Upload folder using huggingface_hub
Browse files- .gitattributes +6 -0
- .summary/0/events.out.tfevents.1717635033.4a1f19113fa6 +3 -0
- .summary/0/events.out.tfevents.1717700366.c11948e043d9 +3 -0
- .summary/0/events.out.tfevents.1717702446.c11948e043d9 +3 -0
- .summary/0/events.out.tfevents.1717708092.c11948e043d9 +3 -0
- .summary/0/events.out.tfevents.1717708590.c11948e043d9 +3 -0
- .summary/0/events.out.tfevents.1717710273.c11948e043d9 +3 -0
- .summary/0/events.out.tfevents.1717711462.c11948e043d9 +3 -0
- .summary/0/events.out.tfevents.1717717783.c11948e043d9 +3 -0
- README.md +42 -0
- a100_100x100.mp4 +3 -0
- a20_40x40.mp4 +3 -0
- a20_b4_40x40.mp4 +3 -0
- a20_r4_40x40.mp4 +3 -0
- a5_25x25.mp4 +0 -0
- checkpoint_p0/best_000870321_14259339264_reward_0.370.pth +3 -0
- checkpoint_p0/checkpoint_000971348_15914565632.pth +3 -0
- checkpoint_p0/checkpoint_000971574_15918268416.pth +3 -0
- config.json +161 -0
- git.diff +0 -0
- replay.mp4 +3 -0
- sf_log.txt +3 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,9 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
a100_100x100.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
a20_40x40.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
a20_b4_40x40.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
a20_r4_40x40.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 40 |
+
replay.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 41 |
+
sf_log.txt filter=lfs diff=lfs merge=lfs -text
|
.summary/0/events.out.tfevents.1717635033.4a1f19113fa6
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ee45b727ef73050650719d6370bb442ed773c108c5efa37345fd1a1195f4cae0
|
| 3 |
+
size 2433024
|
.summary/0/events.out.tfevents.1717700366.c11948e043d9
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1cf6cafad438ea9510a7e6f67df1fb9eb1ff981d02fe2e588bfa5af45242fe1e
|
| 3 |
+
size 861005
|
.summary/0/events.out.tfevents.1717702446.c11948e043d9
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8603f0d9093699a29995f4ae331c45933b2d8ebc4a2f416da7a29df9b07ab809
|
| 3 |
+
size 4371047
|
.summary/0/events.out.tfevents.1717708092.c11948e043d9
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ddeda34714771693362020b4a67d629000665abf6347166f03b8b5b7d43765c3
|
| 3 |
+
size 40
|
.summary/0/events.out.tfevents.1717708590.c11948e043d9
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d5918fd3155796d72554d67c6aa10370132f2c23daf769ed383de0d09a63f693
|
| 3 |
+
size 1183336
|
.summary/0/events.out.tfevents.1717710273.c11948e043d9
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f9de5828b5e679da8fd5921ac0b5419e928d328aabb81b0e0d59e7eac9370a41
|
| 3 |
+
size 823211
|
.summary/0/events.out.tfevents.1717711462.c11948e043d9
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8f8af146e539d3252a7c61c5decc896ba431958591cb8755355b2d90419f9966
|
| 3 |
+
size 4548736
|
.summary/0/events.out.tfevents.1717717783.c11948e043d9
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dd6de054b926c2457045c682a95d24d8fc6385270ebc950aa3b2ff65d547f3d8
|
| 3 |
+
size 218032052
|
README.md
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: sample-factory
|
| 3 |
+
tags:
|
| 4 |
+
- deep-reinforcement-learning
|
| 5 |
+
- reinforcement-learning
|
| 6 |
+
- sample-factory
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
A(n) **APPO** model trained on the **GDY-MettaGrid** environment.
|
| 10 |
+
|
| 11 |
+
This model was trained using Sample-Factory 2.0: https://github.com/alex-petrenko/sample-factory.
|
| 12 |
+
Documentation for how to use Sample-Factory can be found at https://www.samplefactory.dev/
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
## Downloading the model
|
| 16 |
+
|
| 17 |
+
After installing Sample-Factory, download the model with:
|
| 18 |
+
```
|
| 19 |
+
python -m sample_factory.huggingface.load_from_hub -r metta-ai/baseline.v0.2.2
|
| 20 |
+
```
|
| 21 |
+
|
| 22 |
+
|
| 23 |
+
## Using the model
|
| 24 |
+
|
| 25 |
+
To run the model after download, use the `enjoy` script corresponding to this environment:
|
| 26 |
+
```
|
| 27 |
+
python -m <path.to.enjoy.module> --algo=APPO --env=GDY-MettaGrid --train_dir=./train_dir --experiment=baseline.v0.2.2
|
| 28 |
+
```
|
| 29 |
+
|
| 30 |
+
|
| 31 |
+
You can also upload models to the Hugging Face Hub using the same script with the `--push_to_hub` flag.
|
| 32 |
+
See https://www.samplefactory.dev/10-huggingface/huggingface/ for more details
|
| 33 |
+
|
| 34 |
+
## Training with this model
|
| 35 |
+
|
| 36 |
+
To continue training with this model, use the `train` script corresponding to this environment:
|
| 37 |
+
```
|
| 38 |
+
python -m <path.to.train.module> --algo=APPO --env=GDY-MettaGrid --train_dir=./train_dir --experiment=baseline.v0.2.2 --restart_behavior=resume --train_for_env_steps=10000000000
|
| 39 |
+
```
|
| 40 |
+
|
| 41 |
+
Note, you may have to adjust `--train_for_env_steps` to a suitably high number as the experiment will resume at the number of steps it concluded at.
|
| 42 |
+
|
a100_100x100.mp4
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dea77db32fda2241d5541d48313bdedb572bcbb994004f10b4b723e205e2e203
|
| 3 |
+
size 9661669
|
a20_40x40.mp4
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7fd1f4b5380956f834d3df0c89e055f5a3584825d0a85f6cfb2c673ff4c18db8
|
| 3 |
+
size 1949240
|
a20_b4_40x40.mp4
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0ac08b8d66f5475591d97b5c754cbedace2be3b94ec148eabbe08a21013ba57f
|
| 3 |
+
size 1271422
|
a20_r4_40x40.mp4
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8f43b88b0a42e372ea1291e4a67c5a030444c0044118d6ff15e7dbbe04d4342a
|
| 3 |
+
size 2273096
|
a5_25x25.mp4
ADDED
|
Binary file (892 kB). View file
|
|
|
checkpoint_p0/best_000870321_14259339264_reward_0.370.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f5e1a05f7aaf6d4eda8331fd9a9742e5a1e223ed6ba2118cf6f3eb7630b90360
|
| 3 |
+
size 38973114
|
checkpoint_p0/checkpoint_000971348_15914565632.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:342f8d8f91bd2e597164eb4690a493a057fe1e8c079f38f584e061e04f9b37ab
|
| 3 |
+
size 38974422
|
checkpoint_p0/checkpoint_000971574_15918268416.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2615ca4ce26eb958d7a1258895fe25193e4d539a8e91d769af0c6a9e8ecfe89a
|
| 3 |
+
size 38974422
|
config.json
ADDED
|
@@ -0,0 +1,161 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"help": false,
|
| 3 |
+
"algo": "APPO",
|
| 4 |
+
"env": "GDY-MettaGrid",
|
| 5 |
+
"experiment": "p2.metta.4",
|
| 6 |
+
"train_dir": "/workspace/metta/train_dir",
|
| 7 |
+
"restart_behavior": "resume",
|
| 8 |
+
"device": "gpu",
|
| 9 |
+
"seed": 0,
|
| 10 |
+
"num_policies": 1,
|
| 11 |
+
"async_rl": true,
|
| 12 |
+
"serial_mode": false,
|
| 13 |
+
"batched_sampling": false,
|
| 14 |
+
"num_batches_to_accumulate": 2,
|
| 15 |
+
"worker_num_splits": 2,
|
| 16 |
+
"policy_workers_per_policy": 1,
|
| 17 |
+
"max_policy_lag": 50,
|
| 18 |
+
"num_workers": 32,
|
| 19 |
+
"num_envs_per_worker": 2,
|
| 20 |
+
"batch_size": 16384,
|
| 21 |
+
"num_batches_per_epoch": 1,
|
| 22 |
+
"num_epochs": 1,
|
| 23 |
+
"rollout": 256,
|
| 24 |
+
"recurrence": 256,
|
| 25 |
+
"shuffle_minibatches": false,
|
| 26 |
+
"gamma": 0.99,
|
| 27 |
+
"reward_scale": 1.0,
|
| 28 |
+
"reward_clip": 1000.0,
|
| 29 |
+
"value_bootstrap": false,
|
| 30 |
+
"normalize_returns": true,
|
| 31 |
+
"exploration_loss_coeff": 0.002,
|
| 32 |
+
"value_loss_coeff": 0.976,
|
| 33 |
+
"kl_loss_coeff": 0.0,
|
| 34 |
+
"aux_loss_coeff": 0.0,
|
| 35 |
+
"exploration_loss": "symmetric_kl",
|
| 36 |
+
"gae_lambda": 0.95,
|
| 37 |
+
"ppo_clip_ratio": 0.1,
|
| 38 |
+
"ppo_clip_value": 1.0,
|
| 39 |
+
"with_vtrace": false,
|
| 40 |
+
"vtrace_rho": 1.0,
|
| 41 |
+
"vtrace_c": 1.0,
|
| 42 |
+
"optimizer": "adam",
|
| 43 |
+
"adam_eps": 1e-06,
|
| 44 |
+
"adam_beta1": 0.9,
|
| 45 |
+
"adam_beta2": 0.999,
|
| 46 |
+
"max_grad_norm": 4.0,
|
| 47 |
+
"learning_rate": 0.0001,
|
| 48 |
+
"lr_schedule": "constant",
|
| 49 |
+
"lr_schedule_kl_threshold": 0.008,
|
| 50 |
+
"lr_adaptive_min": 1e-06,
|
| 51 |
+
"lr_adaptive_max": 0.01,
|
| 52 |
+
"obs_subtract_mean": 0.0,
|
| 53 |
+
"obs_scale": 1.0,
|
| 54 |
+
"normalize_input": false,
|
| 55 |
+
"normalize_input_keys": null,
|
| 56 |
+
"decorrelate_experience_max_seconds": 150,
|
| 57 |
+
"decorrelate_envs_on_one_worker": true,
|
| 58 |
+
"actor_worker_gpus": [],
|
| 59 |
+
"set_workers_cpu_affinity": true,
|
| 60 |
+
"force_envs_single_thread": false,
|
| 61 |
+
"default_niceness": 0,
|
| 62 |
+
"log_to_file": true,
|
| 63 |
+
"experiment_summaries_interval": 10,
|
| 64 |
+
"flush_summaries_interval": 30,
|
| 65 |
+
"stats_avg": 100,
|
| 66 |
+
"summaries_use_frameskip": true,
|
| 67 |
+
"heartbeat_interval": 20,
|
| 68 |
+
"heartbeat_reporting_interval": 180,
|
| 69 |
+
"train_for_env_steps": 9999999999999,
|
| 70 |
+
"train_for_seconds": 10000000000,
|
| 71 |
+
"save_every_sec": 120,
|
| 72 |
+
"keep_checkpoints": 2,
|
| 73 |
+
"load_checkpoint_path": null,
|
| 74 |
+
"init_checkpoint_path": null,
|
| 75 |
+
"load_checkpoint_kind": "latest",
|
| 76 |
+
"save_milestones_sec": -1,
|
| 77 |
+
"save_best_every_sec": 5,
|
| 78 |
+
"save_best_metric": "reward",
|
| 79 |
+
"save_best_after": 100000,
|
| 80 |
+
"benchmark": false,
|
| 81 |
+
"encoder_mlp_layers": [
|
| 82 |
+
512,
|
| 83 |
+
512
|
| 84 |
+
],
|
| 85 |
+
"encoder_conv_architecture": "convnet_simple",
|
| 86 |
+
"encoder_conv_mlp_layers": [
|
| 87 |
+
512
|
| 88 |
+
],
|
| 89 |
+
"use_rnn": true,
|
| 90 |
+
"rnn_size": 512,
|
| 91 |
+
"rnn_type": "gru",
|
| 92 |
+
"rnn_num_layers": 1,
|
| 93 |
+
"decoder_mlp_layers": [],
|
| 94 |
+
"nonlinearity": "elu",
|
| 95 |
+
"policy_initialization": "orthogonal",
|
| 96 |
+
"policy_init_gain": 1.0,
|
| 97 |
+
"actor_critic_share_weights": true,
|
| 98 |
+
"adaptive_stddev": true,
|
| 99 |
+
"continuous_tanh_scale": 0.0,
|
| 100 |
+
"initial_stddev": 1.0,
|
| 101 |
+
"use_env_info_cache": false,
|
| 102 |
+
"env_gpu_actions": false,
|
| 103 |
+
"env_gpu_observations": true,
|
| 104 |
+
"env_frameskip": 1,
|
| 105 |
+
"env_framestack": 1,
|
| 106 |
+
"pixel_format": "CHW",
|
| 107 |
+
"use_record_episode_statistics": false,
|
| 108 |
+
"episode_counter": false,
|
| 109 |
+
"with_wandb": true,
|
| 110 |
+
"wandb_user": "platypus",
|
| 111 |
+
"wandb_project": "metta",
|
| 112 |
+
"wandb_group": null,
|
| 113 |
+
"wandb_job_type": "SF",
|
| 114 |
+
"wandb_tags": [],
|
| 115 |
+
"with_pbt": false,
|
| 116 |
+
"pbt_mix_policies_in_one_env": true,
|
| 117 |
+
"pbt_period_env_steps": 5000000,
|
| 118 |
+
"pbt_start_mutation": 20000000,
|
| 119 |
+
"pbt_replace_fraction": 0.3,
|
| 120 |
+
"pbt_mutation_rate": 0.15,
|
| 121 |
+
"pbt_replace_reward_gap": 0.1,
|
| 122 |
+
"pbt_replace_reward_gap_absolute": 1e-06,
|
| 123 |
+
"pbt_optimize_gamma": false,
|
| 124 |
+
"pbt_target_objective": "true_objective",
|
| 125 |
+
"pbt_perturb_min": 1.1,
|
| 126 |
+
"pbt_perturb_max": 1.5,
|
| 127 |
+
"env_cfg": "{\"name\": \"GDY-MettaGrid\", \"_target_\": \"env.griddly.mettagrid.gym_env.MettaGridGymEnv\", \"max_action_value\": 10, \"game_builder\": {\"_target_\": \"env.griddly.mettagrid.game_builder.MettaGridGameBuilder\", \"obs_width\": 11, \"obs_height\": 11, \"max_steps\": 1000, \"tile_size\": 16, \"num_agents\": 20, \"objects\": {\"agent\": {\"initial_energy\": 100, \"max_energy\": 500, \"max_inventory\": 5, \"freeze_duration\": 10, \"hp\": 1, \"upkeep\": {\"shield\": 1}}, \"altar\": {\"hp\": 30, \"cooldown\": 2, \"cost\": 100}, \"converter\": {\"hp\": 30, \"cooldown\": 2, \"energy_output\": 50, \"cost\": 0}, \"generator\": {\"hp\": 30, \"cooldown\": 5, \"initial_resources\": 30, \"cost\": 0}, \"wall\": {\"density\": 0.01, \"hp\": 10}}, \"actions\": {\"move\": {\"cost\": 0}, \"rotate\": {\"cost\": 0}, \"jump\": {\"cost\": 3}, \"shield\": {\"cost\": 0}, \"drop\": {\"cost\": 1}, \"use\": {\"cost\": 0}, \"attack\": {\"cost\": 5, \"damage\": 5}}, \"map\": {\"layout\": [[\"base\", \"wild_1\", \"base\"], [\"wild_2\", \"center\", \"wild_2\"], [\"base\", \"wild_1\", \"base\"]], \"room\": {\"width\": 25, \"height\": 25, \"num_agents\": 5, \"objects\": {\"agent\": 5, \"altar\": 1, \"converter\": 3, \"generator\": 15, \"wall\": 10}}, \"wild_1\": {\"width\": 10, \"height\": 15, \"border\": 0, \"objects\": {\"agent\": 0, \"altar\": 1, \"converter\": 1, \"generator\": 5, \"wall\": 5}}, \"wild_2\": {\"width\": 15, \"height\": 10, \"border\": 0, \"objects\": {\"agent\": 0, \"altar\": 1, \"converter\": 1, \"generator\": 5, \"wall\": 5}}, \"center\": {\"width\": 10, \"height\": 10, \"border\": 0, \"objects\": {\"agent\": 0, \"altar\": 2, \"converter\": 5, \"generator\": 10, \"wall\": 20}}, \"base\": {\"width\": 15, \"height\": 15, \"border\": 1, \"objects\": {\"agent\": 5, \"altar\": 1, \"converter\": 3, \"generator\": 5, \"wall\": 5}}}}}",
|
| 128 |
+
"agent_cfg": "{\"_target_\": \"agent.metta_agent.MettaAgent\", \"observation_encoders\": {\"grid_obs\": {\"feature_names\": [], \"normalize_features\": true, \"label_dim\": 4, \"output_dim\": 512, \"layers\": 4}, \"global_vars\": {\"feature_names\": [], \"normalize_features\": true, \"label_dim\": 4, \"output_dim\": 8, \"layers\": 2}, \"last_action\": {\"feature_names\": [\"last_action_id\", \"last_action_val\"], \"normalize_features\": true, \"label_dim\": 4, \"output_dim\": 8, \"layers\": 2}, \"last_reward\": {\"feature_names\": [\"last_reward\"], \"normalize_features\": true, \"label_dim\": 4, \"output_dim\": 8, \"layers\": 2}}, \"fc\": {\"layers\": 3, \"output_dim\": 512}, \"decoder\": {\"_target_\": \"agent.decoder.Decoder\"}, \"core\": {\"_target_\": \"sample_factory.model.core.ModelCoreRNN\", \"rnn_type\": \"gru\", \"rnn_num_layers\": 1, \"rnn_size\": 512}}",
|
| 129 |
+
"command_line": "--aux_loss_coef=0 --recurrence=256 --rollout=256 --value_loss_coeff=0.976 --exploration_loss=symmetric_kl --exploration_loss_coeff=0.002 --policy_initialization=orthogonal --learning_rate=0.0001 --max_policy_lag=50 --nonlinearity=elu --load_checkpoint_kind=latest --normalize_input=False --seed=0 --batch_size=16384 --decorrelate_experience_max_seconds=150 --train_for_env_steps=9999999999999 --with_wandb=True --wandb_user=platypus --wandb_project=metta --experiment=p2.metta.4 --rnn_type=gru --rnn_num_layers=1 --rnn_size=512 --env=GDY-MettaGrid --env_cfg={\"name\": \"GDY-MettaGrid\", \"_target_\": \"env.griddly.mettagrid.gym_env.MettaGridGymEnv\", \"game_builder\": {\"_target_\": \"env.griddly.mettagrid.game_builder.MettaGridGameBuilder\", \"obs_width\": 11, \"obs_height\": 11, \"max_steps\": 1000, \"tile_size\": 16, \"width\": 40, \"height\": 40, \"objects\": {\"agent\": {\"count\": 20, \"initial_energy\": 100, \"max_energy\": 500, \"max_inventory\": 5, \"freeze_duration\": 10, \"hp\": 1, \"upkeep\": {\"shield\": 1}}, \"altar\": {\"hp\": 30, \"count\": 5, \"cooldown\": 2, \"cost\": 100}, \"converter\": {\"hp\": 30, \"count\": 10, \"cooldown\": 2, \"energy_output\": 50, \"cost\": 0}, \"generator\": {\"hp\": 30, \"count\": 50, \"cooldown\": 50, \"initial_resources\": 100, \"cost\": 0}, \"wall\": {\"density\": 0.01, \"hp\": 10}, \"charger\": {\"count\": 10}}, \"actions\": {\"move\": {\"cost\": 0}, \"rotate\": {\"cost\": 0}, \"jump\": {\"cost\": 3}, \"shield\": {\"cost\": 0}, \"drop\": {\"cost\": 1}, \"use\": {\"cost\": 0}, \"attack\": {\"cost\": 5, \"damage\": 5}}}, \"global_features\": [\"conf:agent:energy:initial\", \"conf:agent:energy:max\", \"conf:altar:cooldown\", \"conf:altar:cost\", \"conf:attack:damage\", \"conf:attack:freeze_duration\", \"conf:charger:cooldown\", \"conf:charger:energy\", \"conf:cost:attack\", \"conf:cost:frozen\", \"conf:cost:jump\", \"conf:cost:move\", \"conf:cost:rotate\", \"conf:cost:shield:upkeep\", \"conf:cost:shield\", \"conf:generator:cooldown\", \"conf:gift:energy\"], \"grid_features\": [\"agent\", \"altar\", \"charger\", \"generator\", \"wall\", \"agent:dir\", \"agent:energy\", \"agent:frozen\", \"agent:id\", \"agent:inv:1\", \"agent:inv:2\", \"agent:inv:3\", \"agent:shield\", \"agent:species\", \"altar:ready\", \"charger:bonus\", \"charger:input:1\", \"charger:input:2\", \"charger:input:3\", \"charger:output\", \"charger:ready\", \"generator:ready\", \"generator:resource\"], \"actions\": [\"rotate\", \"move\", \"jump\", \"use\", \"gift\", \"shield\", \"attack\"], \"max_action_value\": 10} --agent_cfg={\"_target_\": \"agent.metta_agent.MettaAgent\", \"observation_encoders\": {\"grid_obs\": {\"feature_names\": [], \"normalize_features\": true, \"label_dim\": 4, \"output_dim\": 512, \"layers\": 4}, \"global_vars\": {\"feature_names\": [], \"normalize_features\": true, \"label_dim\": 4, \"output_dim\": 8, \"layers\": 2}, \"last_action\": {\"feature_names\": [\"last_action_id\", \"last_action_val\"], \"normalize_features\": true, \"label_dim\": 4, \"output_dim\": 8, \"layers\": 2}, \"last_reward\": {\"feature_names\": [\"last_reward\"], \"normalize_features\": true, \"label_dim\": 4, \"output_dim\": 8, \"layers\": 2}}, \"fc\": {\"layers\": 3, \"output_dim\": 512}, \"decoder\": {\"_target_\": \"agent.decoder.Decoder\"}, \"core\": {\"_target_\": \"sample_factory.model.core.ModelCoreRNN\", \"rnn_type\": \"gru\", \"rnn_num_layers\": 1, \"rnn_size\": 512}}",
|
| 130 |
+
"cli_args": {
|
| 131 |
+
"env": "GDY-MettaGrid",
|
| 132 |
+
"experiment": "p2.metta.4",
|
| 133 |
+
"seed": 0,
|
| 134 |
+
"max_policy_lag": 50,
|
| 135 |
+
"batch_size": 16384,
|
| 136 |
+
"rollout": 256,
|
| 137 |
+
"recurrence": 256,
|
| 138 |
+
"exploration_loss_coeff": 0.002,
|
| 139 |
+
"value_loss_coeff": 0.976,
|
| 140 |
+
"aux_loss_coeff": 0.0,
|
| 141 |
+
"exploration_loss": "symmetric_kl",
|
| 142 |
+
"learning_rate": 0.0001,
|
| 143 |
+
"normalize_input": false,
|
| 144 |
+
"decorrelate_experience_max_seconds": 150,
|
| 145 |
+
"train_for_env_steps": 9999999999999,
|
| 146 |
+
"load_checkpoint_kind": "latest",
|
| 147 |
+
"rnn_size": 512,
|
| 148 |
+
"rnn_type": "gru",
|
| 149 |
+
"rnn_num_layers": 1,
|
| 150 |
+
"nonlinearity": "elu",
|
| 151 |
+
"policy_initialization": "orthogonal",
|
| 152 |
+
"with_wandb": true,
|
| 153 |
+
"wandb_user": "platypus",
|
| 154 |
+
"wandb_project": "metta",
|
| 155 |
+
"env_cfg": "{\"name\": \"GDY-MettaGrid\", \"_target_\": \"env.griddly.mettagrid.gym_env.MettaGridGymEnv\", \"game_builder\": {\"_target_\": \"env.griddly.mettagrid.game_builder.MettaGridGameBuilder\", \"obs_width\": 11, \"obs_height\": 11, \"max_steps\": 1000, \"tile_size\": 16, \"width\": 40, \"height\": 40, \"objects\": {\"agent\": {\"count\": 20, \"initial_energy\": 100, \"max_energy\": 500, \"max_inventory\": 5, \"freeze_duration\": 10, \"hp\": 1, \"upkeep\": {\"shield\": 1}}, \"altar\": {\"hp\": 30, \"count\": 5, \"cooldown\": 2, \"cost\": 100}, \"converter\": {\"hp\": 30, \"count\": 10, \"cooldown\": 2, \"energy_output\": 50, \"cost\": 0}, \"generator\": {\"hp\": 30, \"count\": 50, \"cooldown\": 50, \"initial_resources\": 100, \"cost\": 0}, \"wall\": {\"density\": 0.01, \"hp\": 10}, \"charger\": {\"count\": 10}}, \"actions\": {\"move\": {\"cost\": 0}, \"rotate\": {\"cost\": 0}, \"jump\": {\"cost\": 3}, \"shield\": {\"cost\": 0}, \"drop\": {\"cost\": 1}, \"use\": {\"cost\": 0}, \"attack\": {\"cost\": 5, \"damage\": 5}}}, \"global_features\": [\"conf:agent:energy:initial\", \"conf:agent:energy:max\", \"conf:altar:cooldown\", \"conf:altar:cost\", \"conf:attack:damage\", \"conf:attack:freeze_duration\", \"conf:charger:cooldown\", \"conf:charger:energy\", \"conf:cost:attack\", \"conf:cost:frozen\", \"conf:cost:jump\", \"conf:cost:move\", \"conf:cost:rotate\", \"conf:cost:shield:upkeep\", \"conf:cost:shield\", \"conf:generator:cooldown\", \"conf:gift:energy\"], \"grid_features\": [\"agent\", \"altar\", \"charger\", \"generator\", \"wall\", \"agent:dir\", \"agent:energy\", \"agent:frozen\", \"agent:id\", \"agent:inv:1\", \"agent:inv:2\", \"agent:inv:3\", \"agent:shield\", \"agent:species\", \"altar:ready\", \"charger:bonus\", \"charger:input:1\", \"charger:input:2\", \"charger:input:3\", \"charger:output\", \"charger:ready\", \"generator:ready\", \"generator:resource\"], \"actions\": [\"rotate\", \"move\", \"jump\", \"use\", \"gift\", \"shield\", \"attack\"], \"max_action_value\": 10}",
|
| 156 |
+
"agent_cfg": "{\"_target_\": \"agent.metta_agent.MettaAgent\", \"observation_encoders\": {\"grid_obs\": {\"feature_names\": [], \"normalize_features\": true, \"label_dim\": 4, \"output_dim\": 512, \"layers\": 4}, \"global_vars\": {\"feature_names\": [], \"normalize_features\": true, \"label_dim\": 4, \"output_dim\": 8, \"layers\": 2}, \"last_action\": {\"feature_names\": [\"last_action_id\", \"last_action_val\"], \"normalize_features\": true, \"label_dim\": 4, \"output_dim\": 8, \"layers\": 2}, \"last_reward\": {\"feature_names\": [\"last_reward\"], \"normalize_features\": true, \"label_dim\": 4, \"output_dim\": 8, \"layers\": 2}}, \"fc\": {\"layers\": 3, \"output_dim\": 512}, \"decoder\": {\"_target_\": \"agent.decoder.Decoder\"}, \"core\": {\"_target_\": \"sample_factory.model.core.ModelCoreRNN\", \"rnn_type\": \"gru\", \"rnn_num_layers\": 1, \"rnn_size\": 512}}"
|
| 157 |
+
},
|
| 158 |
+
"git_hash": "7864b034e8faa45a8cef257cdd0b4d3b2b2549f3",
|
| 159 |
+
"git_repo_name": "https://github.com/daveey/metta.git",
|
| 160 |
+
"wandb_unique_id": "p2.metta.4_20240605_175031_058004"
|
| 161 |
+
}
|
git.diff
ADDED
|
File without changes
|
replay.mp4
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7fd1f4b5380956f834d3df0c89e055f5a3584825d0a85f6cfb2c673ff4c18db8
|
| 3 |
+
size 1949240
|
sf_log.txt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4e251b8a808735267249b108e57de0871afa1a4236625da2881391376d8fa29e
|
| 3 |
+
size 31982556
|