MattStammers
commited on
Commit
•
e1f84e9
1
Parent(s):
f9934e5
Upload folder using huggingface_hub
Browse files- .gitattributes +2 -0
- .summary/0/events.out.tfevents.1696747807.rhmmedcatt-proliant-ml350-gen10 +3 -0
- .summary/1/events.out.tfevents.1696747807.rhmmedcatt-proliant-ml350-gen10 +3 -0
- README.md +169 -0
- checkpoint_p0/best_000091072_93257728_reward_40.890.pth +3 -0
- checkpoint_p0/checkpoint_000097792_100139008.pth +3 -0
- checkpoint_p0/checkpoint_000098144_100499456.pth +3 -0
- checkpoint_p0/milestones/checkpoint_000008416_8617984.pth +3 -0
- checkpoint_p0/milestones/checkpoint_000016960_17367040.pth +3 -0
- checkpoint_p0/milestones/checkpoint_000025568_26181632.pth +3 -0
- checkpoint_p0/milestones/checkpoint_000034176_34996224.pth +3 -0
- checkpoint_p0/milestones/checkpoint_000042784_43810816.pth +3 -0
- checkpoint_p0/milestones/checkpoint_000051328_52559872.pth +3 -0
- checkpoint_p0/milestones/checkpoint_000059936_61374464.pth +3 -0
- checkpoint_p0/milestones/checkpoint_000068512_70156288.pth +3 -0
- checkpoint_p0/milestones/checkpoint_000077120_78970880.pth +3 -0
- checkpoint_p0/milestones/checkpoint_000085760_87818240.pth +3 -0
- checkpoint_p0/milestones/checkpoint_000094336_96600064.pth +3 -0
- checkpoint_p1/best_000097280_99614720_reward_44.650.pth +3 -0
- checkpoint_p1/checkpoint_000097312_99647488.pth +3 -0
- checkpoint_p1/checkpoint_000097664_100007936.pth +3 -0
- checkpoint_p1/milestones/checkpoint_000008352_8552448.pth +3 -0
- checkpoint_p1/milestones/checkpoint_000016896_17301504.pth +3 -0
- checkpoint_p1/milestones/checkpoint_000025440_26050560.pth +3 -0
- checkpoint_p1/milestones/checkpoint_000034016_34832384.pth +3 -0
- checkpoint_p1/milestones/checkpoint_000042560_43581440.pth +3 -0
- checkpoint_p1/milestones/checkpoint_000051104_52330496.pth +3 -0
- checkpoint_p1/milestones/checkpoint_000059648_61079552.pth +3 -0
- checkpoint_p1/milestones/checkpoint_000068224_69861376.pth +3 -0
- checkpoint_p1/milestones/checkpoint_000076768_78610432.pth +3 -0
- checkpoint_p1/milestones/checkpoint_000085344_87392256.pth +3 -0
- checkpoint_p1/milestones/checkpoint_000093888_96141312.pth +3 -0
- config.json +162 -0
- git.diff +3 -0
- replay.mp4 +3 -0
- sf_log.txt +0 -0
.gitattributes
CHANGED
@@ -33,3 +33,5 @@ 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 |
+
git.diff filter=lfs diff=lfs merge=lfs -text
|
37 |
+
replay.mp4 filter=lfs diff=lfs merge=lfs -text
|
.summary/0/events.out.tfevents.1696747807.rhmmedcatt-proliant-ml350-gen10
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8e866594c9dc30e926418e722a3a4fd6eca8ae6b3f2c35b8e04256e03030678e
|
3 |
+
size 7528010
|
.summary/1/events.out.tfevents.1696747807.rhmmedcatt-proliant-ml350-gen10
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:71670cbfbeef5e86d88d93744d932af9869915a226995cbb79e8db6033d4bdbe
|
3 |
+
size 4355990
|
README.md
ADDED
@@ -0,0 +1,169 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
library_name: sample-factory
|
3 |
+
tags:
|
4 |
+
- deep-reinforcement-learning
|
5 |
+
- reinforcement-learning
|
6 |
+
- sample-factory
|
7 |
+
model-index:
|
8 |
+
- name: APPO
|
9 |
+
results:
|
10 |
+
- task:
|
11 |
+
type: reinforcement-learning
|
12 |
+
name: reinforcement-learning
|
13 |
+
dataset:
|
14 |
+
name: atari_asterix
|
15 |
+
type: atari_asterix
|
16 |
+
metrics:
|
17 |
+
- type: mean_reward
|
18 |
+
value: 14800.00 +/- 10461.83
|
19 |
+
name: mean_reward
|
20 |
+
verified: false
|
21 |
+
---
|
22 |
+
|
23 |
+
A(n) **APPO** model trained on the **atari_asterix** environment.
|
24 |
+
|
25 |
+
This model was trained using Sample-Factory 2.0: https://github.com/alex-petrenko/sample-factory.
|
26 |
+
Documentation for how to use Sample-Factory can be found at https://www.samplefactory.dev/
|
27 |
+
|
28 |
+
|
29 |
+
## Downloading the model
|
30 |
+
|
31 |
+
After installing Sample-Factory, download the model with:
|
32 |
+
```
|
33 |
+
python -m sample_factory.huggingface.load_from_hub -r MattStammers/APPO-atari_asterix
|
34 |
+
```
|
35 |
+
|
36 |
+
|
37 |
+
## About the Model
|
38 |
+
|
39 |
+
This model as with all the others in the benchmarks was trained initially asynchronously un-seeded to 10 million steps for the purposes of setting a sample factory async baseline for this model on this environment but only 3/57 made it.
|
40 |
+
|
41 |
+
The aim is to reach state-of-the-art (SOTA) performance on each atari environment. I will flag the models with SOTA when they reach at or near these levels.
|
42 |
+
|
43 |
+
The hyperparameters used in the model are the ones I have pushed to my fork of sample-factory: https://github.com/MattStammers/sample-factory. Given that https://huggingface.co/edbeeching has kindly shared his.
|
44 |
+
I saved time and energy by using many of his tuned hyperparameters to maximise performance. However, he used 2 billion training steps. I have started as explained above at 10 million then moved to 100m to see how performance goes:
|
45 |
+
```
|
46 |
+
hyperparameters = {
|
47 |
+
"device": "gpu",
|
48 |
+
"seed": 1234,
|
49 |
+
"num_policies": 2,
|
50 |
+
"async_rl": true,
|
51 |
+
"serial_mode": false,
|
52 |
+
"batched_sampling": true,
|
53 |
+
"num_batches_to_accumulate": 2,
|
54 |
+
"worker_num_splits": 1,
|
55 |
+
"policy_workers_per_policy": 1,
|
56 |
+
"max_policy_lag": 1000,
|
57 |
+
"num_workers": 16,
|
58 |
+
"num_envs_per_worker": 2,
|
59 |
+
"batch_size": 1024,
|
60 |
+
"num_batches_per_epoch": 8,
|
61 |
+
"num_epochs": 4,
|
62 |
+
"rollout": 128,
|
63 |
+
"recurrence": 1,
|
64 |
+
"shuffle_minibatches": false,
|
65 |
+
"gamma": 0.99,
|
66 |
+
"reward_scale": 1.0,
|
67 |
+
"reward_clip": 1000.0,
|
68 |
+
"value_bootstrap": false,
|
69 |
+
"normalize_returns": true,
|
70 |
+
"exploration_loss_coeff": 0.0004677351413,
|
71 |
+
"value_loss_coeff": 0.5,
|
72 |
+
"kl_loss_coeff": 0.0,
|
73 |
+
"exploration_loss": "entropy",
|
74 |
+
"gae_lambda": 0.95,
|
75 |
+
"ppo_clip_ratio": 0.1,
|
76 |
+
"ppo_clip_value": 1.0,
|
77 |
+
"with_vtrace": false,
|
78 |
+
"vtrace_rho": 1.0,
|
79 |
+
"vtrace_c": 1.0,
|
80 |
+
"optimizer": "adam",
|
81 |
+
"adam_eps": 1e-05,
|
82 |
+
"adam_beta1": 0.9,
|
83 |
+
"adam_beta2": 0.999,
|
84 |
+
"max_grad_norm": 0.0,
|
85 |
+
"learning_rate": 0.0003033891184,
|
86 |
+
"lr_schedule": "linear_decay",
|
87 |
+
"lr_schedule_kl_threshold": 0.008,
|
88 |
+
"lr_adaptive_min": 1e-06,
|
89 |
+
"lr_adaptive_max": 0.01,
|
90 |
+
"obs_subtract_mean": 0.0,
|
91 |
+
"obs_scale": 255.0,
|
92 |
+
"normalize_input": true,
|
93 |
+
"normalize_input_keys": [
|
94 |
+
"obs"
|
95 |
+
],
|
96 |
+
"decorrelate_experience_max_seconds": 0,
|
97 |
+
"decorrelate_envs_on_one_worker": true,
|
98 |
+
"actor_worker_gpus": [],
|
99 |
+
"set_workers_cpu_affinity": true,
|
100 |
+
"force_envs_single_thread": false,
|
101 |
+
"default_niceness": 0,
|
102 |
+
"log_to_file": true,
|
103 |
+
"experiment_summaries_interval": 3,
|
104 |
+
"flush_summaries_interval": 30,
|
105 |
+
"stats_avg": 100,
|
106 |
+
"summaries_use_frameskip": true,
|
107 |
+
"heartbeat_interval": 10,
|
108 |
+
"heartbeat_reporting_interval": 60,
|
109 |
+
"train_for_env_steps": 100000000,
|
110 |
+
"train_for_seconds": 10000000000,
|
111 |
+
"save_every_sec": 120,
|
112 |
+
"keep_checkpoints": 2,
|
113 |
+
"load_checkpoint_kind": "latest",
|
114 |
+
"save_milestones_sec": 1200,
|
115 |
+
"save_best_every_sec": 5,
|
116 |
+
"save_best_metric": "reward",
|
117 |
+
"save_best_after": 100000,
|
118 |
+
"benchmark": false,
|
119 |
+
"encoder_mlp_layers": [
|
120 |
+
512,
|
121 |
+
512
|
122 |
+
],
|
123 |
+
"encoder_conv_architecture": "convnet_atari",
|
124 |
+
"encoder_conv_mlp_layers": [
|
125 |
+
512
|
126 |
+
],
|
127 |
+
"use_rnn": false,
|
128 |
+
"rnn_size": 512,
|
129 |
+
"rnn_type": "gru",
|
130 |
+
"rnn_num_layers": 1,
|
131 |
+
"decoder_mlp_layers": [],
|
132 |
+
"nonlinearity": "relu",
|
133 |
+
"policy_initialization": "orthogonal",
|
134 |
+
"policy_init_gain": 1.0,
|
135 |
+
"actor_critic_share_weights": true,
|
136 |
+
"adaptive_stddev": false,
|
137 |
+
"continuous_tanh_scale": 0.0,
|
138 |
+
"initial_stddev": 1.0,
|
139 |
+
"use_env_info_cache": false,
|
140 |
+
"env_gpu_actions": false,
|
141 |
+
"env_gpu_observations": true,
|
142 |
+
"env_frameskip": 4,
|
143 |
+
"env_framestack": 4,
|
144 |
+
}
|
145 |
+
|
146 |
+
```
|
147 |
+
|
148 |
+
|
149 |
+
|
150 |
+
## Using the model
|
151 |
+
|
152 |
+
To run the model after download, use the `enjoy` script corresponding to this environment:
|
153 |
+
```
|
154 |
+
python -m sf_examples.atari.enjoy_atari --algo=APPO --env=atari_asterix --train_dir=./train_dir --experiment=APPO-atari_asterix
|
155 |
+
```
|
156 |
+
|
157 |
+
|
158 |
+
You can also upload models to the Hugging Face Hub using the same script with the `--push_to_hub` flag.
|
159 |
+
See https://www.samplefactory.dev/10-huggingface/huggingface/ for more details
|
160 |
+
|
161 |
+
## Training with this model
|
162 |
+
|
163 |
+
To continue training with this model, use the `train` script corresponding to this environment:
|
164 |
+
```
|
165 |
+
python -m sf_examples.atari.train_atari --algo=APPO --env=atari_asterix --train_dir=./train_dir --experiment=APPO-atari_asterix --restart_behavior=resume --train_for_env_steps=10000000000
|
166 |
+
```
|
167 |
+
|
168 |
+
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.
|
169 |
+
|
checkpoint_p0/best_000091072_93257728_reward_40.890.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:62d0d8b26a75bf9f375d6c9e0cd03d5e430f33d9a1fade6b2aa313876091456b
|
3 |
+
size 20740275
|
checkpoint_p0/checkpoint_000097792_100139008.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:04bd1640ae685f668c143c2474f9764a8f3245fb493b902ad67e2f1617b92557
|
3 |
+
size 20740611
|
checkpoint_p0/checkpoint_000098144_100499456.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:56d8fe9de9336e5cd869563a7f403be4a31c5c67dab0af5aecf52b70754472a6
|
3 |
+
size 20740611
|
checkpoint_p0/milestones/checkpoint_000008416_8617984.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:22841dee48b300dc96a4847bb5d345c70e899bcdc763e3861e24c763301b853e
|
3 |
+
size 20741467
|
checkpoint_p0/milestones/checkpoint_000016960_17367040.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7dd005a02db925d77c74341241df916ac1270512779e3247b6643f0c512f2f75
|
3 |
+
size 20741523
|
checkpoint_p0/milestones/checkpoint_000025568_26181632.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:11e1fcea59cc8fee2c3685ebd3ff7e759362981f3de3279105cd888ef5ebf1cf
|
3 |
+
size 20741523
|
checkpoint_p0/milestones/checkpoint_000034176_34996224.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f6f002c4a62882ee197f69fac8c8b253de4039d90c5fec6eaf891958479b6f1f
|
3 |
+
size 20741523
|
checkpoint_p0/milestones/checkpoint_000042784_43810816.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b94dd1e705d9f621acda45798cadde8f31b089d15e12d5e223a149e72709b970
|
3 |
+
size 20741523
|
checkpoint_p0/milestones/checkpoint_000051328_52559872.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e0c232ba29be5351a623cc817ed9ef93eddce104ac5141e2ab1821c9e204f8ad
|
3 |
+
size 20741523
|
checkpoint_p0/milestones/checkpoint_000059936_61374464.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1a9268f9255cecfecb79232d4463ae1208124fa7c658498fb41d72abe79c9cc6
|
3 |
+
size 20741523
|
checkpoint_p0/milestones/checkpoint_000068512_70156288.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9c66f0c662ed6ce1a95758ea461d5f15a0081191f37728a0547ccfd493f386cd
|
3 |
+
size 20741523
|
checkpoint_p0/milestones/checkpoint_000077120_78970880.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:da6e20c37cd51346a0a9a43a7625b5794fc4474497a973428dbef6d72aafc8c4
|
3 |
+
size 20741523
|
checkpoint_p0/milestones/checkpoint_000085760_87818240.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2513fb1f79fc508b346ee39d81fd206435e353fa2b6c0646c873c42ef04859cd
|
3 |
+
size 20741523
|
checkpoint_p0/milestones/checkpoint_000094336_96600064.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:863075962b0bc7143b3211148e8f7aeee2e2f147a71f01095fcab1736bed6fdd
|
3 |
+
size 20741523
|
checkpoint_p1/best_000097280_99614720_reward_44.650.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5dbda331cc752f528237e62e4edee4944c7edec283b8ad9ee8592aa46458e113
|
3 |
+
size 20740275
|
checkpoint_p1/checkpoint_000097312_99647488.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:02b515df11f44b090cc917fe1a472a26b7738fcc39ae9482b06df4541b2abc15
|
3 |
+
size 20740611
|
checkpoint_p1/checkpoint_000097664_100007936.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:215b2ea50132ec511848df2426df713c1e63cdccb2fe7ac715cee5d1ef10a08e
|
3 |
+
size 20740611
|
checkpoint_p1/milestones/checkpoint_000008352_8552448.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4f3c7a6cd7127308e7a36cd8e93e7a5dfda76738a4a8ecfa85cd26820cd17a20
|
3 |
+
size 20741467
|
checkpoint_p1/milestones/checkpoint_000016896_17301504.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:39b754c2eabb52d7a99e4d8b0ab4c1a5f28b5fab597a2aa7150f10ebd4bc8821
|
3 |
+
size 20741523
|
checkpoint_p1/milestones/checkpoint_000025440_26050560.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:79f575a970a1848106649b8c8b80d11c00c8e8371eb827b7a13c8b096235d91c
|
3 |
+
size 20741523
|
checkpoint_p1/milestones/checkpoint_000034016_34832384.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:61789aff687e80ffcfcb7daf79e22c62383e2f13c21645c02cfe7b095441afac
|
3 |
+
size 20741523
|
checkpoint_p1/milestones/checkpoint_000042560_43581440.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e3aeac808ebb2249e0d1348154129ea9d38c026e0b8c0aab1bf3edb39fb8fa44
|
3 |
+
size 20741523
|
checkpoint_p1/milestones/checkpoint_000051104_52330496.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e012493557f3c64fe2a1e7979e7251504fbbcee2f1736aafcad278bf2f942e31
|
3 |
+
size 20741523
|
checkpoint_p1/milestones/checkpoint_000059648_61079552.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:17558fc8f22e1833c12649fe23c11aec94b896c78e17e6cda9ddf0b9277be336
|
3 |
+
size 20741523
|
checkpoint_p1/milestones/checkpoint_000068224_69861376.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5fdfaf4bfb38e09557568e16d7c2d0a92df1121d8bfaaf82c11c34a82762696a
|
3 |
+
size 20741523
|
checkpoint_p1/milestones/checkpoint_000076768_78610432.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:40f00e7ad6bca026b13deac6bcb2a0c2e1c6486c672b54400d6a4c719c9b0eba
|
3 |
+
size 20741523
|
checkpoint_p1/milestones/checkpoint_000085344_87392256.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fa858674562b483547fe55deb405c18f68df8393c30ed4c4b177aae8d78ed957
|
3 |
+
size 20741523
|
checkpoint_p1/milestones/checkpoint_000093888_96141312.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8a53ad4d041d31468cbaa1c221f687d438f684eaeef811053efb9cd967c59fb5
|
3 |
+
size 20741523
|
config.json
ADDED
@@ -0,0 +1,162 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"help": false,
|
3 |
+
"algo": "APPO",
|
4 |
+
"env": "atari_asterix",
|
5 |
+
"experiment": "atari_asterix_APPO",
|
6 |
+
"train_dir": "./train_atari",
|
7 |
+
"restart_behavior": "restart",
|
8 |
+
"device": "gpu",
|
9 |
+
"seed": 1234,
|
10 |
+
"num_policies": 2,
|
11 |
+
"async_rl": true,
|
12 |
+
"serial_mode": false,
|
13 |
+
"batched_sampling": true,
|
14 |
+
"num_batches_to_accumulate": 2,
|
15 |
+
"worker_num_splits": 1,
|
16 |
+
"policy_workers_per_policy": 1,
|
17 |
+
"max_policy_lag": 1000,
|
18 |
+
"num_workers": 16,
|
19 |
+
"num_envs_per_worker": 2,
|
20 |
+
"batch_size": 1024,
|
21 |
+
"num_batches_per_epoch": 8,
|
22 |
+
"num_epochs": 4,
|
23 |
+
"rollout": 128,
|
24 |
+
"recurrence": 1,
|
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": 0.0004677351413,
|
32 |
+
"value_loss_coeff": 0.5,
|
33 |
+
"kl_loss_coeff": 0.0,
|
34 |
+
"exploration_loss": "entropy",
|
35 |
+
"gae_lambda": 0.95,
|
36 |
+
"ppo_clip_ratio": 0.1,
|
37 |
+
"ppo_clip_value": 1.0,
|
38 |
+
"with_vtrace": false,
|
39 |
+
"vtrace_rho": 1.0,
|
40 |
+
"vtrace_c": 1.0,
|
41 |
+
"optimizer": "adam",
|
42 |
+
"adam_eps": 1e-05,
|
43 |
+
"adam_beta1": 0.9,
|
44 |
+
"adam_beta2": 0.999,
|
45 |
+
"max_grad_norm": 0.0,
|
46 |
+
"learning_rate": 0.0003033891184,
|
47 |
+
"lr_schedule": "linear_decay",
|
48 |
+
"lr_schedule_kl_threshold": 0.008,
|
49 |
+
"lr_adaptive_min": 1e-06,
|
50 |
+
"lr_adaptive_max": 0.01,
|
51 |
+
"obs_subtract_mean": 0.0,
|
52 |
+
"obs_scale": 255.0,
|
53 |
+
"normalize_input": true,
|
54 |
+
"normalize_input_keys": [
|
55 |
+
"obs"
|
56 |
+
],
|
57 |
+
"decorrelate_experience_max_seconds": 0,
|
58 |
+
"decorrelate_envs_on_one_worker": true,
|
59 |
+
"actor_worker_gpus": [],
|
60 |
+
"set_workers_cpu_affinity": true,
|
61 |
+
"force_envs_single_thread": false,
|
62 |
+
"default_niceness": 0,
|
63 |
+
"log_to_file": true,
|
64 |
+
"experiment_summaries_interval": 3,
|
65 |
+
"flush_summaries_interval": 30,
|
66 |
+
"stats_avg": 100,
|
67 |
+
"summaries_use_frameskip": true,
|
68 |
+
"heartbeat_interval": 10,
|
69 |
+
"heartbeat_reporting_interval": 60,
|
70 |
+
"train_for_env_steps": 100000000,
|
71 |
+
"train_for_seconds": 10000000000,
|
72 |
+
"save_every_sec": 120,
|
73 |
+
"keep_checkpoints": 2,
|
74 |
+
"load_checkpoint_kind": "latest",
|
75 |
+
"save_milestones_sec": 1200,
|
76 |
+
"save_best_every_sec": 5,
|
77 |
+
"save_best_metric": "reward",
|
78 |
+
"save_best_after": 100000,
|
79 |
+
"benchmark": false,
|
80 |
+
"encoder_mlp_layers": [
|
81 |
+
512,
|
82 |
+
512
|
83 |
+
],
|
84 |
+
"encoder_conv_architecture": "convnet_atari",
|
85 |
+
"encoder_conv_mlp_layers": [
|
86 |
+
512
|
87 |
+
],
|
88 |
+
"use_rnn": false,
|
89 |
+
"rnn_size": 512,
|
90 |
+
"rnn_type": "gru",
|
91 |
+
"rnn_num_layers": 1,
|
92 |
+
"decoder_mlp_layers": [],
|
93 |
+
"nonlinearity": "relu",
|
94 |
+
"policy_initialization": "orthogonal",
|
95 |
+
"policy_init_gain": 1.0,
|
96 |
+
"actor_critic_share_weights": true,
|
97 |
+
"adaptive_stddev": false,
|
98 |
+
"continuous_tanh_scale": 0.0,
|
99 |
+
"initial_stddev": 1.0,
|
100 |
+
"use_env_info_cache": false,
|
101 |
+
"env_gpu_actions": false,
|
102 |
+
"env_gpu_observations": true,
|
103 |
+
"env_frameskip": 4,
|
104 |
+
"env_framestack": 4,
|
105 |
+
"pixel_format": "CHW",
|
106 |
+
"use_record_episode_statistics": true,
|
107 |
+
"with_wandb": true,
|
108 |
+
"wandb_user": "matt-stammers",
|
109 |
+
"wandb_project": "atari_APPO",
|
110 |
+
"wandb_group": "atari_asterix",
|
111 |
+
"wandb_job_type": "SF",
|
112 |
+
"wandb_tags": [
|
113 |
+
"atari"
|
114 |
+
],
|
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 |
+
"command_line": "--algo=APPO --env=atari_asterix --experiment=atari_asterix_APPO --num_policies=2 --restart_behavior=restart --train_dir=./train_atari --train_for_env_steps=100000000 --seed=1234 --num_workers=16 --num_envs_per_worker=2 --num_batches_per_epoch=8 --async_rl=true --batched_sampling=true --batch_size=1024 --max_grad_norm=0 --learning_rate=0.0003033891184 --heartbeat_interval=10 --heartbeat_reporting_interval=60 --save_milestones_sec=1200 --num_epochs=4 --exploration_loss_coeff=0.0004677351413 --with_wandb=true --wandb_user=matt-stammers --wandb_project=atari_APPO --wandb_group=atari_asterix --wandb_job_type=SF --wandb_tags=atari",
|
128 |
+
"cli_args": {
|
129 |
+
"algo": "APPO",
|
130 |
+
"env": "atari_asterix",
|
131 |
+
"experiment": "atari_asterix_APPO",
|
132 |
+
"train_dir": "./train_atari",
|
133 |
+
"restart_behavior": "restart",
|
134 |
+
"seed": 1234,
|
135 |
+
"num_policies": 2,
|
136 |
+
"async_rl": true,
|
137 |
+
"batched_sampling": true,
|
138 |
+
"num_workers": 16,
|
139 |
+
"num_envs_per_worker": 2,
|
140 |
+
"batch_size": 1024,
|
141 |
+
"num_batches_per_epoch": 8,
|
142 |
+
"num_epochs": 4,
|
143 |
+
"exploration_loss_coeff": 0.0004677351413,
|
144 |
+
"max_grad_norm": 0.0,
|
145 |
+
"learning_rate": 0.0003033891184,
|
146 |
+
"heartbeat_interval": 10,
|
147 |
+
"heartbeat_reporting_interval": 60,
|
148 |
+
"train_for_env_steps": 100000000,
|
149 |
+
"save_milestones_sec": 1200,
|
150 |
+
"with_wandb": true,
|
151 |
+
"wandb_user": "matt-stammers",
|
152 |
+
"wandb_project": "atari_APPO",
|
153 |
+
"wandb_group": "atari_asterix",
|
154 |
+
"wandb_job_type": "SF",
|
155 |
+
"wandb_tags": [
|
156 |
+
"atari"
|
157 |
+
]
|
158 |
+
},
|
159 |
+
"git_hash": "5fff97c2f535da5987d358cdbe6927cccd43621e",
|
160 |
+
"git_repo_name": "not a git repository",
|
161 |
+
"wandb_unique_id": "atari_asterix_APPO_20231008_075004_357423"
|
162 |
+
}
|
git.diff
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3357904f421d3f4924836316b1741bf64d5dd0e807d5e80ac07059b4c52a7008
|
3 |
+
size 14426734
|
replay.mp4
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4892f6c3fa69ff9f157bd69b450237b96c02eb18b707ab8a7e00b37e12b5a2f9
|
3 |
+
size 3889614
|
sf_log.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|