Skip to content
GitHub
View on GitHub

Qwen3-32B

API reference for Qwen3_32B

from modal_training_gym.common.models.qwen3_32b import Qwen3_32B

Qwen3-32B (32 billion parameters) from Alibaba.

Inherits from: HFModelConfiguration, ModelConfiguration

FieldTypeDefaultDescription
model_namestr"Qwen/Qwen3-32B"HuggingFace repo ID or other model identifier. Default "".
model_path`strNone`None
architecture`ModelArchitectureNone`None
training`ModelTrainingConfigNone`ModelTrainingConfig(gpu_type='H100', n_nodes=4, tensor_model_parallel_size=1, pipeline_model_parallel_size=1, context_parallel_size=1, sequence_parallel=False, expert_model_parallel_size=1, moe_permute_fusion=False, moe_grouped_gemm=False, moe_shared_expert_overlap=False, moe_aux_loss_coeff=0.0, lora_rank=8, lora_alpha=16, target_modules='all-linear', merge_lora=False)

Download or materialize weights into the model volume.

Source: modal_training_gym/common/models/qwen3_32b.py