Spaces:
Runtime error
Runtime error
etownsupport
commited on
Commit
•
0951ee0
1
Parent(s):
4e0c5c9
Upload 5 files
Browse files- Dockerfile +27 -0
- app.py +10 -0
- etown_mxbai/__init__.py +7 -0
- etown_mxbai/router.py +70 -0
- requirements.txt +6 -0
Dockerfile
ADDED
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Use an official Python runtime as a parent image
|
2 |
+
FROM python:3.10.9
|
3 |
+
|
4 |
+
# Set the working directory in the container to /app
|
5 |
+
WORKDIR /app
|
6 |
+
|
7 |
+
# Create a directory for Hugging Face cache and set broad permissions
|
8 |
+
RUN mkdir -p /app/hf_cache
|
9 |
+
RUN chmod -R 777 /app/hf_cache
|
10 |
+
|
11 |
+
# Set environment variable for Hugging Face home
|
12 |
+
ENV HF_HOME=/app/hf_cache
|
13 |
+
|
14 |
+
# Copy the requirements file into the container at /app
|
15 |
+
COPY ./requirements.txt /app/requirements.txt
|
16 |
+
|
17 |
+
# Install any needed packages specified in requirements.txt
|
18 |
+
RUN pip install --no-cache-dir --upgrade -r /app/requirements.txt
|
19 |
+
|
20 |
+
# Copy the rest of the application into the container at /app
|
21 |
+
COPY . /app
|
22 |
+
|
23 |
+
# Make port 7860 available to the world outside this container
|
24 |
+
EXPOSE 7860
|
25 |
+
|
26 |
+
# Define the command to run the app using uvicorn
|
27 |
+
CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]
|
app.py
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from etown_mxbai import app
|
2 |
+
|
3 |
+
# # etown_mxbai/app.py
|
4 |
+
# from fastapi import FastAPI
|
5 |
+
|
6 |
+
# app = FastAPI()
|
7 |
+
|
8 |
+
# @app.get("/")
|
9 |
+
# async def read_root():
|
10 |
+
# return {"message": "Hello World"}
|
etown_mxbai/__init__.py
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from fastapi import FastAPI
|
2 |
+
|
3 |
+
app = FastAPI(title="mixedbread-ai/mxbai-embed-large-v1 embeddings")
|
4 |
+
|
5 |
+
from etown_mxbai import router
|
6 |
+
|
7 |
+
|
etown_mxbai/router.py
ADDED
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from pydantic import BaseModel
|
2 |
+
from fastapi.middleware.cors import CORSMiddleware
|
3 |
+
from fastapi.responses import JSONResponse
|
4 |
+
# from sentence_transformers import SentenceTransformer
|
5 |
+
# from sentence_transformers.util import cos_sim
|
6 |
+
from typing import List
|
7 |
+
import os, platform, time
|
8 |
+
from transformers import AutoTokenizer
|
9 |
+
import fastembed
|
10 |
+
from fastembed import SparseEmbedding, SparseTextEmbedding, TextEmbedding
|
11 |
+
import numpy as np
|
12 |
+
|
13 |
+
|
14 |
+
sparse_model_name = "prithvida/Splade_PP_en_v1"
|
15 |
+
sparse_model = SparseTextEmbedding(model_name=sparse_model_name, batch_size=32)
|
16 |
+
|
17 |
+
class Validation(BaseModel):
|
18 |
+
prompt: List[str]
|
19 |
+
|
20 |
+
from etown_mxbai import app
|
21 |
+
|
22 |
+
app.add_middleware(
|
23 |
+
CORSMiddleware,
|
24 |
+
allow_origins=["*"],
|
25 |
+
allow_credentials=True,
|
26 |
+
allow_methods=["*"],
|
27 |
+
allow_headers=["*"],
|
28 |
+
)
|
29 |
+
|
30 |
+
@app.post("/api/generate", summary="Generate embeddings", tags=["Generate"])
|
31 |
+
def inference(item: Validation):
|
32 |
+
try:
|
33 |
+
start_time = time.time()
|
34 |
+
embeddings = list(sparse_model.embed(item.prompt, batch_size=5)) # Assuming 'model' is defined elsewhere
|
35 |
+
|
36 |
+
serializable_embeddings = []
|
37 |
+
for embedding in embeddings:
|
38 |
+
# Assuming embedding object has attributes values and indices
|
39 |
+
if isinstance(embedding, SparseEmbedding):
|
40 |
+
values = embedding.values
|
41 |
+
indices = embedding.indices
|
42 |
+
serializable_embeddings.append({
|
43 |
+
"values": values.tolist() if isinstance(values, np.ndarray) else values,
|
44 |
+
"indices": indices.tolist() if isinstance(indices, np.ndarray) else indices
|
45 |
+
})
|
46 |
+
else:
|
47 |
+
# Fallback for other types of embeddings
|
48 |
+
serializable_embeddings.append({
|
49 |
+
"values": embedding.tolist() if isinstance(embedding, np.ndarray) else str(embedding),
|
50 |
+
"indices": list(range(len(embedding))) if isinstance(embedding, (np.ndarray, list)) else []
|
51 |
+
})
|
52 |
+
|
53 |
+
end_time = time.time()
|
54 |
+
time_taken = end_time - start_time # Calculate the time taken
|
55 |
+
|
56 |
+
return JSONResponse(content={
|
57 |
+
"embeddings": serializable_embeddings,
|
58 |
+
"time_taken": f"{time_taken:.2f} seconds",
|
59 |
+
"Number_of_sentence_processed": len(item.prompt), # Assuming you want to count words, not characters
|
60 |
+
"Model_response_space" : "prithvida/Splade_PP_en_v1",
|
61 |
+
"status_code" : 200
|
62 |
+
})
|
63 |
+
except Exception as e:
|
64 |
+
print(f"An error occurred: {str(e)}") # Simple print statement for logging; consider using proper logging
|
65 |
+
return JSONResponse(content={
|
66 |
+
"error": "An error occurred during processing.",
|
67 |
+
"details": str(e),
|
68 |
+
"Model_response_space" : "prithvida/Splade_PP_en_v1",
|
69 |
+
"status_code" : 500
|
70 |
+
})
|
requirements.txt
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
fastapi
|
2 |
+
uvicorn
|
3 |
+
requests
|
4 |
+
pydantic
|
5 |
+
transformers
|
6 |
+
fastembed
|