experimental[minor]: adds mixtral wrapper (#17423)

**Description:** Adds a chat wrapper for Mixtral models using the
[prompt
template](https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1#instruction-format).

---------

Co-authored-by: Bagatur <baskaryan@gmail.com>

libs/experimental/langchain_experimental/chat_models/__init__.py

@@ -17,10 +17,11 @@ an interface where "chat messages" are the inputs and outputs.
     AIMessage, BaseMessage, HumanMessage
 """  # noqa: E501
 
-from langchain_experimental.chat_models.llm_wrapper import Llama2Chat, Orca, Vicuna
-
-__all__ = [
-    "Llama2Chat",
-    "Orca",
-    "Vicuna",
-]
+from langchain_experimental.chat_models.llm_wrapper import (
+    Llama2Chat,
+    Mixtral,
+    Orca,
+    Vicuna,
+)
+
+__all__ = ["Llama2Chat", "Orca", "Vicuna", "Mixtral"]

libs/experimental/langchain_experimental/chat_models/llm_wrapper.py

@@ -148,6 +148,23 @@ class Llama2Chat(ChatWrapper):
     usr_0_end: str = " [/INST]"
 
 
+class Mixtral(ChatWrapper):
+    """See https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1#instruction-format"""  # noqa: E501
+
+    @property
+    def _llm_type(self) -> str:
+        return "mixtral"
+
+    sys_beg: str = "<s>[INST] "
+    sys_end: str = "\n"
+    ai_n_beg: str = " "
+    ai_n_end: str = " </s>"
+    usr_n_beg: str = " [INST] "
+    usr_n_end: str = " [/INST]"
+    usr_0_beg: str = ""
+    usr_0_end: str = " [/INST]"
+
+
 class Orca(ChatWrapper):
     """Wrapper for Orca-style models."""
 

libs/experimental/tests/unit_tests/chat_models/test_llm_wrapper_mixtral.py

@@ -0,0 +1,31 @@
+import pytest
+from langchain.schema import AIMessage, HumanMessage, SystemMessage
+
+from langchain_experimental.chat_models import Mixtral
+from tests.unit_tests.chat_models.test_llm_wrapper_llama2chat import FakeLLM
+
+
+@pytest.fixture
+def model() -> Mixtral:
+    return Mixtral(llm=FakeLLM())
+
+
+@pytest.fixture
+def model_cfg_sys_msg() -> Mixtral:
+    return Mixtral(llm=FakeLLM(), system_message=SystemMessage(content="sys-msg"))
+
+
+def test_prompt(model: Mixtral) -> None:
+    messages = [
+        SystemMessage(content="sys-msg"),
+        HumanMessage(content="usr-msg-1"),
+        AIMessage(content="ai-msg-1"),
+        HumanMessage(content="usr-msg-2"),
+    ]
+
+    actual = model.predict_messages(messages).content  # type: ignore
+    expected = (
+        "<s>[INST] sys-msg\nusr-msg-1 [/INST] ai-msg-1 </s> [INST] usr-msg-2 [/INST]"  # noqa: E501
+    )
+
+    assert actual == expected