flash-250715",
+        # "model": "seed-1-6-flash-250715",
 
         "text": [
             "天空为什么这么蓝",

llm_eval_script/byteplus.py

@@ -38,13 +38,19 @@ def get_args():
 
     deepseek-v3
     deepseek-v3-250324
+
+    skylark-pro-250415
+    skylark-lite-250215
+
     """
     parser = argparse.ArgumentParser()
     parser.add_argument(
         "--model_name",
         # default="seed-1-6-250615",
-        default="seed-1-6-flash-250615",
+        # default="seed-1-6-flash-250615",
         # default="deepseek-v3-250324",
+        # default="skylark-pro-250415",
+        default="skylark-lite-250215",
         type=str
     )
     parser.add_argument(

llm_eval_script/byteplus_chat.py

@@ -44,8 +44,10 @@ def get_args():
     parser.add_argument(
         "--model_name",
         # default="seed-1-6-250615",
-        default="seed-1-6-flash-250615",
+        # default="seed-1-6-flash-250615",
         # default="deepseek-v3-250324",
+        default="skylark-pro-250415",
+        # default="skylark-lite-250215",
         type=str
     )
     parser.add_argument(