Qwen3-30B-A3B
"# Qwen3-30B-A3B\n\n```python\nfrom modal_training_gym.common.models.qwen3_30b import Qwen3_30B\n```\n\nAlibaba Qwen3-30B-A3B MoE model with 30B total and 3B active parameters.\n\n**Attributes**\n\n<div class=\"tg-param\">\n<p><strong>model_name</strong> <code>str</code></p>\n<p>Hugging Face repository ID. <span class=\"tg-param-default\">Default: \"Qwen/Qwen3-30B-A3B\"</span></p>\n</div>\n\n<div class=\"tg-param\">\n<p><strong>architecture</strong> <code>ModelArchitecture | None</code></p>\n<p>Megatron architecture parameters for this model. <span class=\"tg-param-default\">Default: ModelArchitecture(...)</span></p>\n</div>\n\n<div class=\"tg-param\">\n<p><strong>response_parser</strong> <code>Callable[[str], ParsedResponse] | None</code></p>\n<p>Parser for generated text. <span class=\"tg-param-default\">Default: parse_qwen3_response</span></p>\n</div>\n\n## `download`\n\n```python\ndownload() -> None\n```\n\nDownload or materialize weights into the model volume.\n\n## `parse_response`\n\n```python\nparse_response(text: str) -> ParsedResponse\n```\n\nParse model text with `response_parser`.\n\nWithout a configured parser, the model text becomes `ParsedResponse.content`.\n\n**Returns**\n\nParsed model output.\n"
from modal_training_gym.common.models.qwen3_30b import Qwen3_30BAlibaba Qwen3-30B-A3B MoE model with 30B total and 3B active parameters.
Attributes
model_name str
Hugging Face repository ID. Default: "Qwen/Qwen3-30B-A3B"
architecture ModelArchitecture | None
Megatron architecture parameters for this model. Default: ModelArchitecture(...)
response_parser Callable[[str], ParsedResponse] | None
Parser for generated text. Default: parse_qwen3_response
download
Section titled “download”download() -> NoneDownload or materialize weights into the model volume.
parse_response
Section titled “parse_response”parse_response(text: str) -> ParsedResponseParse model text with response_parser.
Without a configured parser, the model text becomes ParsedResponse.content.
Returns
Parsed model output.