train-template.yaml

###################################################################################################
# DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
#
# This material is based upon work supported by the Under Secretary of Defense for Research and
# Engineering under Air Force Contract No. FA8702-15-D-0001. Any opinions, findings, conclusions
# or recommendations expressed in this material are those of the author(s) and do not necessarily
# reflect the views of the Under Secretary of Defense for Research and Engineering.
#
# (c) 2020 Massachusetts Institute of Technology.
#
# MIT Proprietary, Subject to FAR52.227-11 Patent Rights - Ownership by the contractor (May 2014)
#
# The software/firmware is provided to you on an As-Is basis
#
# Delivered to the U.S. Government with Unlimited Rights, as defined in DFARS Part 252.227-7013
# or 7014 (Feb 2014). Notwithstanding any copyright notice, U.S. Government rights in this work
# are defined by DFARS 252.227-7013 or DFARS 252.227-7014 as detailed above. Use of this work other
# than as specifically authorized by the U.S. Government may violate any copyrights that exist in
# this work.
###################################################################################################

# Example goseek configuration
# Anything in <BOLD_BRACKETS> needs to be replaced with an appropriate path.

local_dir: "~/ray_results"

env_config: {
        FILENAME: <TESSE_EXECUTABLE_PATH>,
        SCENES: [1, 2, 3, 4],
        episode_length: 400,
        video_log_path: "./videos",
        collision_reward: 0,
        target_found_reward: 1,
        success_dist: 2,
        no_collisions: False,
        modalities: ["RGB_LEFT", "SEGMENTATION", "DEPTH"],
        visible_targets: False,
        use_esdf_data: True,
        esdf_data: {'1': <SCENE_1_ESDF_PATH>,
                    '2': <SCENE_2_ESDF_PATH>,
                    '3': <SCENE_3_ESDF_PATH>,
                    '4': <SCENE_4_ESDF_PATH>,
                    },
        scene_graph: {"1": <SCENE_1_DSG_PATH>,
                      "2": <SCENE_2_DSG_PATH>,
                      "3": <SCENE_3_DSG_PATH>,
                      "4": <SCENE_4_DSG_PATH>,
                    },
    }

model: {
        custom_model: gnn_actor_critic,
        custom_model_config: {
            fields: {
                nodes: GRAPH_NODES,
                edges: GRAPH_EDGES,
                node_shapes: GRAPH_NODE_SHAPE,
                edge_shapes: GRAPH_EDGE_SHAPE,
            },
            in_features: 10,
            graph_conv_features: [64, 64, 64],
            n_layer_nodes: 8,
            embedding_size: [512],
        },
        max_seq_len: 400,
        vf_share_layers : false
    }

# training
lr : 1.0e-4
gamma : 0.99
lambda : .9
vf_loss_coeff : 0.25
clip_param : 0.3
vf_clip_param : 25
entropy_coeff : 0.01

# batch
train_batch_size : 1024 # 4096
sgd_minibatch_size : 128
num_sgd_iter : 30
rollout_fragment_length: 400

# env
env : goseek
framework : torch
num_workers : 0
num_gpus : 1
num_envs_per_worker : 1
num_gpus_per_worker: 0.25
log_level : INFO

ckpt_save_freq_timesteps: 10000

evaluation_interval : 50
evaluation_num_episodes : 50
evaluation_num_workers: 1
evaluation_config: {
    env_config: {
            SCENES: [5],
            RANK: 3,
            FILENAME: <TESSE_EXECUTABLE_PATH>,
            esdf_data: {'5': <SCENE_5_ESDF_PATH>,
            scene_graph: {"5": <SCENE_5_DSG_PATH>,
        }
    }