{ "auto_loss_scaling": false, "device_iterations": 2, "embedding_serialization_factor": 4, "enable_half_partials": true, "executable_cache_dir": "/tmp/exe_cache/language_modeling", "execute_encoder_on_cpu_for_generation": false, "gradient_accumulation_steps": 16, "inference_device_iterations": 5, "inference_replication_factor": 1, "ipus_per_replica": 4, "layers_per_ipu": [ 0, 4, 4, 4 ], "matmul_proportion": 0.25, "optimizer_state_offchip": true, "optimum_version": "1.5.1", "output_mode": "final", "recompute_checkpoint_every_layer": true, "replicated_tensor_sharding": false, "replication_factor": 1, "seed": 42, "sharded_execution_for_inference": false, "transformers_version": "4.20.1" }