summaryrefslogtreecommitdiffstats
path: root/g4f/Provider/HuggingFace.py
diff options
context:
space:
mode:
Diffstat (limited to 'g4f/Provider/HuggingFace.py')
-rw-r--r--g4f/Provider/HuggingFace.py12
1 files changed, 8 insertions, 4 deletions
diff --git a/g4f/Provider/HuggingFace.py b/g4f/Provider/HuggingFace.py
index a3741196..74957862 100644
--- a/g4f/Provider/HuggingFace.py
+++ b/g4f/Provider/HuggingFace.py
@@ -1,11 +1,14 @@
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
@@ -22,7 +25,7 @@ class HuggingFace(AsyncGeneratorProvider, ProviderModelMixin):
'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",
@@ -76,7 +79,7 @@ class HuggingFace(AsyncGeneratorProvider, ProviderModelMixin):
}
if api_key is not None:
headers["Authorization"] = f"Bearer {api_key}"
-
+
params = {
"return_full_text": False,
"max_new_tokens": max_new_tokens,
@@ -84,7 +87,7 @@ class HuggingFace(AsyncGeneratorProvider, ProviderModelMixin):
**kwargs
}
payload = {"inputs": format_prompt(messages), "parameters": params, "stream": stream}
-
+
async with ClientSession(
headers=headers,
connector=get_connector(connector, proxy)
@@ -106,6 +109,7 @@ 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])
@@ -114,4 +118,4 @@ def format_prompt(messages: Messages) -> str:
for idx, message in enumerate(messages)
if message["role"] == "assistant"
])
- return f"{history}<s>[INST] {question} [/INST]" \ No newline at end of file
+ return f"{history}<s>[INST] {question} [/INST]"