summaryrefslogtreecommitdiffstats
path: root/g4f/Provider/HuggingFace.py
diff options
context:
space:
mode:
authorzukixa <56563509+zukixa@users.noreply.github.com>2024-08-29 08:03:32 +0200
committerzukixa <56563509+zukixa@users.noreply.github.com>2024-08-29 08:03:32 +0200
commitbda2d679275ebd6bc2a3aa84f8fbc93967bead63 (patch)
tree8a36ace98ab138e1eff134a5ed8891fd3c817b5b /g4f/Provider/HuggingFace.py
parent8 providers improved (diff)
downloadgpt4free-bda2d679275ebd6bc2a3aa84f8fbc93967bead63.tar
gpt4free-bda2d679275ebd6bc2a3aa84f8fbc93967bead63.tar.gz
gpt4free-bda2d679275ebd6bc2a3aa84f8fbc93967bead63.tar.bz2
gpt4free-bda2d679275ebd6bc2a3aa84f8fbc93967bead63.tar.lz
gpt4free-bda2d679275ebd6bc2a3aa84f8fbc93967bead63.tar.xz
gpt4free-bda2d679275ebd6bc2a3aa84f8fbc93967bead63.tar.zst
gpt4free-bda2d679275ebd6bc2a3aa84f8fbc93967bead63.zip
Diffstat (limited to 'g4f/Provider/HuggingFace.py')
-rw-r--r--g4f/Provider/HuggingFace.py51
1 files changed, 42 insertions, 9 deletions
diff --git a/g4f/Provider/HuggingFace.py b/g4f/Provider/HuggingFace.py
index 6634aa75..a3741196 100644
--- a/g4f/Provider/HuggingFace.py
+++ b/g4f/Provider/HuggingFace.py
@@ -1,20 +1,17 @@
from __future__ import annotations
-
import json
from aiohttp import ClientSession, BaseConnector
-
from ..typing import AsyncResult, Messages
from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
from .helper import get_connector
from ..errors import RateLimitError, ModelNotFoundError
from ..requests.raise_for_status import raise_for_status
-
class HuggingFace(AsyncGeneratorProvider, ProviderModelMixin):
url = "https://huggingface.co/chat"
working = True
needs_auth = True
supports_message_history = True
- default_model = "mistralai/Mixtral-8x7B-Instruct-v0.1"
+ default_model = "meta-llama/Meta-Llama-3.1-70B-Instruct"
models = [
'meta-llama/Meta-Llama-3.1-70B-Instruct',
'meta-llama/Meta-Llama-3.1-405B-Instruct-FP8',
@@ -22,10 +19,30 @@ class HuggingFace(AsyncGeneratorProvider, ProviderModelMixin):
'mistralai/Mixtral-8x7B-Instruct-v0.1',
'NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO',
'01-ai/Yi-1.5-34B-Chat',
- 'mistralai/Mistral-7B-Instruct-v0.2',
+ 'mistralai/Mistral-7B-Instruct-v0.3',
'microsoft/Phi-3-mini-4k-instruct',
]
+ model_aliases = {
+ "llama-3.1-70b": "meta-llama/Meta-Llama-3.1-70B-Instruct",
+ "llama-3.1-405b": "meta-llama/Meta-Llama-3.1-405B-Instruct-FP8",
+ "command-r-plus": "CohereForAI/c4ai-command-r-plus",
+ "mixtral-8x7b": "mistralai/Mixtral-8x7B-Instruct-v0.1",
+ "mixtral-8x7b": "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",
+ "yi-1.5-34b": "01-ai/Yi-1.5-34B-Chat",
+ "mistral-7b": "mistralai/Mistral-7B-Instruct-v0.3",
+ "phi-3-mini-4k": "microsoft/Phi-3-mini-4k-instruct",
+ }
+
+ @classmethod
+ def get_model(cls, model: str) -> str:
+ if model in cls.models:
+ return model
+ elif model in cls.model_aliases:
+ return cls.model_aliases[model]
+ else:
+ return cls.default_model
+
@classmethod
async def create_async_generator(
cls,
@@ -40,10 +57,26 @@ class HuggingFace(AsyncGeneratorProvider, ProviderModelMixin):
temperature: float = 0.7,
**kwargs
) -> AsyncResult:
- model = cls.get_model(model) if not model else model
- headers = {}
+ model = cls.get_model(model)
+ headers = {
+ 'accept': '*/*',
+ 'accept-language': 'en',
+ 'cache-control': 'no-cache',
+ 'origin': 'https://huggingface.co',
+ 'pragma': 'no-cache',
+ 'priority': 'u=1, i',
+ 'referer': 'https://huggingface.co/chat/',
+ 'sec-ch-ua': '"Not)A;Brand";v="99", "Google Chrome";v="127", "Chromium";v="127"',
+ 'sec-ch-ua-mobile': '?0',
+ 'sec-ch-ua-platform': '"macOS"',
+ 'sec-fetch-dest': 'empty',
+ 'sec-fetch-mode': 'cors',
+ 'sec-fetch-site': 'same-origin',
+ 'user-agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/127.0.0.0 Safari/537.36',
+ }
if api_key is not None:
headers["Authorization"] = f"Bearer {api_key}"
+
params = {
"return_full_text": False,
"max_new_tokens": max_new_tokens,
@@ -51,6 +84,7 @@ class HuggingFace(AsyncGeneratorProvider, ProviderModelMixin):
**kwargs
}
payload = {"inputs": format_prompt(messages), "parameters": params, "stream": stream}
+
async with ClientSession(
headers=headers,
connector=get_connector(connector, proxy)
@@ -72,7 +106,6 @@ class HuggingFace(AsyncGeneratorProvider, ProviderModelMixin):
yield chunk
else:
yield (await response.json())[0]["generated_text"].strip()
-
def format_prompt(messages: Messages) -> str:
system_messages = [message["content"] for message in messages if message["role"] == "system"]
question = " ".join([messages[-1]["content"], *system_messages])
@@ -81,4 +114,4 @@ def format_prompt(messages: Messages) -> str:
for idx, message in enumerate(messages)
if message["role"] == "assistant"
])
- return f"{history}<s>[INST] {question} [/INST]"
+ return f"{history}<s>[INST] {question} [/INST]" \ No newline at end of file