From 2a909e634b72880dc7b57ca5886b02e8ccf5bcde Mon Sep 17 00:00:00 2001
From: Kepler <yuqitw@126.com>
Date: Fri, 27 Dec 2024 20:23:46 +0800
Subject: [PATCH] feat: support Ernie-lite-pro-128k (#12161)

Co-authored-by: bigfish49 <bigfish49@126.com>
---
 .../model_providers/wenxin/_common.py         |  1 +
 .../wenxin/llm/ernie-lite-pro-128k.yaml       | 42 +++++++++++++++++++
 2 files changed, 43 insertions(+)
 create mode 100644 api/core/model_runtime/model_providers/wenxin/llm/ernie-lite-pro-128k.yaml

diff --git a/api/core/model_runtime/model_providers/wenxin/_common.py b/api/core/model_runtime/model_providers/wenxin/_common.py
index c77a499982..1247a11fe8 100644
--- a/api/core/model_runtime/model_providers/wenxin/_common.py
+++ b/api/core/model_runtime/model_providers/wenxin/_common.py
@@ -122,6 +122,7 @@ class _CommonWenxin:
         "bge-large-zh": "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/embeddings/bge_large_zh",
         "tao-8k": "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/embeddings/tao_8k",
         "bce-reranker-base_v1": "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/reranker/bce_reranker_base",
+        "ernie-lite-pro-128k": "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/ernie-lite-pro-128k",
     }
 
     function_calling_supports = [
diff --git a/api/core/model_runtime/model_providers/wenxin/llm/ernie-lite-pro-128k.yaml b/api/core/model_runtime/model_providers/wenxin/llm/ernie-lite-pro-128k.yaml
new file mode 100644
index 0000000000..4f5832c859
--- /dev/null
+++ b/api/core/model_runtime/model_providers/wenxin/llm/ernie-lite-pro-128k.yaml
@@ -0,0 +1,42 @@
+model: ernie-lite-pro-128k
+label:
+  en_US: Ernie-Lite-Pro-128K
+model_type: llm
+features:
+  - agent-thought
+model_properties:
+  mode: chat
+  context_size: 128000
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+    min: 0.1
+    max: 1.0
+    default: 0.8
+  - name: top_p
+    use_template: top_p
+  - name: min_output_tokens
+    label:
+      en_US: "Min Output Tokens"
+      zh_Hans: "最小输出Token数"
+    use_template: max_tokens
+    min: 2
+    max: 2048
+    help:
+      zh_Hans: 指定模型最小输出token数
+      en_US: Specifies the lower limit on the length of generated results.
+  - name: max_output_tokens
+    label:
+      en_US: "Max Output Tokens"
+      zh_Hans: "最大输出Token数"
+    use_template: max_tokens
+    min: 2
+    max: 2048
+    default: 2048
+    help:
+      zh_Hans: 指定模型最大输出token数
+      en_US: Specifies the upper limit on the length of generated results. If the generated results are truncated, you can increase this parameter.
+  - name: presence_penalty
+    use_template: presence_penalty
+  - name: frequency_penalty
+    use_template: frequency_penalty