Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,16 +1,15 @@
|
|
1 |
-
import uvicorn
|
2 |
import os
|
|
|
3 |
from flask import Flask, request, jsonify
|
4 |
from huggingface_hub import InferenceClient
|
5 |
|
6 |
app = Flask(__name__)
|
7 |
|
8 |
-
#
|
9 |
-
|
10 |
-
API_URL = "https://api-inference.huggingface.co/models/mistralai/Mixtral-8x7B-Instruct-v0.1"
|
11 |
-
# API_KEY = os.getenv("API_KEY", "your_default_api_key") # Default API_KEY can be set here
|
12 |
-
API_KEY = os.getenv("API_KEY")
|
13 |
|
|
|
|
|
14 |
|
15 |
def format_prompt(message, custom_instructions=None):
|
16 |
prompt = ""
|
@@ -34,9 +33,8 @@ def normalize_text(text):
|
|
34 |
text = text.replace(british, american)
|
35 |
|
36 |
return text
|
37 |
-
|
38 |
def Mistral7B(prompt, instructions, api_key, temperature=0.1, max_new_tokens=2, top_p=0.95, repetition_penalty=1.0):
|
39 |
-
global API_URL
|
40 |
try:
|
41 |
temperature = max(float(temperature), 1e-2)
|
42 |
top_p = float(top_p)
|
@@ -52,13 +50,12 @@ def Mistral7B(prompt, instructions, api_key, temperature=0.1, max_new_tokens=2,
|
|
52 |
custom_instructions = instructions
|
53 |
formatted_prompt = format_prompt(prompt, custom_instructions)
|
54 |
|
55 |
-
client = InferenceClient(
|
56 |
-
response = client.text_generation(formatted_prompt, **generate_kwargs)
|
57 |
return response
|
58 |
except Exception as e:
|
59 |
return str(e)
|
60 |
|
61 |
-
|
62 |
@app.route("/generate-text", methods=["POST"])
|
63 |
def generate_text():
|
64 |
data = request.json
|
@@ -75,4 +72,4 @@ def generate_text():
|
|
75 |
return jsonify({"response": response}), 200
|
76 |
|
77 |
if __name__ == "__main__":
|
78 |
-
uvicorn.run(app, host="0.0.0.0", port=int(os.getenv("PORT", 8000)))
|
|
|
|
|
1 |
import os
|
2 |
+
import uvicorn
|
3 |
from flask import Flask, request, jsonify
|
4 |
from huggingface_hub import InferenceClient
|
5 |
|
6 |
app = Flask(__name__)
|
7 |
|
8 |
+
# Fixed API URL
|
9 |
+
MODEL_ID = "mistralai/Mixtral-8x7B-Instruct-v0.1"
|
|
|
|
|
|
|
10 |
|
11 |
+
# Use environment variable for the API key
|
12 |
+
API_KEY = os.getenv("API_KEY")
|
13 |
|
14 |
def format_prompt(message, custom_instructions=None):
|
15 |
prompt = ""
|
|
|
33 |
text = text.replace(british, american)
|
34 |
|
35 |
return text
|
36 |
+
|
37 |
def Mistral7B(prompt, instructions, api_key, temperature=0.1, max_new_tokens=2, top_p=0.95, repetition_penalty=1.0):
|
|
|
38 |
try:
|
39 |
temperature = max(float(temperature), 1e-2)
|
40 |
top_p = float(top_p)
|
|
|
50 |
custom_instructions = instructions
|
51 |
formatted_prompt = format_prompt(prompt, custom_instructions)
|
52 |
|
53 |
+
client = InferenceClient(token=api_key)
|
54 |
+
response = client.text_generation(formatted_prompt, model=MODEL_ID, **generate_kwargs)
|
55 |
return response
|
56 |
except Exception as e:
|
57 |
return str(e)
|
58 |
|
|
|
59 |
@app.route("/generate-text", methods=["POST"])
|
60 |
def generate_text():
|
61 |
data = request.json
|
|
|
72 |
return jsonify({"response": response}), 200
|
73 |
|
74 |
if __name__ == "__main__":
|
75 |
+
uvicorn.run(app, host="0.0.0.0", port=int(os.getenv("PORT", 8000)))
|