Fimeg/letta-server
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

feat: support deepseek models (#821)

Browse Source 
Co-authored-by: Charles Packer <packercharles@gmail.com>
Co-authored-by: Sarah Wooders <sarahwooders@gmail.com>
Co-authored-by: Shubham Naik <shub@memgpt.ai>
Co-authored-by: Shubham Naik <shub@letta.com>
This commit is contained in:
Kevin Lin
2025-02-18 15:28:01 -08:00
committed by GitHub
parent 6d49dc1ac5
commit 98f0062416
18 changed files with 709 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
letta/local_llm/chat_completion_proxy.py
View File

@@ -14,7 +14,7 @@ from letta.local_llm.grammars.gbnf_grammar_generator import create_dynamic_model
from letta.local_llm.koboldcpp.api import get_koboldcpp_completion
from letta.local_llm.llamacpp.api import get_llamacpp_completion
from letta.local_llm.llm_chat_completion_wrappers import simple_summary_wrapper
from letta.local_llm.lmstudio.api import get_lmstudio_completion
from letta.local_llm.lmstudio.api import get_lmstudio_completion, get_lmstudio_completion_chatcompletions
from letta.local_llm.ollama.api import get_ollama_completion
from letta.local_llm.utils import count_tokens, get_available_wrappers
from letta.local_llm.vllm.api import get_vllm_completion
@@ -141,11 +141,24 @@ def get_chat_completion(
f"Failed to convert ChatCompletion messages into prompt string with wrapper {str(llm_wrapper)} - error: {str(e)}"
)
# get the schema for the model
"""
if functions_python is not None:
model_schema = generate_schema(functions)
else:
model_schema = None
"""
# Run the LLM
try:
result_reasoning = None
if endpoint_type == "webui":
result, usage = get_webui_completion(endpoint, auth_type, auth_key, prompt, context_window, grammar=grammar)
elif endpoint_type == "webui-legacy":
result, usage = get_webui_completion_legacy(endpoint, auth_type, auth_key, prompt, context_window, grammar=grammar)
elif endpoint_type == "lmstudio-chatcompletions":
result, usage, result_reasoning = get_lmstudio_completion_chatcompletions(endpoint, auth_type, auth_key, model, messages)
elif endpoint_type == "lmstudio":
result, usage = get_lmstudio_completion(endpoint, auth_type, auth_key, prompt, context_window, api="completions")
elif endpoint_type == "lmstudio-legacy":
@@ -214,7 +227,7 @@ def get_chat_completion(
index=0,
message=Message(
role=chat_completion_result["role"],
content=chat_completion_result["content"],
content=result_reasoning if result_reasoning is not None else chat_completion_result["content"],
tool_calls=(
[ToolCall(id=get_tool_call_id(), type="function", function=chat_completion_result["function_call"])]
if "function_call" in chat_completion_result