/opt/chainguard/nemo-skills/venv/bin/litellm

15:01:14 - LiteLLM:WARNING: get_model_cost_map.py:290 - LiteLLM: Failed to fetch remote model cost map from https://raw.githubusercontent.com/BerriAI/litellm/main/model_prices_and_context_window.json: [Errno -3] Temporary failure in name resolution. Falling back to local backup.
Usage: litellm [OPTIONS] [CLI_ARGS]...

Options:
  --host TEXT                     Host for the server to listen on.
  --port INTEGER                  Port to bind the server to.
  --num_workers INTEGER           Number of worker processes for uvicorn /
                                  gunicorn, or Granian worker processes
                                  (--workers). Default is 1 (from
                                  DEFAULT_NUM_WORKERS_LITELLM_PROXY). With
                                  --run_granian, use --granian_threads for
                                  runtime threads per worker.
  --granian_threads INTEGER RANGE
                                  Only with --run_granian: runtime threads per
                                  worker process (Granian --runtime-threads /
                                  GRANIAN_RUNTIME_THREADS). Omit to use
                                  Granian's default (1).  [x>=1]
  --api_base TEXT                 API base URL.
  --api_version TEXT              For azure - pass in the api version.
  -m, --model TEXT                The model name to pass to litellm expects
  --alias TEXT                    The alias for the model - use this to give a
                                  litellm model name (e.g. "huggingface/codell
                                  ama/CodeLlama-7b-Instruct-hf") a more user-
                                  friendly name ("codellama")
  --add_key TEXT                  The model name to pass to litellm expects
  --headers TEXT                  headers for the API call
  --save                          Save the model-specific config
  --debug                         To debug the input
  --detailed_debug                To view detailed debug logs
  --use_queue                     To use celery workers for async endpoints
  --temperature FLOAT             Set temperature for the model
  --max_tokens INTEGER            Set max tokens for the model
  --request_timeout INTEGER       Set timeout in seconds for completion calls
  --drop_params                   Drop any unmapped params
  --add_function_to_prompt        If function passed but unsupported, pass it
                                  as prompt
  -c, --config TEXT               Path to the proxy configuration file (e.g.
                                  config.yaml). Usage `litellm --config
                                  config.yaml`
  --max_budget FLOAT              Set max budget for API calls - works for
                                  hosted models like OpenAI, TogetherAI,
                                  Anthropic, etc.`
  --telemetry BOOLEAN             Helps us know if people are using this
                                  feature. Turn this off by doing `--telemetry
                                  False`
  --log_config TEXT               Path to the logging configuration file
  --setup                         Run the interactive setup wizard to
                                  configure providers and generate a config
                                  file
  -v, --version                   Print LiteLLM version
  --health                        Make a chat/completions request to all llms
                                  in config.yaml
  --test                          proxy chat completions url to make a test
                                  request to
  --test_async                    Calls async endpoints /queue/requests and
                                  /queue/response
  --iam_token_db_auth             Connects to RDS DB with IAM token
  --num_requests INTEGER          Number of requests to hit async endpoint
                                  with
  --run_gunicorn                  Starts proxy via gunicorn, instead of
                                  uvicorn (better for managing multiple
                                  workers)
  --run_hypercorn                 Starts proxy via hypercorn, instead of
                                  uvicorn (supports HTTP/2)
  --run_granian                   Starts proxy via Granian (Rust ASGI server)
                                  instead of uvicorn. Requires Python 3.10+
                                  and the `granian` package.
  --ssl_keyfile_path TEXT         Path to the SSL keyfile. Use this when you
                                  want to provide SSL certificate when
                                  starting proxy
  --ssl_certfile_path TEXT        Path to the SSL certfile. Use this when you
                                  want to provide SSL certificate when
                                  starting proxy
  --ciphers TEXT                  Ciphers to use for the SSL setup.
  --use_prisma_db_push            Use prisma db push instead of prisma migrate
                                  for database schema updates
  --local                         for local debugging
  --skip_server_startup           Skip starting the server after setup (useful
                                  for migrations only)
  --keepalive_timeout INTEGER     Set the uvicorn keepalive timeout in seconds
                                  (uvicorn timeout_keep_alive parameter)
  --timeout_worker_healthcheck INTEGER
                                  Set the uvicorn worker health-check timeout
                                  in seconds (uvicorn
                                  timeout_worker_healthcheck parameter).
                                  Requires uvicorn>=0.37.0. Only applies when
                                  running uvicorn directly with
                                  --num_workers>1; ignored under
                                  --run_gunicorn / --run_hypercorn.
  --max_requests_before_restart INTEGER
                                  Restart worker after this many requests
                                  (uvicorn: limit_max_requests, gunicorn:
                                  max_requests)
  --max_requests_before_restart_jitter INTEGER
                                  Stagger worker restarts by adding a random
                                  amount in [0, jitter] to
                                  --max_requests_before_restart so workers do
                                  not recycle at the same time (uvicorn:
                                  limit_max_requests_jitter, requires
                                  uvicorn>=0.41.0; gunicorn:
                                  max_requests_jitter). Has no effect without
                                  --max_requests_before_restart.
  --limit_concurrency INTEGER RANGE
                                  Set uvicorn's concurrency limit. Uvicorn
                                  counts both active tasks and accepted
                                  connections and returns HTTP 503 after the
                                  limit is reached. Idle connections can
                                  consume capacity, so use upstream
                                  connection/header timeouts and per-client
                                  connection limits. Only applies to uvicorn
                                  (ignored under --run_gunicorn /
                                  --run_hypercorn / --run_granian).  [x>=1]
  --enforce_prisma_migration_check
                                  Exit with error if database migration fails
                                  on startup.
  --use_v2_migration_resolver     Opt into the v2 migration resolver. Avoids
                                  the diff-and-force recovery path that can
                                  cause schema thrashing during rolling
                                  deploys where two LiteLLM versions contend
                                  for the same DB. Default is the v1 resolver.
  --reload                        Enable uvicorn hot reload (dev only). Also
                                  reloads when the --config YAML file changes.
                                  Incompatible with --num_workers>1,
                                  --run_gunicorn, and --run_hypercorn.
  --help                          Show this message and exit.