docs(supported_embeddings.md): add doc on provider-specific params for embedding models

docs/my-website/docs/embedding/supported_embedding.md

@@ -85,6 +85,17 @@ print(query_result[:5])
 </Tabs>
 
 ## Input Params for `litellm.embedding()`
+
+
+:::info
+
+Any non-openai params, will be treated as provider-specific params, and sent in the request body as kwargs to the provider.
+
+[**See Reserved Params**](https://github.com/BerriAI/litellm/blob/2f5f85cb52f36448d1f8bbfbd3b8af8167d0c4c8/litellm/main.py#L3130)
+
+[**See Example**](#example)
+:::
+
 ### Required Fields
 
 - `model`: *string* - ID of the model to use. `model='text-embedding-ada-002'`
@@ -363,3 +374,62 @@ All models listed here https://docs.voyageai.com/embeddings/#models-and-specific
 | voyage-01 | `embedding(model="voyage/voyage-01", input)` | 
 | voyage-lite-01 | `embedding(model="voyage/voyage-lite-01", input)` | 
 | voyage-lite-01-instruct | `embedding(model="voyage/voyage-lite-01-instruct", input)` | 
+
+## Provider-specific Params
+
+
+:::info
+
+Any non-openai params, will be treated as provider-specific params, and sent in the request body as kwargs to the provider.
+
+[**See Reserved Params**](https://github.com/BerriAI/litellm/blob/2f5f85cb52f36448d1f8bbfbd3b8af8167d0c4c8/litellm/main.py#L3130)
+:::
+
+### **Example**
+
+Cohere v3 Models have a required parameter: `input_type`, it can be one of the following four values:
+
+- `input_type="search_document"`: (default) Use this for texts (documents) you want to store in your vector database
+- `input_type="search_query"`: Use this for search queries to find the most relevant documents in your vector database
+- `input_type="classification"`: Use this if you use the embeddings as an input for a classification system
+- `input_type="clustering"`: Use this if you use the embeddings for text clustering
+
+https://txt.cohere.com/introducing-embed-v3/
+
+<Tabs>
+<TabItem value="sdk" label="SDK">
+
+```python
+from litellm import embedding
+os.environ["COHERE_API_KEY"] = "cohere key"
+
+# cohere call
+response = embedding(
+    model="embed-english-v3.0", 
+    input=["good morning from litellm", "this is another item"], 
+    input_type="search_document" # 👈 PROVIDER-SPECIFIC PARAM
+)
+```
+</TabItem>
+<TabItem value="proxy" label="PROXY">
+
+**via config**
+
+```yaml
+model_list:
+  - model_name: "cohere-embed"
+    litellm_params:
+      model: embed-english-v3.0
+      input_type: search_document # 👈 PROVIDER-SPECIFIC PARAM
+```
+
+**via request**
+
+```bash
+curl -X POST 'http://0.0.0.0:4000/v1/embeddings' \
+-H 'Authorization: Bearer sk-54d77cd67b9febbb' \
+-H 'Content-Type: application/json' \
+-D '{"input": ["Are you authorized to work in United States of America?"], "model": "cohere-embed", "input_type": "search_document}'
+```
+</TabItem>
+</Tabs>

docs/my-website/docs/providers/cohere.md

@@ -68,7 +68,7 @@ response = embedding(
 ```
 
 ### Setting - Input Type for v3 models
-v3 Models have a required parameter: `input_type`, it can be one of the following four values:
+v3 Models have a required parameter: `input_type`. LiteLLM defaults to `search_document`. It can be one of the following four values:
 
 - `input_type="search_document"`: (default) Use this for texts (documents) you want to store in your vector database
 - `input_type="search_query"`: Use this for search queries to find the most relevant documents in your vector database
@@ -76,6 +76,8 @@ v3 Models have a required parameter: `input_type`, it can be one of the followin
 - `input_type="clustering"`: Use this if you use the embeddings for text clustering
 
 https://txt.cohere.com/introducing-embed-v3/
+
+
 ```python
 from litellm import embedding
 os.environ["COHERE_API_KEY"] = "cohere key"

litellm/tests/test_embedding.py

@@ -1,18 +1,21 @@
-import sys, os
+import os
+import sys
 import traceback
+
+import openai
 import pytest
 from dotenv import load_dotenv
-import openai
 
 load_dotenv()
 
 sys.path.insert(
     0, os.path.abspath("../..")
 )  # Adds the parent directory to the system path
-import litellm
-from litellm import embedding, completion, completion_cost
 from unittest.mock import MagicMock, patch
 
+import litellm
+from litellm import completion, completion_cost, embedding
+
 litellm.set_verbose = False
 
 
@@ -232,6 +235,7 @@ def test_cohere_embedding():
         response = embedding(
             model="embed-english-v2.0",
             input=["good morning from litellm", "this is another item"],
+            input_type="search_query",
         )
         print(f"response:", response)
 
@@ -486,16 +490,16 @@ def test_mistral_embeddings():
 
 
 def test_watsonx_embeddings():
-    
-    def mock_wx_embed_request(method:str, url:str, **kwargs):
+
+    def mock_wx_embed_request(method: str, url: str, **kwargs):
         mock_response = MagicMock()
         mock_response.status_code = 200
         mock_response.headers = {"Content-Type": "application/json"}
         mock_response.json.return_value = {
-                "model_id": "ibm/slate-30m-english-rtrvr",
-                "created_at": "2024-01-01T00:00:00.00Z",
-                "results": [ {"embedding": [0.0]*254} ],
-                "input_token_count": 8
+            "model_id": "ibm/slate-30m-english-rtrvr",
+            "created_at": "2024-01-01T00:00:00.00Z",
+            "results": [{"embedding": [0.0] * 254}],
+            "input_token_count": 8,
         }
         return mock_response
 
@@ -505,7 +509,7 @@ def test_watsonx_embeddings():
             response = litellm.embedding(
                 model="watsonx/ibm/slate-30m-english-rtrvr",
                 input=["good morning from litellm"],
-                token="secret-token"
+                token="secret-token",
             )
         print(f"response: {response}")
         assert isinstance(response.usage, litellm.Usage)
@@ -514,26 +518,27 @@ def test_watsonx_embeddings():
     except Exception as e:
         pytest.fail(f"Error occurred: {e}")
 
+
 @pytest.mark.asyncio
 async def test_watsonx_aembeddings():
 
     def mock_async_client(*args, **kwargs):
 
         mocked_client = MagicMock()
-        
+
         async def mock_send(request, *args, stream: bool = False, **kwags):
             mock_response = MagicMock()
             mock_response.status_code = 200
             mock_response.headers = {"Content-Type": "application/json"}
             mock_response.json.return_value = {
-                    "model_id": "ibm/slate-30m-english-rtrvr",
-                    "created_at": "2024-01-01T00:00:00.00Z",
-                    "results": [ {"embedding": [0.0]*254} ],
-                    "input_token_count": 8
+                "model_id": "ibm/slate-30m-english-rtrvr",
+                "created_at": "2024-01-01T00:00:00.00Z",
+                "results": [{"embedding": [0.0] * 254}],
+                "input_token_count": 8,
             }
             mock_response.is_error = False
             return mock_response
-            
+
         mocked_client.send = mock_send
 
         return mocked_client
@@ -544,7 +549,7 @@ async def test_watsonx_aembeddings():
             response = await litellm.aembedding(
                 model="watsonx/ibm/slate-30m-english-rtrvr",
                 input=["good morning from litellm"],
-                token="secret-token"
+                token="secret-token",
             )
         print(f"response: {response}")
         assert isinstance(response.usage, litellm.Usage)