diff --git a/README.md b/README.md index 9344c0f22..684d5de73 100644 --- a/README.md +++ b/README.md @@ -226,6 +226,7 @@ curl 'http://0.0.0.0:4000/key/generate' \ | [deepinfra](https://docs.litellm.ai/docs/providers/deepinfra) | ✅ | ✅ | ✅ | ✅ | | [perplexity-ai](https://docs.litellm.ai/docs/providers/perplexity) | ✅ | ✅ | ✅ | ✅ | | [Groq AI](https://docs.litellm.ai/docs/providers/groq) | ✅ | ✅ | ✅ | ✅ | +| [Deepseek](https://docs.litellm.ai/docs/providers/deepseek) | ✅ | ✅ | ✅ | ✅ | | [anyscale](https://docs.litellm.ai/docs/providers/anyscale) | ✅ | ✅ | ✅ | ✅ | | [IBM - watsonx.ai](https://docs.litellm.ai/docs/providers/watsonx) | ✅ | ✅ | ✅ | ✅ | ✅ | [voyage ai](https://docs.litellm.ai/docs/providers/voyage) | | | | | ✅ | diff --git a/docs/my-website/docs/providers/deepseek.md b/docs/my-website/docs/providers/deepseek.md new file mode 100644 index 000000000..678561eca --- /dev/null +++ b/docs/my-website/docs/providers/deepseek.md @@ -0,0 +1,54 @@ +# Deepseek +https://deepseek.com/ + +**We support ALL Deepseek models, just set `deepseek/` as a prefix when sending completion requests** + +## API Key +```python +# env variable +os.environ['DEEPSEEK_API_KEY'] +``` + +## Sample Usage +```python +from litellm import completion +import os + +os.environ['DEEPSEEK_API_KEY'] = "" +response = completion( + model="deepseek/deepseek-chat", + messages=[ + {"role": "user", "content": "hello from litellm"} + ], +) +print(response) +``` + +## Sample Usage - Streaming +```python +from litellm import completion +import os + +os.environ['DEEPSEEK_API_KEY'] = "" +response = completion( + model="deepseek/deepseek-chat", + messages=[ + {"role": "user", "content": "hello from litellm"} + ], + stream=True +) + +for chunk in response: + print(chunk) +``` + + +## Supported Models - ALL Deepseek Models Supported! +We support ALL Deepseek models, just set `deepseek/` as a prefix when sending completion requests + +| Model Name | Function Call | +|--------------------------|------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| deepseek-chat | `completion(model="deepseek/deepseek-chat", messages)` | +| deepseek-coder | `completion(model="deepseek/deepseek-chat", messages)` | + + diff --git a/docs/my-website/sidebars.js b/docs/my-website/sidebars.js index f5777d6e7..d00d853a0 100644 --- a/docs/my-website/sidebars.js +++ b/docs/my-website/sidebars.js @@ -134,6 +134,7 @@ const sidebars = { "providers/ollama", "providers/perplexity", "providers/groq", + "providers/deepseek", "providers/fireworks_ai", "providers/vllm", "providers/xinference", diff --git a/litellm/__init__.py b/litellm/__init__.py index 5e745668b..4f72504f6 100644 --- a/litellm/__init__.py +++ b/litellm/__init__.py @@ -361,6 +361,7 @@ openai_compatible_endpoints: List = [ "api.deepinfra.com/v1/openai", "api.mistral.ai/v1", "api.groq.com/openai/v1", + "api.deepseek.com/v1", "api.together.xyz/v1", ] @@ -369,6 +370,7 @@ openai_compatible_providers: List = [ "anyscale", "mistral", "groq", + "deepseek", "deepinfra", "perplexity", "xinference", @@ -523,6 +525,7 @@ provider_list: List = [ "anyscale", "mistral", "groq", + "deepseek", "maritalk", "voyage", "cloudflare", diff --git a/litellm/main.py b/litellm/main.py index de35dbfd0..acd7f9b90 100644 --- a/litellm/main.py +++ b/litellm/main.py @@ -305,6 +305,7 @@ async def acompletion( or custom_llm_provider == "deepinfra" or custom_llm_provider == "perplexity" or custom_llm_provider == "groq" + or custom_llm_provider == "deepseek" or custom_llm_provider == "text-completion-openai" or custom_llm_provider == "huggingface" or custom_llm_provider == "ollama" @@ -982,6 +983,7 @@ def completion( or custom_llm_provider == "deepinfra" or custom_llm_provider == "perplexity" or custom_llm_provider == "groq" + or custom_llm_provider == "deepseek" or custom_llm_provider == "anyscale" or custom_llm_provider == "mistral" or custom_llm_provider == "openai" @@ -2168,7 +2170,7 @@ def completion( """ assume input to custom LLM api bases follow this format: resp = requests.post( - api_base, + api_base, json={ 'model': 'meta-llama/Llama-2-13b-hf', # model name 'params': { @@ -2565,6 +2567,7 @@ async def aembedding(*args, **kwargs): or custom_llm_provider == "deepinfra" or custom_llm_provider == "perplexity" or custom_llm_provider == "groq" + or custom_llm_provider == "deepseek" or custom_llm_provider == "fireworks_ai" or custom_llm_provider == "ollama" or custom_llm_provider == "vertex_ai" @@ -3085,6 +3088,7 @@ async def atext_completion(*args, **kwargs): or custom_llm_provider == "deepinfra" or custom_llm_provider == "perplexity" or custom_llm_provider == "groq" + or custom_llm_provider == "deepseek" or custom_llm_provider == "fireworks_ai" or custom_llm_provider == "text-completion-openai" or custom_llm_provider == "huggingface" diff --git a/litellm/utils.py b/litellm/utils.py index a938a0ba8..0b4aa0e87 100644 --- a/litellm/utils.py +++ b/litellm/utils.py @@ -4929,6 +4929,7 @@ def get_optional_params( and custom_llm_provider != "anyscale" and custom_llm_provider != "together_ai" and custom_llm_provider != "groq" + and custom_llm_provider != "deepseek" and custom_llm_provider != "mistral" and custom_llm_provider != "anthropic" and custom_llm_provider != "cohere_chat" @@ -5614,6 +5615,29 @@ def get_optional_params( if seed is not None: optional_params["seed"] = seed + elif custom_llm_provider == "deepseek": + supported_params = get_supported_openai_params( + model=model, custom_llm_provider=custom_llm_provider + ) + _check_valid_arg(supported_params=supported_params) + + if frequency_penalty is not None: + optional_params["frequency_penalty"] = frequency_penalty + if max_tokens is not None: + optional_params["max_tokens"] = max_tokens + if presence_penalty is not None: + optional_params["presence_penalty"] = presence_penalty + if stop is not None: + optional_params["stop"] = stop + if stream is not None: + optional_params["stream"] = stream + if temperature is not None: + optional_params["temperature"] = temperature + if logprobs is not None: + optional_params["logprobs"] = logprobs + if top_logprobs is not None: + optional_params["top_logprobs"] = top_logprobs + elif custom_llm_provider == "openrouter": supported_params = get_supported_openai_params( model=model, custom_llm_provider=custom_llm_provider @@ -5946,6 +5970,19 @@ def get_supported_openai_params(model: str, custom_llm_provider: str): "response_format", "seed", ] + elif custom_llm_provider == "deepseek": + return [ + # https://platform.deepseek.com/api-docs/api/create-chat-completion + "frequency_penalty", + "max_tokens", + "presence_penalty", + "stop", + "stream", + "temperature", + "top_p", + "logprobs", + "top_logprobs", + ] elif custom_llm_provider == "cohere": return [ "stream", @@ -6239,8 +6276,12 @@ def get_llm_provider( # groq is openai compatible, we just need to set this to custom_openai and have the api_base be https://api.groq.com/openai/v1 api_base = "https://api.groq.com/openai/v1" dynamic_api_key = get_secret("GROQ_API_KEY") + elif custom_llm_provider == "deepseek": + # deepseek is openai compatible, we just need to set this to custom_openai and have the api_base be https://api.deepseek.com/v1 + api_base = "https://api.deepseek.com/v1" + dynamic_api_key = get_secret("DEEPSEEK_API_KEY") elif custom_llm_provider == "fireworks_ai": - # fireworks is openai compatible, we just need to set this to custom_openai and have the api_base be https://api.groq.com/openai/v1 + # fireworks is openai compatible, we just need to set this to custom_openai and have the api_base be https://api.fireworks.ai/inference/v1 if not model.startswith("accounts/fireworks/models"): model = f"accounts/fireworks/models/{model}" api_base = "https://api.fireworks.ai/inference/v1" @@ -6303,6 +6344,9 @@ def get_llm_provider( elif endpoint == "api.groq.com/openai/v1": custom_llm_provider = "groq" dynamic_api_key = get_secret("GROQ_API_KEY") + elif endpoint == "api.deepseek.com/v1": + custom_llm_provider = "deepseek" + dynamic_api_key = get_secret("DEEPSEEK_API_KEY") return model, custom_llm_provider, dynamic_api_key, api_base # check if model in known model provider list -> for huggingface models, raise exception as they don't have a fixed provider (can be togetherai, anyscale, baseten, runpod, et.) @@ -6901,6 +6945,11 @@ def validate_environment(model: Optional[str] = None) -> dict: keys_in_environment = True else: missing_keys.append("GROQ_API_KEY") + elif custom_llm_provider == "deepseek": + if "DEEPSEEK_API_KEY" in os.environ: + keys_in_environment = True + else: + missing_keys.append("DEEPSEEK_API_KEY") elif custom_llm_provider == "mistral": if "MISTRAL_API_KEY" in os.environ: keys_in_environment = True