databricks configure fail;: 400 Bad Request. Message: Error: Model response did not respect the required format

[livshitsa]

configure of databricks provider fails for databricks-meta-llama-3-3-70b-instruct
here is the error

Enter a model from that provider:
│ databricks-meta-llama-3-3-70b-instruct
│
◇ Request failed: Request failed with status: 400 Bad Request. Message: Error: Model response did not respect the required format. Please consider retrying or using a more straightforward prompt.

│
└ Failed to configure provider: init chat completion request with tool did not succeed.

[salman1993]

thanks for reporting this! i can replicate this on my side. will update this thread once we put in a fix.

[salman1993]

we are running into this error with databricks model databricks-meta-llama-3-3-70b-instruct but works with other models such as gpt-4o.

> curl https://YOUR_HOST/serving-endpoints/databricks-meta-llama-3-3-70b-instruct/invocations \
  -u token:$DATABRICKS_TOKEN \
  -X POST \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
        {
            "role": "user",
            "content": "What is the weather like in Paris today?"
        }
    ],
    "tools": [
        {
            "type": "function",
            "function": {
                "name": "get_weather",
                "description": "Get current temperature for a given location.",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "location": {
                            "type": "string",
                            "description": "City and country e.g. Bogotá, Colombia"
                        }
                    },
                    "required": [
                        "location"
                    ],
                    "additionalProperties": false
                },
                "strict": true
            }
        }
    ]
}'

{"id":"chatcmpl_eef44b42-5ecf-43dc-a9f6-0084718ebb25","object":"chat.completion","created":1739035458,"model":"meta-llama-3.3-70b-instruct-121024","choices":[{"index":0,"message":{"role":"assistant","content":null,"tool_calls":[{"id":"call_52d676b6-b6e8-43e7-867b-396c48e77ae7","type":"function","function":{"name":"get_weather","arguments":"{\"location\": \"Paris, France\"}"}}]},"finish_reason":"tool_calls","logprobs":null}],"usage":{"prompt_tokens":719,"completion_tokens":16,"total_tokens":735}}

[salman1993]

this PR should fix this: #1162

[salman1993]

you can try installing the canary:

curl -fsSL https://github.com/block/goose/releases/download/stable/download_cli.sh | CANARY=true bash

[livshitsa]

thanks for fixing; the configure works now, but using goose session gives the error, i just entered "what can you do" :
│ Configure Providers
│
◇ Which model provider should we use?
│ Databricks
│
● DATABRICKS_HOST is set via environment variable
│
◇ Would you like to save this value to your keyring?
│ Yes
│
● Saved DATABRICKS_HOST to config file
│
◇ Enter a model from that provider:
│ databricks-meta-llama-3-3-70b-instruct
│
◐ Checking your configuration... └ Configuration saved successfully
maccomputer % goose session
starting session | provider: databricks model: databricks-meta-llama-3-3-70b-instruct
logging to /Users/communicator/.config/goose/sessions/KNEzs19j.jsonl

Goose is running! Enter your instructions, or try asking what goose can do.

( O)> what can you do
◒ Nurturing neural nets... 2025-02-08T19:19:05.116652Z ERROR goose::agents::truncate: Error: Request failed: Request failed with status: 400 Bad Request. Message: Bad request: rpc error: code = InvalidArgument desc = Invalid JSON schema - could not process schema

at crates/goose/src/agents/truncate.rs:279

Ran into this error: Request failed: Request failed with status: 400 Bad Request. Message: Bad request: rpc error: code = InvalidArgument desc = Invalid JSON schema - c
ould not process schema
.

Please retry if you think this is a transient or recoverable error.

[salman1993]

what extensions are you using?

[livshitsa]

what extensions are you using?

just dev tools

[salman1993]

i could replicate this. seems like an error on Databricks model with tool calling but we're trying to debug it with Databricks