Files
2026-07-13 13:22:34 +08:00

94 lines
2.0 KiB
Python

from mlflow.ml_package_versions import GENAI_FLAVOR_TO_MODULE_NAME, NON_GENAI_FLAVOR_TO_MODULE_NAME
# NB: Kinesis PutRecords API has a limit of 500 records per request
BATCH_SIZE = 500
BATCH_TIME_INTERVAL_SECONDS = 10
MAX_QUEUE_SIZE = 1000
MAX_WORKERS = 1
CONFIG_STAGING_URL = "https://config-staging.mlflow-telemetry.io"
CONFIG_URL = "https://config.mlflow-telemetry.io"
UI_CONFIG_STAGING_URL = "https://d34z9x6fp23d2z.cloudfront.net"
UI_CONFIG_URL = "https://d139nb52glx00z.cloudfront.net"
RETRYABLE_ERRORS = [
429, # Throttled
500, # Interval Server Error
]
UNRECOVERABLE_ERRORS = [
400, # Bad Request
401, # Unauthorized
403, # Forbidden
404, # Not Found
]
GENAI_MODULES = {
"agno",
"anthropic",
"autogen",
"chromadb",
"crewai",
"dspy",
"faiss",
"google.genai", # gemini
"groq",
"haystack",
"langchain",
"langgraph",
"langsmith",
"litellm",
"llama_cpp",
"llama_index.core",
"milvus",
"mistralai",
"openai",
"pinecone",
"pydantic_ai",
"qdrant",
"ragas",
"semantic_kernel",
"smolagents",
"vllm",
"weaviate",
} | set(GENAI_FLAVOR_TO_MODULE_NAME.values())
NON_GENAI_MODULES = {
# Classic ML
"catboost",
"h2o",
"lightgbm",
"optuna",
"prophet",
"pyspark.ml",
"sklearn",
"spacy",
"statsmodels",
"xgboost",
# Deep Learning
"accelerate",
"bitsandbytes",
"deepspeed",
"diffusers",
"fastai",
"flash_attn",
"flax",
"jax",
"keras",
"lightning",
"mxnet",
"paddle",
"peft",
"sentence_transformers",
"tensorflow",
"timm",
"torch",
"transformers",
} | set(NON_GENAI_FLAVOR_TO_MODULE_NAME.values()) - {"pyspark"}
MODULES_TO_CHECK_IMPORT = GENAI_MODULES | NON_GENAI_MODULES
# fallback config to use for UI telemetry in case fetch fails
FALLBACK_UI_CONFIG = {
"disable_ui_telemetry": True,
"disable_ui_events": [],
"ui_rollout_percentage": 0,
}