phoenix/litellm
1
0
Fork 0
forked from phoenix/litellm-mirror
Code Pull requests Activity

Added support for the deepseek api

Browse source
This commit is contained in:
Paul Gauthier 2024-05-07 11:44:03 -07:00
parent ee1b1fe4f8
commit 90eb0ea022
6 changed files with 114 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

1
README.md
View file

@ -226,6 +226,7 @@ curl 'http://0.0.0.0:4000/key/generate' \
| [deepinfra](https://docs.litellm.ai/docs/providers/deepinfra) | ✅ | ✅ | ✅ | ✅ | | [deepinfra](https://docs.litellm.ai/docs/providers/deepinfra) | ✅ | ✅ | ✅ | ✅ |
| [perplexity-ai](https://docs.litellm.ai/docs/providers/perplexity) | ✅ | ✅ | ✅ | ✅ | | [perplexity-ai](https://docs.litellm.ai/docs/providers/perplexity) | ✅ | ✅ | ✅ | ✅ |
| [Groq AI](https://docs.litellm.ai/docs/providers/groq) | ✅ | ✅ | ✅ | ✅ | | [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) | ✅ | ✅ | ✅ | ✅ | | [anyscale](https://docs.litellm.ai/docs/providers/anyscale) | ✅ | ✅ | ✅ | ✅ |
| [IBM - watsonx.ai](https://docs.litellm.ai/docs/providers/watsonx) | ✅ | ✅ | ✅ | ✅ | ✅ | [IBM - watsonx.ai](https://docs.litellm.ai/docs/providers/watsonx) | ✅ | ✅ | ✅ | ✅ | ✅
| [voyage ai](https://docs.litellm.ai/docs/providers/voyage) | | | | | ✅ | | [voyage ai](https://docs.litellm.ai/docs/providers/voyage) | | | | | ✅ |

54
docs/my-website/docs/providers/deepseek.md Normal file
View file

@ -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)` |

1
docs/my-website/sidebars.js
View file

@ -134,6 +134,7 @@ const sidebars = {
"providers/ollama", "providers/ollama",
"providers/perplexity", "providers/perplexity",
"providers/groq", "providers/groq",
"providers/deepseek",
"providers/fireworks_ai", "providers/fireworks_ai",
"providers/vllm", "providers/vllm",
"providers/xinference", "providers/xinference",

3
litellm/__init__.py
View file

@ -361,6 +361,7 @@ openai_compatible_endpoints: List = [
"api.deepinfra.com/v1/openai", "api.deepinfra.com/v1/openai",
"api.mistral.ai/v1", "api.mistral.ai/v1",
"api.groq.com/openai/v1", "api.groq.com/openai/v1",
"api.deepseek.com/v1",
"api.together.xyz/v1", "api.together.xyz/v1",
] ]
@ -369,6 +370,7 @@ openai_compatible_providers: List = [
"anyscale", "anyscale",
"mistral", "mistral",
"groq", "groq",
"deepseek",
"deepinfra", "deepinfra",
"perplexity", "perplexity",
"xinference", "xinference",
@ -523,6 +525,7 @@ provider_list: List = [
"anyscale", "anyscale",
"mistral", "mistral",
"groq", "groq",
"deepseek",
"maritalk", "maritalk",
"voyage", "voyage",
"cloudflare", "cloudflare",

4
litellm/main.py
View file

@ -305,6 +305,7 @@ async def acompletion(
or custom_llm_provider == "deepinfra" or custom_llm_provider == "deepinfra"
or custom_llm_provider == "perplexity" or custom_llm_provider == "perplexity"
or custom_llm_provider == "groq" or custom_llm_provider == "groq"
or custom_llm_provider == "deepseek"
or custom_llm_provider == "text-completion-openai" or custom_llm_provider == "text-completion-openai"
or custom_llm_provider == "huggingface" or custom_llm_provider == "huggingface"
or custom_llm_provider == "ollama" or custom_llm_provider == "ollama"
@ -982,6 +983,7 @@ def completion(
or custom_llm_provider == "deepinfra" or custom_llm_provider == "deepinfra"
or custom_llm_provider == "perplexity" or custom_llm_provider == "perplexity"
or custom_llm_provider == "groq" or custom_llm_provider == "groq"
or custom_llm_provider == "deepseek"
or custom_llm_provider == "anyscale" or custom_llm_provider == "anyscale"
or custom_llm_provider == "mistral" or custom_llm_provider == "mistral"
or custom_llm_provider == "openai" or custom_llm_provider == "openai"
@ -2565,6 +2567,7 @@ async def aembedding(*args, **kwargs):
or custom_llm_provider == "deepinfra" or custom_llm_provider == "deepinfra"
or custom_llm_provider == "perplexity" or custom_llm_provider == "perplexity"
or custom_llm_provider == "groq" or custom_llm_provider == "groq"
or custom_llm_provider == "deepseek"
or custom_llm_provider == "fireworks_ai" or custom_llm_provider == "fireworks_ai"
or custom_llm_provider == "ollama" or custom_llm_provider == "ollama"
or custom_llm_provider == "vertex_ai" 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 == "deepinfra"
or custom_llm_provider == "perplexity" or custom_llm_provider == "perplexity"
or custom_llm_provider == "groq" or custom_llm_provider == "groq"
or custom_llm_provider == "deepseek"
or custom_llm_provider == "fireworks_ai" or custom_llm_provider == "fireworks_ai"
or custom_llm_provider == "text-completion-openai" or custom_llm_provider == "text-completion-openai"
or custom_llm_provider == "huggingface" or custom_llm_provider == "huggingface"

51
litellm/utils.py
View file

@ -4929,6 +4929,7 @@ def get_optional_params(
and custom_llm_provider != "anyscale" and custom_llm_provider != "anyscale"
and custom_llm_provider != "together_ai" and custom_llm_provider != "together_ai"
and custom_llm_provider != "groq" and custom_llm_provider != "groq"
and custom_llm_provider != "deepseek"
and custom_llm_provider != "mistral" and custom_llm_provider != "mistral"
and custom_llm_provider != "anthropic" and custom_llm_provider != "anthropic"
and custom_llm_provider != "cohere_chat" and custom_llm_provider != "cohere_chat"
@ -5614,6 +5615,29 @@ def get_optional_params(
if seed is not None: if seed is not None:
optional_params["seed"] = seed 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": elif custom_llm_provider == "openrouter":
supported_params = get_supported_openai_params( supported_params = get_supported_openai_params(
model=model, custom_llm_provider=custom_llm_provider 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", "response_format",
"seed", "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": elif custom_llm_provider == "cohere":
return [ return [
"stream", "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 # 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" api_base = "https://api.groq.com/openai/v1"
dynamic_api_key = get_secret("GROQ_API_KEY") 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": 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"): if not model.startswith("accounts/fireworks/models"):
model = f"accounts/fireworks/models/{model}" model = f"accounts/fireworks/models/{model}"
api_base = "https://api.fireworks.ai/inference/v1" api_base = "https://api.fireworks.ai/inference/v1"
@ -6303,6 +6344,9 @@ def get_llm_provider(
elif endpoint == "api.groq.com/openai/v1": elif endpoint == "api.groq.com/openai/v1":
custom_llm_provider = "groq" custom_llm_provider = "groq"
dynamic_api_key = get_secret("GROQ_API_KEY") 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 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.) # 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 keys_in_environment = True
else: else:
missing_keys.append("GROQ_API_KEY") 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": elif custom_llm_provider == "mistral":
if "MISTRAL_API_KEY" in os.environ: if "MISTRAL_API_KEY" in os.environ:
keys_in_environment = True keys_in_environment = True