Alphafold2 API error: Internal Server Error, Inference error

AI & Data Science NVIDIA NIM Models
,
1

Thank you so much for providing these convenient products. I can use the diffdock API successfully, however, I have difficulties in using alphafold2 model via API.

My test is exactly same with the python and bash example you provided in the webpage:

#!/usr/bin/env python3
import os
import requests
import time
from pathlib import Path

# Variables
key = os.getenv("NVCF_RUN_KEY") or input("Paste the Run Key: ")
url = os.getenv("URL", "https://health.api.nvidia.com/v1/biology/deepmind/alphafold2")
status_url = os.getenv("STATUS_URL", "https://health.api.nvidia.com/v1/status")

sequence = ("MVPSAGQLALFALGIVLAACQALENSTSPLSADPPVAAAVVSHFNDCPDSHTQFCFHGTCRFL"
    "VQEDKPACVCHSGYVGARCEHADLLAVVAASQKKQAITALVVVSIVALAVLIITCVLIHCCQVRKHCEWCR"
    "ALICRHEKPSALLKGRTACCHSETVV"
)
output_file = Path("output.json")

# Initial request
headers = {
    "content-type": "application/json",
    "Authorization": f"Bearer keykeykey",
    "NVCF-POLL-SECONDS": "2",
}
data = {
    "sequence": sequence,
    "algorithm": "mmseqs2",
    "e_value": 0.0001,
    "iterations": 1,
    "databases": ["uniref90", "small_bfd", "mgnify"],
    "relax_prediction": True,
}

print("Making request...")
response = requests.post(url, headers=headers, json=data)
# Check the status code
if response.status_code == 200:
    output_file.write_text(response.text)
    print(f"Response output to file: {output_file}")
elif response.status_code == 202:
    print("Request accepted...")
    # Extract reqId header
    req_id = response.headers.get("nvcf-reqid")

    # Poll the /status endpoint
    while True:
        print("Polling for response...")
        status_response = requests.get(f"{status_url}/{req_id}", headers=headers)

        if status_response.status_code != 202:
            output_file.write_text(status_response.text)
            print(f"Response output to file: {output_file}")
            break

        # Wait before polling again
        time.sleep(2)
else:
    print(f"Unexpected HTTP status: {response.status_code}")
    print(f"Response: {response.text}")

I got the following results:

Making request…
Request accepted…
Polling for response…
Polling for response…
Polling for response…
Polling for response…
Polling for response…
Polling for response…
Polling for response…
Polling for response…
Response output to file: output.json

The output.json is :

{"type":"about:blank","status":500,"title":"Internal Server Error","detail":"Inference error"}

Could you please give me some hints on this ?

2

Hi @chenxin199261 ,Thanks for reporting this, we are taking a look at this issue currently.