pushing model
Browse files- README.md +12 -12
- cleanba_ppo_envpool_impala_atari_wrapper_naturecnn.cleanrl_model +2 -2
- events.out.tfevents.1677098987.ip-26-0-130-11 +0 -0
- events.out.tfevents.1677099833.ip-26-0-136-122 +3 -0
- replay.mp4 +0 -0
- videos/Breakout-v5__cleanba_ppo_envpool_impala_atari_wrapper_naturecnn__1__41dddc95-c7d1-4fc2-9abe-dac84077d081-eval/0.mp4 +0 -0
- videos/Breakout-v5__cleanba_ppo_envpool_impala_atari_wrapper_naturecnn__1__63369269-4fb3-436e-8ec5-03d7a6dee2ec-eval/0.mp4 +0 -0
README.md
CHANGED
@@ -16,7 +16,7 @@ model-index:
|
|
16 |
type: Breakout-v5
|
17 |
metrics:
|
18 |
- type: mean_reward
|
19 |
-
value:
|
20 |
name: mean_reward
|
21 |
verified: false
|
22 |
---
|
@@ -46,7 +46,7 @@ curl -OL https://huggingface.co/vwxyzjn/Breakout-v5-cleanba_ppo_envpool_impala_a
|
|
46 |
curl -OL https://huggingface.co/vwxyzjn/Breakout-v5-cleanba_ppo_envpool_impala_atari_wrapper_naturecnn-seed1/raw/main/pyproject.toml
|
47 |
curl -OL https://huggingface.co/vwxyzjn/Breakout-v5-cleanba_ppo_envpool_impala_atari_wrapper_naturecnn-seed1/raw/main/poetry.lock
|
48 |
poetry install --all-extras
|
49 |
-
python cleanba_ppo_envpool_impala_atari_wrapper_naturecnn.py --
|
50 |
```
|
51 |
|
52 |
# Hyperparameters
|
@@ -54,7 +54,7 @@ python cleanba_ppo_envpool_impala_atari_wrapper_naturecnn.py --local-num-envs 12
|
|
54 |
{'actor_device_ids': [0],
|
55 |
'actor_devices': ['gpu:0'],
|
56 |
'anneal_lr': True,
|
57 |
-
'async_batch_size':
|
58 |
'async_update': 3,
|
59 |
'batch_size': 15360,
|
60 |
'capture_video': False,
|
@@ -66,14 +66,14 @@ python cleanba_ppo_envpool_impala_atari_wrapper_naturecnn.py --local-num-envs 12
|
|
66 |
'exp_name': 'cleanba_ppo_envpool_impala_atari_wrapper_naturecnn',
|
67 |
'gae_lambda': 0.95,
|
68 |
'gamma': 0.99,
|
69 |
-
'global_learner_decices': ['gpu:1', 'gpu:
|
70 |
'hf_entity': '',
|
71 |
-
'learner_device_ids': [1
|
72 |
-
'learner_devices': ['gpu:1'
|
73 |
'learning_rate': 0.00025,
|
74 |
-
'local_batch_size':
|
75 |
-
'local_minibatch_size':
|
76 |
-
'local_num_envs':
|
77 |
'local_rank': 0,
|
78 |
'max_grad_norm': 0.5,
|
79 |
'minibatch_size': 3840,
|
@@ -81,20 +81,20 @@ python cleanba_ppo_envpool_impala_atari_wrapper_naturecnn.py --local-num-envs 12
|
|
81 |
'num_envs': 120,
|
82 |
'num_minibatches': 4,
|
83 |
'num_steps': 128,
|
84 |
-
'num_updates':
|
85 |
'profile': False,
|
86 |
'save_model': True,
|
87 |
'seed': 1,
|
88 |
'target_kl': None,
|
89 |
'test_actor_learner_throughput': False,
|
90 |
'torch_deterministic': True,
|
91 |
-
'total_timesteps':
|
92 |
'track': True,
|
93 |
'update_epochs': 4,
|
94 |
'upload_model': True,
|
95 |
'vf_coef': 0.5,
|
96 |
'wandb_entity': None,
|
97 |
'wandb_project_name': 'cleanRL',
|
98 |
-
'world_size':
|
99 |
```
|
100 |
|
|
|
16 |
type: Breakout-v5
|
17 |
metrics:
|
18 |
- type: mean_reward
|
19 |
+
value: 679.80 +/- 208.17
|
20 |
name: mean_reward
|
21 |
verified: false
|
22 |
---
|
|
|
46 |
curl -OL https://huggingface.co/vwxyzjn/Breakout-v5-cleanba_ppo_envpool_impala_atari_wrapper_naturecnn-seed1/raw/main/pyproject.toml
|
47 |
curl -OL https://huggingface.co/vwxyzjn/Breakout-v5-cleanba_ppo_envpool_impala_atari_wrapper_naturecnn-seed1/raw/main/poetry.lock
|
48 |
poetry install --all-extras
|
49 |
+
python cleanba_ppo_envpool_impala_atari_wrapper_naturecnn.py --distributed --learner-device-ids 1 --track --save-model --upload-model --env-id Breakout-v5 --seed 1
|
50 |
```
|
51 |
|
52 |
# Hyperparameters
|
|
|
54 |
{'actor_device_ids': [0],
|
55 |
'actor_devices': ['gpu:0'],
|
56 |
'anneal_lr': True,
|
57 |
+
'async_batch_size': 20,
|
58 |
'async_update': 3,
|
59 |
'batch_size': 15360,
|
60 |
'capture_video': False,
|
|
|
66 |
'exp_name': 'cleanba_ppo_envpool_impala_atari_wrapper_naturecnn',
|
67 |
'gae_lambda': 0.95,
|
68 |
'gamma': 0.99,
|
69 |
+
'global_learner_decices': ['gpu:1', 'gpu:3'],
|
70 |
'hf_entity': '',
|
71 |
+
'learner_device_ids': [1],
|
72 |
+
'learner_devices': ['gpu:1'],
|
73 |
'learning_rate': 0.00025,
|
74 |
+
'local_batch_size': 7680,
|
75 |
+
'local_minibatch_size': 1920,
|
76 |
+
'local_num_envs': 60,
|
77 |
'local_rank': 0,
|
78 |
'max_grad_norm': 0.5,
|
79 |
'minibatch_size': 3840,
|
|
|
81 |
'num_envs': 120,
|
82 |
'num_minibatches': 4,
|
83 |
'num_steps': 128,
|
84 |
+
'num_updates': 3255,
|
85 |
'profile': False,
|
86 |
'save_model': True,
|
87 |
'seed': 1,
|
88 |
'target_kl': None,
|
89 |
'test_actor_learner_throughput': False,
|
90 |
'torch_deterministic': True,
|
91 |
+
'total_timesteps': 50000000,
|
92 |
'track': True,
|
93 |
'update_epochs': 4,
|
94 |
'upload_model': True,
|
95 |
'vf_coef': 0.5,
|
96 |
'wandb_entity': None,
|
97 |
'wandb_project_name': 'cleanRL',
|
98 |
+
'world_size': 2}
|
99 |
```
|
100 |
|
cleanba_ppo_envpool_impala_atari_wrapper_naturecnn.cleanrl_model
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9a54c3a00b863c60d60fc8d6c2c3cca5c4c49b9b6b95931d9479bfec2bc7e3f6
|
3 |
+
size 6747984
|
events.out.tfevents.1677098987.ip-26-0-130-11
DELETED
File without changes
|
events.out.tfevents.1677099833.ip-26-0-136-122
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:66d528aeae85e6eebb5ed4b5ffbebd397284baf70eb765e14c20747e6d629183
|
3 |
+
size 4754155
|
replay.mp4
CHANGED
Binary files a/replay.mp4 and b/replay.mp4 differ
|
|
videos/Breakout-v5__cleanba_ppo_envpool_impala_atari_wrapper_naturecnn__1__41dddc95-c7d1-4fc2-9abe-dac84077d081-eval/0.mp4
ADDED
Binary file (488 kB). View file
|
|
videos/Breakout-v5__cleanba_ppo_envpool_impala_atari_wrapper_naturecnn__1__63369269-4fb3-436e-8ec5-03d7a6dee2ec-eval/0.mp4
DELETED
Binary file (6.89 kB)
|
|